Professional Documents
Culture Documents
contains 59 pages
ABOUTMID
The interestsof MIDcover the wholefield of marine technology (i.e. all aspects of
engineering technology and sciencerelating tothe sea and to the exploitation and
exploration ofthe sea, both belowand above the seabed).
REVIEW OF REPAIRS TO
OFFSHORE STRUCTURES
AND PIPELINES
PUBLICATION 94/102
Published by MTD
The Marine Technology Directorate Limited
© MTh 1994
ISBN: 1 870553 18 7
2 MTDPublication 94/102
FOREWORD
This report results from a major review of repairs,which was initiated by The MarineTechnology
Directorate Limited as a multi-sponsor project funded by the organisations listed below.
The project was carried out under contractto MTh by Mr M Hordyk,Mr S Morahan and Dr C J
Billington of Billington Osborne-Moss Engineering LW, with sub-contractwork by Mr J J S
Daniel of Jeremy Daniel and Co Ltd. Three of them were involved in the previous study, and their
involvement in the new study provided continuity and a considerable understanding of the changes
that had occurred.
The project was conductedunder the guidance of a Steering Group which comprised:
The raw data on repairs, collected from oil and gas operators as part of the project, remain
confidential to the funding organisations listed above. However, a Lotus 123 diskversion of the
structural repairs data (not identifying any repair to a specific platform or operator) is available
from MTD for those recipients of this report who require further statistical information from the
survey. The report is a summary and analysis of those findings, and it has been released after a
period of confidentiality to the sponsors, following completion of the project.
Every reasonable effort has been madeto ensure that this publication accurately reflects the information collected during the course of
the survey. However, no liability is assumed by MTD, BOMEL or the sponsors for the contents ofthis report,nor does it necessarily
reflect the viewsor policy of any of the parties concerned.
Repairs review 3
1
4 MTD Publication94/102
CONTENTS
Page no.
List of illustrations 6
List of tables 6
Summary 7
1. Introduction 7
2. Scope and methodology 8
3. Structural repairs and strengthening on the North West European
Continental Shelf 12
6. Repair systems used on steel structures located on the North West European
Continental Shelf 27
11. Pipeline repairs and strengthening on the North West European Continental Shelf 40
12. Lessons learnt from completed repair work on the North West European
Continental Shelf 46
15. Recommendations 50
References 51
Repairsreview 5
LIST OF ILLUSTRATIONS
LIST OF TABLES
6 MTD Publication94/102
SUMMARY
The study covered the sub-sea strengthening and repair of structures and pipelines used
for, or associated with, the production, storage and transportation of hydrocarbons in the
marine environment. This report therefore covers repairs on:
The review is the most comprehensive study of its type yet undertaken with 172 repairs
recorded, and it builds on the earlier 1982 review.
The data are analysed in various ways, including: cumulative total of platforms by
construction material per installation year, cumulative repairs per national sector
population, frequency of repairs per structure year versus water depth, causes of
damage,repair types, numberof repairs versus structure age, and reported costs.
The report contains conclusions and makes recommendations based on the findings. A
Lotus 1-2-3 diskette, containing data on the structural repairs, is also available from MTh.
1. INTRODUCTION
In 1983, UEG published a review of repairs to structures in the North Sea'. This was the
first comprehensive study on the causes of damage to offshore structures and it considered
the various techniques used to repair the damage and the experience of using these
techniques for some 60 underwater repairs which had been carried out up to that time.
The review was undertaken in 1982 at a time when the effects of fatigue loading on
offshore structures were becoming apparent and were under active study. It was also at a
time when the technology of repair systems was not widely understood and was held by
only a limited number of engineering and technical contractors. The document was
limited in a small way by a certain reluctance within the offshore industry to discuss
publicly work on repair systems.
The review was widely circulated and received considerable acclaim. It was used by some
operating companies as a primer upon which they developed a corporate strategy for
offshore repairs.
By the end of the 1980s, a numberof operating companies believed that it was time to
undertake a second review. The intervening years had, for several reasons, produced a
new crop of repairs. Many structures were approaching the end of their original design
lives and might thus he more susceptible to damage. The fatigue design of structures had
Repairs review 7
become better understood, but older structures might now be sufferingother problems such
as corrosion. The newer structures with fewer bracing members and possibly with less
structural redundancy, and with members being designed to more precisely defined design
limits, might be introducing new problems. The new survey would be able to identify
these. It would also be possible to review the performance of earlier repairs, some of
which were themselves nearly 20 years old.
The data gathering work for the new study was undertaken in 1992, and the project
extended the first review in the following three ways:
To be included in this review, the primary definition that a repair or strengthening system
had to meet was that the remedial work was not planned during the design stages of the
project. For example, if anodes were replaced as planned, the event was not included.
However, if the anodes were found to be depleting faster than anticipated, the works
required to correct the deficiency in the cathodic protection system would be included.
Although repairs to topsides were specifically excluded from the study, splash-zone
damage was included, even when the repair was above the water line. The working
definition adopted here was that the repair was included if sea conditions affected the
repair activities. Thus some repairs to module support frames were included in the
review. A problem of definition arose when a flare boom was damaged by a vessel
impact, although in that case other structural damage ensured that the repair was included
in the review.
Repairs to non-structural sub-sea elements were included, because they involve all the
complexities of working at or below the sea surface. Often, there are structural
consequences too (pile guides for example attract wave and current loadings, and caissons
8 MTDPublication 94/102
deplete the cathodic protection system). These definitions set the boundaries of the study
work.
It was originally anticipated that all vessel impacts would be recorded. It was argued that
a significantstructural defect or one only requiring a touch-up coat of paint could be the
result of a similar cause. It was therefore important to understand the causes behind all
such incidents. However, during the field work, it became apparent that minor vessel
impacts are very frequent, and that the background to minor events is rarely recorded.
Thus only vessel impacts which resulted in a repair are recorded in this review.
Cracks in the sub-sea structure are also frequently found. If structural analyses show that
the crack is unlikely to have resulted from a static or fatigue underdesign, it is common
practice to grind Out the crack and to inspect it subsequently to monitor its development.
These incidents are part of routine maintenance, and they are also excluded from this
review.
Information was sought from each of the 38 companies. Some companies covered their
national operations from more than one location, and a total of 41 contacts was made to
include all the operating centres. Table 1 gives details of responses to the field survey
work.
Number of interviews 30
Covered by post/telephone etc. 2
Not interviewed - nothing relevant 8
Declinedinterview or appointment 1
TOTAL 41
Operators who had recorded repairs in the 1983 reviewWwere specifically questioned on
the subsequentperformance of those earlier repairs.
Repairsreview 9
The responses on the questionnaires were organised on a database*, and they are analysed
in this review. The responses have been checked against literature surveys, also against
the CODAM database for major structural repairs compiled by the Norwegian Petroleum
Directorate.
A substantial application of this principle occurred when considering repairs outside the
NWECS. A unique cause affected 22 separate installations, resulting in over 100 repair
sites. Again, these were recorded as one event. The principle is extended to fatigue
damage to conductor guide frames. Separate events are recorded only where there is a
difference in the history of the repair (e.g. in the manner in which it was discovered).
The total number of repairs included on the database is discussed in Section 3. The
multiple repairs recorded as one event (such as the fabrication fault repeated in 155
locations) are also listed in that Section.
2.4 Structural repairs outside the North West European Continental Shelf
Knowledge of repairs to 16 facilities outside the NWECS is held by BOMEL and by some
of the operatingcompanies participating in the study. The questionnaire used for the main
structural survey was completed for these repairs by BOMEL or by the individual
participant.
It was not possible to produce statistical analyses based on these repairs, because the
responses are too few and too biased towards a limited numberof respondents. However,
the responses do allow comparisons to be made between structures on the NWECS and
elsewhere in the world. In turn, these help to determine any differences in the causes of
damage or the nature of the repair systems used.
A full survey of all the companies controlling pipelines on the NWECS was not possible,
because of a limitation on project funds. Therefore, a number of operators were selected
to he interviewed. The selection procedure ensured that the operators who were
approached controlled a representative sample of pipelines.
According to the INFIELD database2, a total of 14900 km of pipeline had been laid on
the NWECS by the end of 1991. Shell, Statoil and BP account for 34% of this total, and
10 MTDPublication 94/102
the top ten operators (Shell, Statoil and BP, together with the operators controlling the
next seven largest total lengths of pipeline) account for 68% of the total.
Details of the 10100 km of line laid by the top 10 operators were investigated to ensure
that the parametric variations of pipeline diameter, material, coating, installation method,
trenching and product carried were all adequately covered. The analysis showedthat
limiting the numberto the top 10 companies adequately covered the major variables
chosen.
The interviews were conducted in the same manner as in the structural survey (see Section
2.3). A questionnaire was prepared and is reproduced in AppendixB.
The database prepared from the responses is presented in Appendix C, and analysed in this
review. The responses have been checked against literature surveys and against the
CODAM database for major pipeline incidents compiled by the Norwegian Petroleum
Directorate.
The Piper Alpha disaster has not been included. Repairs to Piper Alpha recorded in the
1983 review1 are included in the database, but none since that date.
The response to the questionnaires was varied. Those interviewed tended to be familiar
with repairs or strengthening systems with which they had been personally involved, but
few of them had continued with this type of work over many years. Some organisations
opened their records fully and gave all the information they had to the study. Others
limited their responses to the interviewees'memory. In some instances, only poor details
were available for rather large repair projects, even key dates being unknown. In such
cases, these details were supplemented as far as possible from BOMEL's in-house records.
Repairs review 11
Some parts of the questionnaires were answered less well than others. Most respondents
knew the background to the repairs and details of their installation. Less detail was
available on any alternative repair systems which had been considered and rejected, and
little was stated about any difficulties with the design or fabrication of the repair elements.
It may be that the fabricators themselves have views on the buildability of the designs of
the repair systems, but the fieldwork for this study was confined to the operators.
The general response to the structural survey was very good, those interviewed recognising
the usefulness of the study and keen to ensure that their work and views were recorded.
This contrasts with the fieldwork for the 1983 review', when sometimes a fear was
expressed that information on repairs could reflect badly on the company involved.
The response to the pipeline survey was more cautious, and in several cases the level of
detail made available was significantly lower than for the structural repairs.
By the end of 1991, there were 352 steel and 23 concrete platforms located on the
NWECS. Table 2 shows the numberof repairs for which information is contained in the
database. Analyses of the data generally relate to the date of the repair or the date on
which the incidentwas discovered. Table 2 shows that this information was not available
for all cases. The relatively low numberof datapoints for the date at which the damage
was caused is explained in part by deterioration caused by fatigue which cannot be
categorised in this way, also by the database including structural upgradings not
necessitated by damage.
Some analyses of the data are related to the frequency of repair, which provides an
indication of the probability that a structure will require a repair in any year. The
frequency of repair is the numberof repairs divided by the total exposure of similar
platforms expressed as structure years; the technique was previously used by Sharp4, and
Table 2 also gives this total life exposure as determined from the INFIELD database.
12 MTDPublication 94/102
Figure 1 shows that fewer repairs were undertaken during the 1980s compared to earlier
years, in relation to the total platform population. A tabulation of the values used to
prepare Figure 1 is presented as Appendix D.
Figure 1 has two peaks, both in the years preceding those in which the two repair reviews
were carried out. This could indicate a tendency to recall best only the most recent
events, as suggested in Section 2.7, or could indicate that more repairs are required after
periods of severe weather. However, a study of meteorological records of north west
Europe could draw no link between storm dates and the number of damaged offshore
structures.
400 50
40
300
Ca
a)
>-
-c
30 C-)
(-I,
CS
200 a-
a)
20 0
a)
-o
E
100 z
10
0 0
1965 1970 1975 1980 1985 1990 1995
Year Of Instatation (LeftHand Axis); Year Of Repair (Right Hand Axis)
... Steel _. Number Of Repairs .._ Concrete
It is thus possible that the repairs recorded in the years between the peaks represent a
lower bound, and that more regular surveys are required to accurately determine the
numberof repairs carried out each year.
Repairsreview 13
Some of the 172 repairs recorded in Table 2 actually consisted of multiple repairs.
Fatigue damage to a single conductorguide frame, for example, generally results in more
than one repair location on that frame. Table 3 shows that considerably more than 172
repairs have been undertaken on the NWECS, hut for the purposes of data analysis repairs
have been recorded in this review as single events if they result from a single cause. To
do otherwise would distort the analyses (e.g. a false conclusion would be drawn on the
importance of fabrication in causing the deterioration of offshore structures if the 155
fabrication faults listed in Table 3, and which resulted from a single poor detail, were
recorded individually).
Figure 2 indicates that the number of repairs completed in Norway and the UK are at a
comparable level when compared to the numberof structures in each sector. The
Netherlands has a low incidence of repair, although this could he a result of the survey
methodology. The itinerary for the 1982 survey was limited to the UK and Scandinavia,
despite the Netherlands having, by 1980, over three times as many platforms as Denmark
which was included on the itinerary. It is possible that the 1992 survey has failed to
identify some early Dutch repairs. The one operator who was not interviewed in 1992
might have added significantly to the total of repairs in the Dutch sector, although if that
operator's platforms are removed from the Dutch sector population, the plot presented for
the Netherlands in Figure 2 is increased only slightly.
14 MTD Publication94/102
The Danish sector also has a low incidence of repair. It is believed that some early repair
work in the Danish sector was not recorded at all. This omission would influence the
statistics. The Irish sector is not included in this analysis, because its two platforms could
not produce any statistically meaningful results. No repairs were reported in the German
sector.
The Dutch and Danish sectors have shallower mean water depths than the northern UK
and Norwegian sectors. National repair frequency might reflect the fact that more repairs
are required for structures located in deeper waters, and Figure 3 comparesthe frequency
of repair with water depth. Structures in deeper water have more structural members than
shallow water structures, and they thus have a greater chance of incurring damage. Figure
3 confirms that the frequency of repair for deeper water structures increases at a greater
rate than the increase in the water depth.
Figure 4 charts the relative performance of steel and concrete structures. It shows how
the annual numberof repairs to steel structures remained static during the 1980s in
proportion to the total population, although there appears to be a slight increase since
1990. From the early to the mid 1980s the numberof repairs to concrete and steel
structures related to their respective populations was more or less equal. Since 1987,
concrete structures have required a higher rate of repair.
0.7
.2 0.6
a)
a
0
a-
•a
a)
0.4
o 0.3
a)
.0
E
0.2
.
0
1965 1970 1975 1980 1985 1990 1995
Repair Year
UK N OK NL
Repairsreview 15
Concrete structures have steel components, including caissons and, in some cases,
substantial conductor guide frame assemblies. Repairs to these elements are included
within the concrete structure repair values presented in Figure 4 (as they are for the steel
structure repair values). Five of the 14 repairs to concrete structures were to steel
elements of this type, hut, even if these repairs are not included in the concrete repair
values, it is shown that neither basic structural material has a repair advantage. This is
perhaps a surprising result in view of the commonly held opinion that concrete is a more
durable material in the marine environment.
Leaks in concrete structures through construction joints and grouting ducts have taken
time to manifest themselves, and this could explain why initially there were fewer repairs
to concretestructures than to steel structures. At least one of the early concrete structures
now has significant problems of deterioration.
Steel skeletal structures consist of members with different levels of structural importance
often designated as primary or secondary structure. Analysis of the repairs database
indicates any increased susceptibility to damage for, say, members of lesser importance,
resulting perhaps from a reduced design input. A similar analysis of concrete structures is
not possible. All of a concrete structure can be considered as primary, hut it is
nonetheless able to tolerate severe local damage (e.g. a hole caused by a vessel impact)
without effect on its structural integrity. The analysisof the elements affected is thus
limited to steel structures.
0.15
(0
a)
>-
a)
0)
a..
(J)
(0
a.
a)
0
0.05
0
0—30 30—60 60—90 90—120 120+
Water Depth (m)
16 MTD Publication94/102
0.6
0
Cu
a-
0
a-
Cu
0.4
Cu
Cl)
a
Cu
0.3
0
a)
.0
E 0.2
z0)
>
0
1965 1970 1975 1980 1985 1990 1995
Repair Year
—s— Concrete structures - Including Steel Components —0-- Steel Structures
—.— Concrete structures - Steel Components Excluded
Table 4 shows that the number of repairs to primary structural elements has declined since
1974, but that since 1985 the numberof repairs to appurtenances has more than tripled.
Primary structure is defined as main structural elements (legs and bracings most usually in
the vertical planes). Secondary structures are typically members which transfer loads to
the main structural elements, such as conductor guide frames. Appurtenances are elements
such as caissons or pile guides, and include their direct supports. Members in the main
structural frame are includedas primary, even if analysis shows that structural integrity is
maintained with the member removed. Repairs to sub-sea structures and those made
necessary by fabrication aids have been shown separately in Table 4.
74-79 24 75 17 8 0 0
80 - 82 38 63 24 13 0 0
83-85 23 61 26 9 4 0
86-88 20 50 25 25 0 0
89 - early 92 33 36 24 27 3 9
Total 138
Note: Two additionalrepairs were earned out before 1974, and for one repairit was
not known precisely which member had been affected
Repairsreview 17
3.5 Changes of approach to repairs since the 1982 survey
Since the first survey in 1982, many papers and articles have been published on the causes
of structural damage and on the various repair methods. Several companies have
developed corporate procedures for evaluating damage to structures. The typical current
reaction to the discovery of damage is to measure the damage, to understand its cause, to
assess its consequences, and then to analyse options including repair options. This
methodical approach is carried out at an appropriate speed, and it contrasts favourably
with responses in the 1982 survey when those interviewed told of over-hasty reaction to
the discovery of damage.
Table 5 demonstrates that most damage is discovered either as it happens (usually the case
with vessel impact) or by routine or non-routine inspection. Routine inspections include
flooded member detection inspections, which may sometimes yield unexpected results but
which are not considered to he chance discoveries, because damage is found in the course
of a planned inspection regime. There are few datapoints within each heading of Table 5,
and no distinct trends are discernible.
In two cases, damage was discovered in non-routine inspections: one made following the
discovery of corrosion on a similar platform, the other a commissioning inspection
following a winter installation break. Chance discoveries of damage were generally made
during visual inspections of other structural elements. Many repairs follow structural re-
analyses to determine the effects of increased topsides loading or of more severe
environmental loads introduced by revisions to design guidance. Damage, as such, is not
discovered, so these repairs are not included in Table 5. Also, the date on which the
damage is caused is generally not known for defects detected by chance, so that few of
these incidents are included in Table 5. The most recentevent recorded as a chance
discovery concerned damage to a cover to a sub-sea housing which was discovered after
installation work resumed following a winter break.
Table 6 demonstrates that chance discoveries of damage are not common, that corrosion
and fatigue damage are usually discovered by routine examination, and that the effect of
most accidents is discovered immediately. In several cases, those interviewed did not
know how the incident was discovered, and these are recorded as unknown in Table 6.
However, it is likely that all five incidents recorded under unknown but caused by
accident (vessel impacts or dropped objects) were known about immediately they occurred.
69-76 13 54 23 0 8 15
77-79 14 79 21 0 0 0
80-82 16 75 6 0 0 19
83-85 9 67 22 11 0 0
86-88 12 50 42 0 0 8
89 - early 92 11 46 27 9 9 9
Total 75
Note: There are only 76 datapoints for which the date that the incident occurred is known.
One of these incidentsof damage occurredbefore 1969.
Corrosion 13 0 62 8 8 23
Fatigue 40 0 63 3 8 28
Vessel impact 37 68 19 0 5 8
Dropped objects 16 75 13 0 0 13
Total 106
Five problems requiring repair were discovered more than 12 months after the damage
was caused. Each of these incidents is unusual and is described below:
• A temporary launch member was removed sub-sea following the placing of the jacket
on the seabed. Damage was caused to the permanent structural members, to which it
had been attached. These damage locations were not routinely surveyedbecause, in
the permanent condition, they were simply straight members with no connecting
nodes.
• An operator took over an existing platform and discoveredcollision damage which had
been caused earlier.
• Damage was discovered during commissioning trials which took place some time after
the facility was first installed.
• Installation damage on a similar facility led to a non-routine inspection of an existing
structure.
Repairsreview 19
5. CAUSES OF DAMAGE TO STRUCTURES ON THE NORTH WEST
EUROPEAN CONTINENTAL SHELF
5.1 General
Several causes of damage have been reported, and the frequency with which they might
be expected on a structure in any given year is presented in Tables 7 and 8. Damage
listed under the heading of 'other" in Tables 7 and 8 includes diver error, scour,
prematurely depleting anodes, a mooring line failure, a loose duct, waves reflected by an
adjacent structure, and a pipeline pushing into a structure.
Fatigue 39 0.0101
Vessel impact 36 0.0094
Dropped object
-installation 8 0.0021
- operating 6 0.0016
Fabrication Fault 12 0.0031
Installation fault 12 0.0031
Corrosion 10 0.0026
Design Fault 9 0.0023
Operatingfault 4 0.0010
Design upgrade 11 0.0029
Other 8 0.0021
Unknown 3 0.0003
Steel components:
corrosion 3 0.0113
fatigue 1 0.0038
- operations 1 0.0038
Constructionfault 3 0.0113
Droppedobject
- installation 0 0
- operations 2 0.0075
Vessel impact 1 0.0038
Design fault 1 0.0038
Other 2 0.0075
Total 14 0.0526
The most distinct trend with time is that repairs necessary to remedy corrosion damage
are increasing.
20 MTDPublication94/102
Date of repair Numberof Percentage oftotalfor eadi timeband:
incidenis Fatigue Vessel Dropped Fabrication Corrosion Otberl
impact object & installation Unknown
74-76 5 40 20 20 0 0 20
77 - 79 19 26 16 21 16 0 21
80-82 38 34 39 5 8 0 13
83 - 85 23 25 17 4 35 4 17
86 -88 20 21 20 10 10 5 30
89 - early 92 34 12 3 21 15 26
Total 139
Figure 5 indicates that, in general, structures installed over the nine years since the 1982
survey appear to have been designed adequately against fatigue. This follows the
extensive research into the fatigue resistance of steel tubular joints undertaken in the late
1970s and early 1980s. The effectiveness with which the fatigue problems associatedwith
the earlier generations of offshore jackets appear to have been controlled is even more
noticeable when the number of fatigue repairs is related to the platform population (see
Figure 6).
The most recently installed structure to have fatigue problems had solid plating across
horizontal guide frame panels. This is a surprisingly fatigue sensitive detail to be included
on a structure installed as late as 1983. The operator of this facility reported that the
designer contributed towards the cost of the repair.
The improved design techniques do not mean that there are no longer any repairs for
fatigue damage. Table 9 shows that fatigue repairs still account for more than 20% of the
annual total, and as time goes by it is probable that other fatigue sensitive details will
manifest themselves on the earlier structures. It is also quite possible that more recently
designed structures will exhibit fatigue damage in future years.
However, the understanding of the fatigue phenomenon in offshore structures is now such
that the gross errors which led to joints with very short fatigue lives in many early
structures have been eliminated.
Repairs review 21
15
C')
10
a
a)
a)
30)
a)
U-
0
0a)
z35
E
0
65 67 69 71 73 75 77 79 81 83 85 87 89 91
66 68 70 72 74 76 78 80 82 84 86 88 90 92
Installation Year
0.12
0.1
C
0
(U
0-
3
(I)
0.08
a
(U
0.06
0.04
0.02
0
65 67 69 71 73 75 77 79 81 83 85 87 89 91
66 68 70 72 74 76 78 80 82 84 86 88 90 92
Installation Year
Enant 2 1 0 1 0 4 0.0010
Operational 13 4 0 1 1 19 0.0046
Work 3 1 1 0 0 5 0.0012
Not known 6 2 1 0 0 9 0.0022
73-80 9 67 11 11 11
81-84 14 71 22 7 0
85-88 6 67 33 0 0
89 - early 92 4 50 25 0 25
Total 33
Repairsreview 23
Details of a submarinecollisionwith a jacket structure have been reported elsewhere,
and the resulting repair to the structure is included here. It is important to note that
submarines navigate sub-sea by various forms of dead reckoning, and that they do not
usually have the opportunity to make visual or sonar contact with a structure.
Operational vessel impacts often occur when wave action sweeps the vessel into the
structure. One impact was thought to have been caused by a standby vessel practising
manoeuvres actually inside the jacket structure. Operational procedures were changed
following this incident.
One Dutch operator reported that he expected one or two vessel impacts per structure year,
90% of which cause damage to the vessel although not to the structure. This anticipated
frequency of 1.0 or 2.0 compares with the overall frequency of 0.0090 taken from Table
10 for those impacts which actually cause damage to the offshore structures. However, a
UK operator in the southern North Sea reported that although vessel impacts are a
constant possibility, there was no expectation at all that structures would suffer impacts on
a regular basis.
5.4 Dropped objects
At least 12 of the 16 instances of damage caused by dropped objects occurred before
1982. Of these, at least seven, and possibly as many as 11, occurred during the main
installation phase. Before 1982, the single known case of a dropped object causing
damage once the offshore facility was in operation concerned a drill collar which
penetrated the drill deck and damaged the jacket structure below the water line.
The three most recent incidents of damage resulted from more complex situations. In one
case, an incorrectly machined thread resulted in a length of conductor droppingaway
through the structure. In another case, a caisson fitted within its guides during construction
was dropped as it was pulled through to its final location during offshore hook-up. In the
third case, damage was caused by a poorly controlled pile hammer hitting a structure.
It appears that over the last decade there has been an improvement in offshore handling
techniques, although it is difficult to draw conclusive trends, because the total numberof
repairs arising from dropped objects is relatively small.
Other serious installation problems have arisen during the movement of jackets, including
skidding them on and off the launch barge and manoeuvring them to mate with pre-
located sub-sea template structures.
24 MTD Publication94/102
Some fabrication faults indicate a failure to apportion responsibilities between the designer
and the fabricator. In one case, serious damage was caused by the designer calling for a
double-sided closure weld but omitting to design and detail welder access manholes. The
fabricator's manhole detail subsequently caused distress to the main structure.
A similar lack of contact between the designer and the fabricator has led to fatigue
problems initiated by wind-induced vibrations of non-supported members during
construction in the fabricator's yard, also to fabricator's walkways remainingfixed to the
structure, causing structural problems once the facility was submerged. One operator
called for a more sophisticated interaction between the designer and the fabricator than
that achieved by current contractual arrangements. This echoed the general message from
an earlier study carried out by BOMEL on integrating the fabricator into the design
process8.
5.7 Corrosion
Corrosion problems resulting in the need for repair have been increasing (see Table 9).
There have been two main problems: a degradation of some cathodic protection systems,
and the specific problem of bi-metallic corrosion. In addition, some impressed current
systems and hybrid systems with impressed currents have been found to be unsatisfactory
and are being replaced by sacrificial anode systems.
The sensitivity of the cathodic protection regime to adjacent structural elements and to the
local environment has not always been sufficiently appreciated during the design. In one
example, open-topped piles (which is how piles are generally left post-installation) were
Repairsreview 25
covered to prevent the cathodic protection system draining to the inside of the piles. In
another example, the effect of the rate of flow of water past the anodes had not been fully
allowed for.
A more severe problem has been caused by the replacement of original cast iron down-
hole caisson pumps with pumps manufactured in non-ferrous metals. This has caused
extreme bi-metallic corrosion to occur, with caisson walls corroded through within a short
period. One operator reported that a pump manufacturer actually stated in his technical
brochure that the caisson would act as the pump's anode. Internally-lined caissons have
also suffered from this form of corrosion, the spiders to the pump having worn away the
lining at their points of contact.
Three concrete repairs have involved the flow of sea water into the structure through a
construction joint or through a grouted prestressing duct. These faults can take time to
manifest themselves before first appearingas a damp patch on the inside face of the
concrete structure and then progressingto a flow of water.
One operator reported that the transition zone between the concretesub-structure and the
steel deck required attention offshore, but insufficient details were given for this work to
be included in the repairs database.
6.1 General
All-the repair techniques described in the 1983 review' remain in use, and no new
techniques have been introduced. Wet weldinghas now been used for a structural repair.
Table 12 indicates that clamps have been used for over 40% of all repairs and for just less
than 50% of all sub-sea repairs. Welding from scaffolding(even in one case from a
dinghy) has been used for 66% of repairs above the water line, and habitat or wet welding
has been used for 35% of repairs beyond 25m and 38% of repairs beyond 50m.
26 MTD Publication94/102
Percentage oftotalfor each band of waterdepth:
Water depth, Number of Mechanical Grouted Air Cofferdain Hyperbaric Wet weld GrouSed Boltsplates
m incidents clamps clamps weld weld weld members
o &above 48 13 17 60 6 0 0 0 4
o - 10 32 31 19 6 16 19 0 9 0
-
10 25 30 (2) 40 (7) 23 0 0 23 7 7 0
-
25 50 21 (3) 24 (5) 24 (5) 0 0 33 (5) 0 19 0
50- 100 7 (1) 0 29 0 0 14(14) 14 14 29
100 & below 6 (1) 33 0 0 0 33(17) 0 17 17
Total 144
Note: The numbersin brackets are the repairsearned outat the mud line and are included within theunbrackctednumber.
Table 13 analyses the use of different repair techniques with time. The only distinct trend
shown is that wet welding has been used since 1986. Another possible trend is that during
the early and mid 1980s there was a move from clamped repairs to welded repairs, but
statistics from the past 8 years indicate a rough parity between the two techniques.
Date of Number of Mechanical Grouted Air Cofferdam Hyperbaric Wet Grouted Bolts
repair incidents clamps clamps weld weld weld weld members plates
74 - 76 8 38 12 8 0 12 0 0 0
77 - 79 21 10 19 24 0 24 0 10 14
80 - 82 38 34 21 18 8 11 0 8 0
83 - 85 20 15 20 15 10 15 0 15 10
86 - 88 17 18 18 24 6 18 12 6 0
89 - 91 21 29 10 38 0 10 10 5 10
Total 125
Note: Some repairs include more than one repair technique, and each technique has been countedindividually in this
Table. The total does not, therefore, tally with thosegiven in Table 2.
Lang et al.9 attempted to rationalise the repair options for fatigue damage to conductor
guide frames on platforms in the Gulf of Mexico. Table 14 summarises their findings.
Weldingtechniques were not included, because it was argued that the quality of repair
would probably be lower than that achieved during the original fabrication. It is generally
necessary to actually improve the fatigue life of a damaged section, and simply repairinga
weld would only get back to the original design.
Repairs review 27
Repair technique Effect of repair Increase in fatigue life, multiplying
- factor
* Theseare typical values. The Stressconcentration factorscan he further reduced by using heavierclampa.
Each repair situation is unique, and physical limitations often preclude certain options,
even before costs are considered. At least one operator spoke of the need to avoid sub-sea
repairs, on grounds not of cost but of safety. This operator did not use divers if it was at
all possible to use analytical techniques to avoid or reduce sub-sea work.
6.2 Clamps
Clamps are split sleeves joined together by longitudinal bolted flanges. Load is
transferred by friction from the original member to the clamp which bridges the load
across the damaged or understrength section. Clamps can be formed in complex shapes to
cover nodal joints in two, three or more planes. Repair clamps as heavy as 22 tonnes and
using 149 bolts have been successfully installed. The efficiency of the clamps in
transferring load has been proven by many full-size laboratory trials.
It is not generally necessary to repair the defect itself when using clamps, because the
clamp provides an alternative load path. However, a careful assessment has to be made to
ensure for example that a crack will remain stable beneath the clamp. This is particularly
important when clamps are designed to share the load with the existing member, a
technique used when there is not enough room for a full-load transferclamp.
The technology of clamp design and installation is well known. Early problems with the
sub-sea sealing of the ends of the clamp to contain the grout have been solved, and bolts
manufactured from L7 steel have overcome the stress corrosion associated with the first
bolts used sub-sea. It is common practice to use long bolts in order to reduce the strain
resulting from fatigue loadings.
28 MTD Publication94/102
Clampsare of three basic forms.
1. A mechanical clamp transfers load by friction generated by stressing together the
contact faces of the original member and the clamp. Thin neoprene or rubber linings
are sometimes introduced between the contact faces to overcome surface imperfections.
The disadvantage with mechanical clamps is that they must be fabricated very
accurately in order to match the members to which they are fixed. This requires
accurate surveying techniques, and it effectively precludes mechanical clamps from
being fixed to even very simple two-dimensional nodes. Figure 7 shows a typical
mechanical clamp.
Table 12 shows that despite the difficulties associatedwith fit-up, more repairs have
used mechanical clamps than any other type of clamp, although many of these have
involved fixing elements such as caissons or riser support brackets.
2. Grouted clamps have an annulus between the original member and the inner face of the
clamp. Once in position, the ends of the clamp are sealed and grout is introduced into
the annulus. The annulus can be sufficiently wide to accommodate buckles and dents
in the original member.
Load is transferred by the bond that develops between the cementitious grout and the
steel surface. Load transfercan be enhanced by weldingshear keys on the steel
surfaces. To date, only one repair has been recorded (in 1980) in which such keys
were added to the original member by hyperbaric welding.
Grouted clamps are suitable to repair complex nodes given the degree of survey error
that can be accommodated within the annulus. An important use of groutedclamps is
to join two members together as a trombone joint (see Figure 7).
3. Stressed grouted clamps are essentially a hybrid of the two clamping systems already
described. They have the same advantage of a loose tolerance fit as a grouted clamp,
but the shear transfer capacity is enhanced by increasing the normal force across the
grout to steel interface by the application of a prestress after the grout has gained
strength. This permits shorter connection lengths than grouted clamps carrying the
same load.
Stressed grouted clamps are generally used to repair or strengthen complexnodes or are
used where there are only short lengths of existing members to which they can be
fitted.
Steelwork or members adjacent to the area to be strengthened can avoid clamps being
used, because a certain length is necessary to transferthe load to the clamp. This problem
is sometimes solved by placing additional members around the affected area in order to
providean alternative load path (see Figure 7).
The design procedure for clamps must assess the effect the clamp has on the local
stiffness of the offshore structure, and it must include any additional wave or current
Repairsreview 29
loadings that the clamp itself attracts. Care must be taken in design to ensure that the
components of the clamp can be manoeuvred to the repair location, and to ensure that the
forces generated by the bolts will not crush or damage the original member.
There have been two reports of fit-up problems, both, surprisingly, with grouted clamps,
not mechanical clamps. In one case, photogrammetric surveys had been undertaken at a
time of poor visibility because of quantities of plankton in the sea. The repair members
were fabricated to inaccurate dimensions determined by this photogrammetric survey.
Another problem arose when a clamp was designed with only.a 25-mm annulus in order
to fit into a restricted space. An adjacent guide cone had been installed out of position, so
that the clamp could not be fitted. Photogrammetric surveyswere made to determine the
true position of the guide cone, and the repair members had to he re-fabricated.
63 Air welds
Table 12 indicates that above the water line most repairs are welded from access
scaffolding. In common with all welding to existing structures, it is sometimes necessary
to brace temporarily across the repair weld to prevent movement of the root gap while the
initial passes of the weld are being placed.
Working from scaffolding renders the work similar to normal topsides structural
modifications. In at least one case, an operator reported that access scaffolding for repair
work had been lost during storms.
Air welding is possible below the water line by introducing a coffer dam around the
damaged area with direct access in the dry from above water. The coffer dam may have
to be complex if it is placed around a sub-sea node or if a new memberhas to pass
through the coffer dam for welding to the original structure. The main advantage is that
the welding and testing personnel do not need to be experienced divers.
Many problems have occurred with the coffer dam seals against hydrostatic pressure. On
one repair, a weldingchamber was designed for work on both ends of a horizontal
member at a depth of 6m. The chamberwas designed to act either as a coffer dam with
access to the surface or as a hyperbaric chamber. The first node was repaired using the
chamber as a coffer dam, hut the seals becamedamaged. The second node was welded
using the chamber hyperbarically, and the work progressed more speedily. Because the
pressure within the hyperbaric chamber balances the external hydrostatic pressure, the
seals were not as critical to the operation as when the chamber was acting as a coffer
dam.
MTDPubLication94/102
STRESSED GROUTED CLAMP
CLAMPS USED TO
PROVIDE ALTERNATIVE
LOAD PATH
/ /fIj ANNULUS ACCEPTS
TOLERANCEGAPS
TROMBONE JOINT
ADJACENT STEEL
PREVENTS USE
OF NODECLAMP
WHICH
BE ATTACHED A'
Repairsreview 31
Tables 12 and 13 (page 27) confirm that coffer dams are not much used.
Hyperbaric chambers can be made of steel or a flexible material. The chamber, or habitat,
incorporates all the necessary welding and life support systems, and it is usually purpose
built for each repair location. Considerable engineering resources must be committed to
the design and construction of each habitat. The 1983 review1 recorded a reluctance by
some operators to consider hyperbaric welding, and it noted concerns on the quality of
welds produced under high ambient pressure. The technology of all the habitat systems is
now proven, and no misgivings were recorded during the 1992 survey.
The most recent major hyperbarically welded repair'° was to a weld joining a 1800-mm
diameter 45-mm thick brace into the leg bottle between pile sleeves at a depth of 182 m.
The weld had a through-thickness crack. The loads were considered to be too large for
clamps or cover plates, and the profiles were probably too complex. The repair was
successful, but the operator reported difficulties in finding qualified welder divers. Also,
the induction heating system formed water condensate at the weld location.
Wet welding offers the possibility of diverless repairs. It has been used extensively in the
Gulf of Mexico, where the steels in general use for offshore structures have a lower
carbon content than those used in the North Sea. With increasing depth, there is an
increase in the porosity and a consequent reduction in the strength of wet welds. This
reduced strength can be engineered out by extending the length of the weld (e.g. by using
castellated sleeves).
Four wet welded repairs have been reported. The first two connected anodes to non-
structural elements, and the third (in 1987) connected a bolted clamp to a conductor guide
frame. Only one defect was found in this repair in a total weld length of 69m.
The first, and to date only, repair to a structural member on the NWECS was completed in
1990 and is fully reported11 elsewhere. A vessel impact had fractured a weld joining a
diagonal brace to a leg at a depth of lim. Analyses showed that the structure was highly
redundant, and it could tolerate the loss not only of this diagonal brace but also of the one
adjacent to it.
The operator carried out extensive research into the background and practicalities of
connecting a new brace to the existing structure by a wet weld. Twelve welder divers
were employed, and three subsequent inspections have yet to discover a defect. Few
operators on the NWECS are yet considering wet weldingas a routine structural repair
technique.
The internal grouting of chord members can significantly reduce the stress concentration
factors at tubular joints. It can also be used to increase member strength and stability, and
to improve impact resistance. It is a relatively simple and inexpensive technique, and 11
repairs of this type have been recorded on the repairs database.
Most of the five bolted repairs recorded on the database have concerned non-structural
patch plates covering holes in existing members. On one important exception, the repair
plates were connected to existing 100-mm thick steel plate formed into complex profilesat
the 100-rn deep bottle leg'2. A hyperbaric welded repair had been rejected for several
reasons, including doubts on the metallurgical integrity of the welded zone.
The removal of members has been reported as an acceptable repair technique. In one
case, the gradual expansion of a pipeline was forcing a riser J-tube into the jacket. The
removal of members allowed the J-tube space to deform, and it relieved the load on the
jacket structure. The members which were removed were designed primarily for the pull-
in of the flow lines. Structurally acceptable alternative load paths are obviously a
prerequisite of this type of repair.
It had been anticipated that there would be an increasingneed for structures to be repaired
as they aged. Figure 8 suggests that damage is gradually identified over the early years of
the life of a structure, and that once this is repaired the structure operates with fewer
problems. There is no evidence from the database that older structures have an increasing
25
20
•a
U,
C)
15
cc
0
ci)
10
z
0
0 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25
Structure AgeAt Discovery (Years)
Figure 8 Numberof repairs versus structure age at discovery
Repairs review 33
requirement for repair, nor are certain types of damage (e.g. corrosion) more prevalent for
older structures. Table 9 (on page 21) shows an increasing incidence of problems arising
from corrosion, but older structures are not affected any more than any other structure age
group.)
This situation implies that the repairs perform satisfactorily, and the survey of the 61
repairs listed in the 1983 review confirms that none of these repairs has itself required
any further remedial attention. This statement excludes temporary repairs undertaken to
protect a damaged area over a winter season before the permanent repair is installed.
Some of the clamps identified in the earlier review have had their bolts replaced as a
precautionary measure. The original Macalloy bolts were considered vulnerable to stress
corrosion, and they were replaced with either Monel or L7 steel bolts. (Monel bolts were
used in the early 1980s, but L7 bolts were then used, because they are significantly
cheaper.) In some cases, short bolts have been replacedwith extension collars and long
bolts.
The only example in the database of a repair itself requiring subsequent remedial attention
is where vertical members sprang apart on being cut through in order to allow an
extension spool piece to be inserted. The initial work was undertaken to counteract the
subsidence of the Ekofisk field. A repair was required to cater for the eccentric loads
applied to the spool piece.
1. A vessel impact caused cracks and local damage to a 400-mm thick wall. Water was
able to pass through the cracked wall. Resin and caulkingwere used to seal the
damaged area externally. A 200-mm deep section of wall was removed internally and
recast, a coffer dam was placed against the outer wall, and the outer 200mm of wall
was then removed and recast. Resin injection points were cast into the new concrete to
ensure a good bond between the old and new material. The operator reported
difficulties in sealing the coffer dam.
2. A leak across a construction joint at a depth of 88m manifested itself some years after
construction, and it was finally repaired 10 years after construction. Initial repair
attempts were to inject resin into the joint, but they were only temporarily successful.
Subsequent optic fibre examinations showed voids within the 1200-mm wide wall. The
voids were injected with cementitious grout, epoxy grout being used for the final
contact with the old concrete. The work was extremely difficult to carry out, because it
involved working against the full hydrostatic head.
4. Damage to a 500-mm thick cell roof slab at a depth of 80m was caused by a dropped
object. A 300-mm deep hole was formed in the concrete, and water flowed through the
slab. A permanent repair was made by placing prepacked aggregate within the hole,
covering the hole with a steel plate, and injecting grout to restore the original concrete
profile.
5. Cracks in external shear walls (with hydrostatic pressure on both sides) were injected
with resin. In some cases, the crack was jacked apart before the resin was placed, so
that, on removal of the jacks, the resin was compressed into the crack voids. Ballast
was added to the structure to prevent the cracks reopening under wave loading.
6. A scour hole extending 4m beneath the structure was formed at a pipelineentry point.
The void was grouted and subsequently protected by rock dump.
Thus, in general, concrete structures offshore are repaired in the same way as concrete
structures onshore, by replacing damaged concrete with sound material, and by injecting
cracks with resin. Epoxy or resin compounds are often used to ensure that there is a good
bond between the fresh concrete and the original material. The hydrostatic head across the
concrete elements can be very difficultto deal with, and valved systems are used when
drilling injection holes into areas which might connect to the sea.
Details of 16 repairs outside the NWECS are recorded on the structural repairs database,
but not included in this report (see Foreword). In only two of these cases were the
questionnaires completed by the operators. Information on the remaining repairs was
gathered from BOMEL's records.
Oil and gas companies operate on a world-wide basis, and technical information gathered
by a central location in Europe or the US is disseminated to all their operating arms.
Company personnel working offshore Africa or Asia often have North Sea or other
relevant experience, and are therefore aware of the repair techniques used there. The
design contractors specialising in repairs also operate on a world-wide basis, and their
personnel have used their knowledge when working locally outside of Europe and the
USA.
Given this transfer of information, the repair technology is the same in all regions of the
world in which the multi-national oil and gas companies operate. In other areas of the
world, such as Russia or China, differentrepair techniques may have evolved, but the
project resources did not allow this aspect to be examined during this study.
The causes of damage differ widely, and 10 cases are described below to illustrate a cross
section. In one case, unique techniques were used to increase the capacity of piles while
they continued to support a working facility. This is also described.
Repairs review 35
1. A reappraisal of the strength of the foundation system to four structures located in
70m of water in the Bass Strait, Australia, led to the provision of sub-sea templates
from which raking latticed struts rose at 45° to mate with the deck levels of the
original structures. These struts and template structures were designed to resist the
horizontal forces applied to the platform.
Each of the four structures had eight legs and eight piles of some Im in diameter
penetrating the sea bed by 30m. The subsea templates, designed to strengthen the
original structures, each had eight piles of 2.lm diameter penetrating the sea bed by
more than 30m. In one case, two smaller templates and raking struts were used so as
to sit either side of a flow line which would have been directlybeneath a single
template. The template and raking struts were installed in 1986 and 1987.
2. Two-stage insert piles with an overall length of 150m were designed for a platform
off the west coast of Australia in lOOm of water, but far less skin friction was
measured during their installation than had been assumed in the design. A concrete
plug was designed to increase the end-bearing capacity of each pile and to transferthe
load to the inside face of the lower (insert) pile. However, there was still insufficient
skin friction for the uplift design case, and the strengthening system was extended to
include an under-ream below the toe of the insert pile.
Strengthening options which were rejected include improving the pile toe soil
properties by permanent soil freezing and by treating the soil with epoxy. The
solution chosen was to use epoxy to stabilise the soil around some of the piles during
the under-reaming operation.
The concrete plug was reinforced with a cage of reinforcing bars, and it contained
cement replacements to reduce the heat of hydration. Full-scale laboratory tests were
undertaken to confirm the capacity of the plug and its ability to transfer load to the
pile. The under-reaming was undertaken from the live platform in 1985.
3. Fatigue cracks were discovered 10 years after the installation of a jacket structure
offshore of the west coast of the US in 60m of water. The jacket had been fabricated
in Japan, and it had suffered fatigue damage during its journey across the Pacific.
Two additional members attached by stressed grouted clamps are one repair option
currently being considered by the operator.
4. A crack extendingfor half the circumference of a leg was found near a mudline node
of a jacket offshore of the west coast of the US in 17m of water. Seismic loads were
thought to have caused the damage. The repair was accomplished by grouting the
pile within the leg, and by providing a stressed grouted clamp to the node.
5. The conductorguide frames of nine wellhead platforms offshore south east Asia were
damaged by a typhoon in 1989. The typhoon created 12-rn waves, but the platforms
had only been designed for 4-rn wave heights. The structural members are thin
walled with a maximum thickness of 9mm, and some stress concentration factors
exceeded 20. Clamps were used to repair the damage.
36 MTDPublication 94/102
6. A series of eight stubs supporting an impressed current riser corroded through. The
stubs ran the length of the structure in 70m of water in the Bass Strait, Australia.
The corrosion mechanism was not determined. The stubs were replacedwith
mechanical clamps in 1982.
7. K-nodes on a structure in 70m of water in the Bass Strait, Australia, were detailed
with gussets. These gave rise to high stress concentrations, which, in 1986, were
relieved by filling the through members with grout.
9. Piling operations caused anodes to fall off and cracks to form in the parent metal
adjacent to the anode stubs on a jacket in 85m of water offshore west Africa. The
anode stubs had been welded on locally. 100-mm diameter holes were found in
primary and secondary members at four levels throughout the depth of the structures.
Repairs in 1988 included groutedclamps surroundingthe holes and patch plates.
10. Lang et a1.9 report how the loss of a conductor guide frame at a depth of 6m in 1975
led to a study of 28 platforms in the Gulf of Mexico. Of these platforms, 28 had
fatigue problems resulting from vertical wave loading. Frames as deep as 69m were
affected. Four of the subsequent repairs are included in the repairs database. The
only unusual repair system was to tie the conductors directly to the structural frame,
but these repairs failed quickly, because the restraint on relative movement between
the conductors and guide frame produced higher forces than expected.
During the 1983 review', there was a reluctance by operators to discuss the cost of
repairs. For those costs which were reported, there were large differences between what
appeared to be similar repairs. It was assumed that market conditions accounted for these
differences, and the questionnaires for the 1992 survey therefore concentrated on
timescales rather than costs. There was still only a limited response (see Table 15), and
there was no response at all on the time taken to fabricate the repair assembliesused on
the NWECS.
Such costs as were reported are also included in Table 15. In some cases, precise total
costs were hard to define because divers and vessels employed for other work were
diverted temporarily to deal with repair tasks.
Repair options are not always determined by costs alone. A recent commission
undertaken by BOMEL concerneda repair which could not be installed from the jacket
Repairsreview 37
(which was the cheapestoption), simply because there were no bed spaces for the repair
crew. However, the operator had a diving support vessel on charter, and the repair was
thus designed around the capabilities of that vessel. This type of situation makes it very
difficult to comparecosts of apparently similar repairs.
Figure 9 shows that the total offshore time tends to be higher for repairs carried out at or
near the water line. This may be because these repairs are more susceptible to poor
weather, or because the cost of deep sub-sea work imposes a tighter discipline. Lang et
al. providea useful background of costs and time for a repairs contract which included
the removal of a conductor guide frame, the internal grouting of tubulars, and the
installation of stressed grouted clamps in the Gulf of Mexico at depths of 37 and 68m.
Details are repeated in Tables 16a and 16b.
BOMEL produced two sets of 1992 cost data for structural repairs offshore Thailand
which demonstrate the balance between initial capital and whole life costs. The lower
capital cost was for a certain minimum number of repair clamps and a relatively high level
of inspection over the 20-year life of the jacket, compared to the costs for a greater
numberof clamps and a relatively low level of inspection. The amounts are given in
Table 17, in which no allowance has been made for the cost of the money.
90
• NWECSREPAIRS
80 — • 0 REPAIRSOUTSIDEEUROPE
70_
60_
TOTAL OFFSHORE
TIME IN DAYS 50 —
0
40 —
30_
2O_
0
10 —
0_ •
I I I I I I I I I I
O 20 40 60 80 100 120 140 160 180 200
Figure 9 Comparison ofrepair water depth with total offshore repair time
38 MTDPublication p4/102
Time, days Costs, £k
Year of repair Cause of damage Damaged members Water Repair technique Design Fabrication Offshore Offshore Design Fabrication Offshore Total
depth,m water total installation
Table 17 Costs associated with the level of repair workand the extent of the subsequent
inspection
11.1 General
In all, 18 repairs are recorded on the pipeline repairs database, which covers 68% of the
total length of pipeline laid on the NWECS (see Section 2.5). Only those repairs requiring
replacement of sections of the line or the use of a clamping system are recorded. Pipeline
upheavals and spanning are dealt with as routine maintenance items, and they are not
included.
The information gathered from operators in the Norwegian sector compares well with the
Norwegian Petroleum Directorate's CODAM database. A 1990 study'3 on pipeline loss
of containment recorded, hut did not list, 63 incidents which resulted in a loss of
containment. Assuming these all required repair, this indicates a higher incident per km of
pipeline than discovered during the current study. Operators in the UK sector of the
NWECS are obliged to provide full details of all pipeline repairs to the UK Health and
Safety Executive, hut the information is not publicly available, and it could not be used to
enhance the pipeline database in this review.
Elevenof the repairs were to pipelines of less than 400mm diameter, and the wall
thickness of repaired lines varied between 9 and 19mm.
1. Although only 27% of the pipelines laid on the NWECS are less than 400mm in
diameter, at least 69% of the reported repairs fell within this limit. The proportion
remains at 69%, even if risers are excluded from the repairs database.
2. 74% of pipelines on the NWECS carry gas, but for the 12 repairs for which the
product was reported, 58% carried oil, 25% carried water and 17% carried gas. It is
believed that oil is also carried in three of the pipelines for which no productwas
reported. This finding corresponds to investigations into pipeline failures in the Gulf
of Mexico'4.
11.2 The discoveryof damage
Six of the 18 reported repairs occurred during installation, and at least three of these were
known about and repaired immediately. Seven of the 18 were detected in routine
examination (two internal and three external), three were discovered by chance, and one
was discovered by detecting a loss of pressure and a loss of the delivered product. There
was one strengthening required as a result of a pipeline upgrade, and two repairs were of a
preventative nature.
Two operators reported difficulties in using inspection pigs following the discovery of
damage. One stuck in the first bend after launch, and it had to be cut out of the pipeline
on the deck of the offshore installation. The quality of lighting on a video pig was
reported to be too poor to he able to see any defects.
Two of the chance discoveries were by operational vessels spotting an oil slick or bubbles
on the surface, and the third chance discovery was of the product overflowing from a J-
tube on a platform. It was reported that leak detection systems are only approximate
devices, and that they do not detect small pipeline leaks.
Followingdiscovery of damage, three pipelines were shut down. In one case, temporary
use was made of a test line, but production through the other two lines was stopped for 3
months and for 2 months, respectively, until the repairs were completed. The operating
pressure in two lines was reduced following discovery of damage, by 27% in one case and
by 50% in the other.
Repairsreview 41
113 Installation damage
Two cases of damage were reported to have been caused by trenchers being dropped onto
the line. These incidents occurred in 1976 and 1978, at a time when structuraldamage
was also being caused by dropped objects during installation (see Table 9, page 21).
In one case, a leak developed at a flange, possibly as a result of excessive bolt torques
being applied. Another pipeline was laid to an incorrect radius, probably as a result of
being tensioned too hard during laying. The as-laid position would have clashed with the
anchor spread of a vessel working over an adjacent sub-sea manifold, and 400m of pipe
was hauled to the correct position after an initial cut had been made. Buckles were then
found at the pipe ends when the 100-m long insert piece was offered up. The operator
reported that there are no commonly available analytical techniques to determine the
forces to he applied to move the pipe on the sea bed, and it is possible that the forces
were too large and caused the pipe to buckle. The 100-rn insert was too long a unit to be
handled successfully, and eventually two 50-rn lengths were installed. The inserts were
hyperbarically welded.
Two trawl boards hit pipelines during their installation. In one case, the contractor pulled
the line hack, repaired the damage, recoated the pipe, and relaid it. Subsequently, two
buckles were found, both of which required 72-rn sections of line to be inserted, connected
by hyperbaric welds. In the second case of trawl damage, the pipeline was bent through
30°. This was also repaired with a hyperbarically-welded insert piece.
All three cases of external corrosion relate to risers. In two cases, the repair was
precautionary. In one of these cases, a coffer dam was used to repair corrosion damage in
the splash zone, hut details of the repair technique are not known. The second case is
more interesting: it was reported that monotonic hydrogen from the oil product diffused
through the 510-mm diameter steel riser and became trapped between the steel and the
Monel sheathing to the riser. This was reported to have led to the rupture of the Monel
sheathing.
The third case of external corrosion was discovered when oil was seen overflowing from a
J-tuhe containing a riser. The annulus between the J-tuhe and the riser should have been
filled with inhibited water, but instead it was filled with sea water. A pin hole leak was
found in the riser. A sleeve was placed over the damaged section.
Four cases of internal corrosion were reported. The first led to the replacement of the
813-mm diameter, 19-mm wall thickness Forties pipeline which is documented'
elsewhere. Intelligent pigs had discovered a 2-mm wall loss at some locations, increasing
to 4mm at welds. The design pressure of 142 bar was reduced to 115 bar in 1987 after
the corrosionwas first discovered. Further pressure de-rating would have made the line
uneconomic.
The 169-km line was replaced in its entirety with a 914-mm diameter line, primarily
because the shutdown period for this option was 14 days less than the best of the
alternatives.
The second case concerned two infield 150-mm diameter flexible flowlines, which were
each initially used to carry productbut subsequently carried water for injection. Internal
calliper runs showed significant corrosion, and in 1991 both lines were replaced with
polyethylene lined pipe.
In the third case, a 254-mm diameter, 11-mm wall thickness concrete-coated pipeline with
a 5 to 10 bar operating pressure was used both to carry sea water for injection and for
discharging oily waste. It was cleanedevery 3 weeks with a scraper pig. In 1990, a crew
member of a supply boat saw bubbles on a dead calm sea. Divers detected the bubbles
but found no damage and could not find the source. Video pigs showed nothing
internally. The line carried gas pressurised to 50 bar which blew off the concrete coating.
Two distinct pits, like bullet holes, were found at the lowest part of the pipeline (the six
o'clock position). One was through thickness, the other was 5 to 6mm deep.Other pits
were found nearby. The cause was reported to be sulphate-reducing bacteria, which
developed as the dual use of the pipe brought fresh sea water and oily waste into contact.
Despitebrush and foam cleansing before the video pig was deployed, a sludge remained at
the pipeline low point, obscuring the defects. A proprietary clamp was used to repair the
line which is now hydro-tested at 40 bar every 3 weeks.
The fourth case of internal corrosion occurred in a 168-mm diameter, 9-mm wall thickness
infield line carrying filtered and treated sea water at 143 bar. The volumes of product
measured on despatch were not received. A 250-mm long rupture was found in the pipe
which was lying in a 7-rn wide, 2-rn deep crater in the sea bed. Pipeline samples showed
that only 2mm of wall thickness remained. The reduction was possibly caused by
bacterial corrosion. The rupture was thought to have been caused by operational surges.
A proprietary clamp was used to repair the pipe, and the operating pressure was halved to
71 bar.
It may be relevant that in three of the four cases of internal corrosion there had either
been a change of use, or the line carried different products at different times, also that in
three of the four cases sea water is carried in the line.
Apart from the cases of damage caused during installation (see Section 11.3), only one
case of trawl damage was reported. This line is coated in concrete, and it is covered in
rock dump except at the damage location where, because of the proximity of valves, it is
protected by an epoxy coat alone. A flexible line now by-passes the damaged area.
Repairsreview 43
11.7 Repair systems
Five hyperbaric repairs were reported, four with insert pieces and one with a patch plate.
Proprietary mechanical clamps were used for three repairs, and mechanical clamps were
purpose designed for two repairs. Four pipelines were replaced, and one damagedarea
was by-passed. A sleeve was placed around one riser. The repair systems used in two
cases were not recorded.
There are insufficient data to attempt to find any links between the type of repair and
water depth or operating pressure.
A comparisonbetween Tables 15 and 18 indicates that the time within which pipelines are
repaired following the discovery of damage is of the same order as the time taken to
repair offshore structures.
As with all offshore costings, there are differences in the stated costs for apparently
similar repairs. The costs of pipeline and structural repairs compare only at an order of
cost level. For example, the 1986 repair of corrosion damage to a riser was said to have
cost £500k, and in 1988 the cost of replacing a caisson was said to be £250k.
An interesting aspect of one riser repair is that two alternative repair systems were run in
parallel. This resulted from doubts on certain installation aspects of the first repair
method. The alternative was set up to reduce any possibility of further extendingthe
production downtime. As much money was spent on the second scheme by the time it
was abandoned as was spent in total on the preferred scheme.
No total costs were obtained for hyperbaric repairs to pipelines nor for repairs with
proprietary mechanical clamps, although one clamp was said to have cost as little as £2k.
44 MTD Publication94/102
Cost, £k
Year of repair Repair technique Time, days Design Material and Installation Total
fabrication
(alternative replacement riser scheme run in parallel - costs recorded to time scheme was abandoned) 80 200J
12.1 General
The questionnaires asked operators to state what lessons they had learnt from their repair
work, and these are grouped below. Some of the opinions have been reported earlier in
the text, but they are repeated here for completeness. These items are not intended to be
considered as study conclusions, only to represent the perceptions of practitioners in the
field of offshore repairs.
Another operator reported difficulty in dealing with the interaction of the tolerances in
sub-sea survey readings, and was concerned that the readings could never be confirmed.
One operator pointed out that the original structure must be re-analysed to incorporate the
local stiffening introduced by each repair system. Another spoke of the need to reduce
overhead welding in wet weld repairs.
A lesson reported by two operators was that, on groundsof safety, sub-sea work should
only be undertaken if essential.
The problem of sealing coffer dams was reported as a lesson learnt, and the usefulness of
onshore fit-up trials and preparation time was emphasised. One operator reported finding
underwater fit-up trials in a dock very worthwhile.
Coarse threaded bolts were reported to be easier to use under water sub-sea than fine
threaded bolts.
MTDPublication94/102
The importance of keeping welder diver teams together was mentioned, and another lesson
reported was that heat removal was no longer considered to be a problem in habitat work.
Several operators considered that insufficient attention was being paid to secondary
structural elements, and that some details were left to the fabricator to draw up without
proper checks on the design.
12.7 Fabrication of the original structure
It was stated that older jackets currently see far more inspection sub-sea than they ever did
in the fabricator's yard. Two operators called for improved standards of fabrication, and
one considered there was a need for a fuller interaction between designers and fabricators.
Seven operators were unhappy about the quality of the as-built information supplied to
them.
Three operators mentioned the inadequacies of dynamic positioning controls, and one
reported that a dynamic positioning failure might be expected once every 2 years of a
vessel's life. This does not necessarily conflict with Table 9 (see page 21), which gives
the frequency of repairs resulting from work vessels impacts compared to the total life
exposure of structures on the NWECS. Work vessels are at close quarters to structures for
relatively short periods only, not for the full life exposure period.
One operator considered it wrong for the operator to have direct control over all the sub-
contracts of repair work (design, fabrication, material procurement, etc.), because this did
not provideany incentive for sub-contractors to perform well.
One operator spoke of the need to understand the limitations of pigs, another of the need
to introduce automaticcontrols to prevent pressure surges. The basic design of one
proprietary clamping system was criticised, because it was difficult to positiona torque
wrench for some of the bolts.
Repairs review 47
13. CONCLUSIONS - STRUCTURAL REPAIRS
The conclusions have been drawn from the analyses presented in this review. Where
appropriate, the Section containing the arguments which support each conclusion is listed,
and appropriate Tables or Figures are noted.
(1) 172 recorded structural repairs on the NWECS is a sufficiently large number to
have confidence that any trends which are found from the reported data are
significant.
(2) During the period considered, the pro rata repair rate was found to be similar for
steel and concrete structures, despite the widespread perception that concrete is a
more durable material than steel. Some concrete problems take time to manifest
themselves. (Section 3.3 and Figure 4, page 17)
(3) The need to repair appurtenances is increasing, while the need to repair primary
structural members is decreasing. (Section 3.4 and Table 4, page 17)
(4) Consistently over time, most structural damage is discovered by inspection or as the
damage occurs. Apart from a single case of damage found following a winter
installation break, the last chance discovery of structural damage was over 12 years
ago. (Section 4 and Table 5, page 19)
(5) Fatigue damage sufficient to require a repair (other than grinding) has not been
discovered to date on any steel structure installed on the NWECS later than 1983.
However, fatigue damage on earlier structures continues to require repair.
(Section 5.2, Table 9, page 21, Figures 5 and 6, page 22)
(6) All structures have a range of fatigue lives over their weldedjoints. Therefore, it is
likely that where joints with fatigue lives lower than the operational life (which may
be greater than original design life) have not been strengthened, it is likely that
further fatigue problems will manifestthemselves with time. Despite computational
advances, the approximations necessary in the design process may result in newer
structures still experiencing fatigue problems of some kind while operational.
(Section 5.2, page 21)
(7) Based on all vessel impacts which have led to damage requiringa repair and
assuminga 25-yearstructure design life, one in every five structures on the
NWECS can he expected to have an impact which leads to a repair. Alternatively,
this can he translated as three structures per year incurring vessel impact damage
which result in a repair. These average values are reducing with time.
(Section 5.3, Tables 10 and 11, page 23)
(8) Damage caused by errant vessel impacts can be severe, and such impacts have the
potential to cause the catastrophic failure of offshore installations. (Section 5.3 and
Table 10, page 23)
(11) Corrosion damage has increased over the 1980s, and a particularproblem has been
identified of bi-metallic corrosion occurring between new pumps manufactured in
non-ferrous metals and the steel caissons within which they are placed.
(Section 5.7, page 25)
(12) Repair techniques reported in the 1983 review remain in use, with detailed changes
to ensure more efficient and reliable installations. Apart from wet welding, no new
repair techniques have emerged over the last decade. (Section 6, page 26)
(13) Clamp systems are used for approximately 50% of all sub-sea repairs, and welded
systems for about 35% of all sub-sea repairs. Above the water line, welded repairs
account for 66%, and clamped repairs for 30% of the total. (Section 6.1, page 26)
(14) For fatigue damage, the internal grouting of members and the use of stressed
grouted clamps have been reported as being the most successful repair techniques.
(Section 6.1, page 26)
(15) The difficulties of sealing coffer dams against hydrostatic pressure have not reduced
with time. Hyperbaric welding technologies have improved over the last decade,
and the only problems reported concern condensate water associatedwith induction
heating. (Section 6.3, page 30)
(16) Wet welding has been used as a- repair medium on the NWECS since 1986, and in
1990 it was used for the first time in repairinga primary member.
(Section 6.4, page 32)
(17) The inspection and re-analyses of structures have gradually identified their design,
fabrication and installation faults. Once these are repaired, the structures have a
declining need for remedial work. Although some North Sea installations are nearly
30 years old, the greatest age reported for a structure requiring repair is 18 years.
(Section 7 and Figure 8, page 33)
(18) Repairsmade necessary by the effects of corrosion have not been found to be more
prevalent on older structures. (Section 7, page 33)
(19) All repairs reported in the 1983 review have performed satisfactorily, apart from
some bolts having been replaced. Specifications for bolts (includingcorrosion
protection) have been developed to ensure long-term performance. Individual
repairs have successfully withstood up to 22 years service. (Section 7, page 33)
49
Repairsreview
(20) Concretestructures offshore are essentially repaired in the same manner as concrete
structures onshore, although with the significant difference that the hydrostatic
pressure acting across structural members can lead to complex repair operations.
(Section 8, page 34)
(21) Structures operated by international oil and gas companies make use of the same
repair techniques throughout the world. (Section 9, page 35)
(22) Some repairs outside the NWECS have been caused by. local conditions or by local
practices (e.g. typhoons or using explosives adjacent to structures as a fishing
method. (Section 9, page 35)
(23) Structural repairs at or near the water line can take three or four times longer to
complete offshore than repairs at depth. (Section 10, page 37)
The conclusions below have been drawn from the analyses presented in Section 11 (page
40).
(1) A disproportionately large number of pipelines with diameters less than 400mm
require repair.
(3) Internal corrosion damage to pipelines tends to occur when two products are carried
consecutively in the same line, or when there is a change in the productover time.
Internal corrosion is also a feature of lines carrying sea water.
15. RECOMMENDATIONS
(1) Errant vessels present a serious hazard to offshore installations, and improved
methods to minimise the risk of impact should be investigated, including
determining the capabilities of standby ships to detect and warn off such vessels.
(2) The damage caused by work vessels in proportion to their time on location should
be investigated to determine whether they pose a particular threat to fixed offshore
installations.
So MTDPublication94/102
(3) Attention given to the design of secondary members and other attachments should
be as great as that given to primary members. This recommendation was also made
in the 1983 review.
(4) Accurate as-built drawings are essential to producestructural repair systems, and
procedures should be introduced to ensure that such drawings are not neglected as a
result of project pressures. Steel material properties should be measured and
recorded during fabrication, so as to allow detailed fracture mechanics calculations
of the original structure to be undertaken in order to assess the as-damaged
structural condition.
(5) Jacket designers should recognise the likely costs of repair over the life of a
structure and consider the potential benefits of designing damage-tolerant structures
which incorporate alternative load paths, thus minimising the extent of repair works.
(6) Close links should be established between designers and fabricators to ensure that
all details, from primary welds through to minor attachments, are designed so as to
be easily fabricated, also that no fabrication practices are undertaken without
understanding their effect on the long-term structural performance.
(7) The propensity to internalcorrosion of pipelines which carry sea water and of
pipelineswhich carry a changed or changing product should be investigated to
determine pipelines which are at risk from sudden failure.
(9) Container vessels with walls that also act as structural members should incorporate
automatic pressure release systems if an operational malfunction could cause
collapse or bursting.
REFERENCES
3. SMITh, D. J.
Project management of subsidence and Ekofisk jacking project
Offshore Technology Conference (Houston), 1988, Paper OTC 5655
4. SHARP, J. V.
Ageing offshore structures - a review of recent UK research
Offshore Mechanics and Arctic Engineering Conference (Calgary), 1992
Repairsreview 51
5. LAMB, H.
Historic storms of the North Sea, British Isles and Northwest Europe
Cambridge University Press (Cambridge), 1991
6. TEBBEYF, I. E.
Damage and repair trends in fixed steel offshore structures
Paper presented at the International Conference on Weld Failures(London), 1988
7. SVEEN, D.
Oseberg B jacket - damage assessment and repair after submarine collision
Society of Petroleum Engineers Offshore Europe Conference (Aberdeen), 1989,
Paper SPE 19269
8. HORDYK, M.
Offshore structures : design/fabrication interface
UEG (Underwater Engineering Group) Report UR37, 1988
10. COTTRILL, A.
Comex puts the heat on BP Magnus repair
Offshore Engineer (London), October 1990
13. Pipeline and riser loss of containment study - 1990, prepared for the United
Kingdom Offshore Operators Association Ltd (UKOOA)
0TH Report No. 91 337, Health and Safety Executive Books (London), 1991
14. MANDKE, J. S.
Evaluation of offshore pipeline failure data for Gulf of Mexico
Paper presented at the Offshore Mechanics and Arctic Engineering Conference
(Houston), 1990
52 MTDPublication 94/102
APPENDIX A
QUESTIONNAIRE USED FOR STRUCTURALREPAIRS
llitSSa Description of what was done, hnw and why(including sketches if appropriate), and also details of
any design solutions considered and rejected(and why).
Current apemlor:
Slrrrclurv name/locasion:
Date installed:
ine
____ Which factors ted to the nacceas of the
How adequate was the offshore survey data (globally and operation (e.g. planning, personnel, weather,
Previous operators with dales: soyport vessels crc)?
locally)?
Howwas it discovered? Whatwere the elapsed time and manhoursfor the design wnrk
and its compeneers parts?
When was it diseovneed? What has hero the inputby, or the response
of, the Certifying Authority?
Pabricalion
Did the fahricator have sufficient time to complete his contoact? stage of the repair/strengthening project.
Was lime u critical element of the fabricators' oantract?
Ut
APPENDIX B
QUESTIONNAIREUSED FOR PIPELINE REPAIRS
Ut
a 2
REPAIRS TO flSEA PIPElINES
REPAIRS TO SUB-SEAPIPELINES
OPERATORS REPORT- BASIC DATA
OPERATORS REPORT - TIlE REPAIR
if
Descriptionof what was done, how and why (inctnding nkenchm npprnprlale), and also details of
Currentoperator any design sotntiom considered and rejected (and why)?
Water depth at repair pesition: Was the design carried out in-home or by Which factors introdneed prohtenm?
(specialist) cousattant?
Nature of seabedat repair position: Whichsnppert vessel wm usedand whatwan
How adequate were the uoailable design the period of charter?
Product carried by pipeline (temperature, gnidrokrersonnrt eupertise?
pressure, phase,Intl bore, ete): Whatwas the elupned time and manhours for
What input wuu there from the Certifyiag the inslalutionactivities?
Continnousty operated: Authority?
Snhaeqnentperformance
What support vesset capabilities were
Natore of Damanr (respond by sketch if nignilrcant to the design, eg. dynamic What type of inspection (and nI what
more appropriate) positioning, sea-keeping, cranage capacity, frequency) does the repair have?
diver requirements, etc.?
Whatwas the damage?
Were the vewel facilities and inolallutiou
contractors' needs properly consideredduring Has anything been found dnring these
droign.? inspections?
Howwas it caused?
What was the elapsed time and manhoursfor
the design work and its component parts?