1

23-rd World Gas Congress, Amsterdam

ESTIMATION OF RESIDUAL PIPELINE LIFE

BASED ON IN-LINE INSPECTION DATA

BEKKER M.V. UKRTRANSGAS Affiliated company of NAFTOGAZ OF UKRAINE National JSC

BILYK S.F., GAZEXPORT, Ltd., Representative Office in Ukraine
ROZGONIUK V.V., GAZEXPORT, Ltd., Representative Office in Ukraine
ORYNIAK I.V., Institute for Strength Problems, Nat.Acad.Sci. of Ukraine

1
 2

Abstract

The paper provides the results of in-line inspection for complex evaluation of technical
conditions of a gas pipelines. The evaluation is based on the intensity of corrosion processes at
separate pipeline sections, defined with two inspections separated with a five-year period. Analysis of
the results of different inspections discovered the problem of defects identification. The obtained
results emphasise the problem of accuracy of defects recording and possibility of their identification
according to several in-line inspections separated in time.

2
 3

TABLE OF CONTENTS

1. Method of in-line inspection to evaluate technical conditions of a gas pipeline

2. Analysis of the results
3. Complex evaluation of technical conditions of a gas pipeline
4. Evaluation of residual life of a pipeline
5. Probability of a pipeline failure

3
 4

1. METHOD OF IN-LINE INSPECTION TO EVALUATE TECHNICAL CONDITIONS

OF A GAS PIPELINE
Transit pipelines belong to highly hazardous facilities. To provide the sufficient level of their
safety, Gas Pipeline Operators have to fulfil a number of procedures and processes, such as in-line
inspection. The data obtained according to this method together with the data of other types of
inspection base the ground for the Operator’s programmes of pipeline rehabilitation. The paper
attempts to compare the results of two in-line inspections divided with a five-year period.
There are different methods of pipeline diagnostics: in-line inspection (II) with intelligent pigs
(IP) , intensive measurement (IM) inspection, prospecting digging methods, etc. For more reliable
estimation of the pipeline condition, the Gas Pipeline Operator needs results of surveys carried out by
different methods. II techniques stand out from other diagnosis methods by their swiftness and the
scope of information they provide. That’s why the subject of this report is an attempt of more detailed
assessment of both diagnostic data and pipe condition using results of the said techniques.
This work is done on the basis of II data obtained at Dolyna-Rososh 100 km section of Dolyna-
Uzhgorod-State Border-2 (DUSB-2) pipeline. Data used for analysis involve results of two
examinations of the said pipeline section: first survey carried out in 1997 and resurvey made in 2002
with IPs produced by Rosen Technologies.
Inspection gives us information about defects dimensions: their depth (calculated from the
measured residual pipe wall thickness), length and width values. Inspection also gives us some
general defect characteristics, such as its nature (mechanical damage, corrosive metal loss, etc.), and
the pipe cross-section angle of defect location.
In order to estimate changes in pipeline condition within five years between two pipeline
surveys in 1997 and in 2002, we’ll concentrate only on dimensions of pipeline defects.

2. ANALYSIS OF THE RESULTS

To analyze the results of two pipeline diagnostics, let’s present them graphically. Figure 1
shows dependence of number of defects from pipeline length.
140

119
120

100
Number of defects

80

60
49
44
40
27
23
19
20 16
11 9 10
8
5 5 3 5
1 2 0 1 0
0
10 20 30 40 50 60 70 80 90 100
Pipeline length, km

Inspections of 2002 Inspections of 1997

Figure 1: Defects according to in-line inspections in 1997 and 2002

4
 5

To conclude, the number of pipeline defects increased to 306 in 2002 against 52 defects
reported in 1997. It should be stressed out that 52 defects were recorded in 1997, imply, so-called, a
post-overhaul sampling; i.e., these defects are those remained after the pipe repair and renovation
works according to results of the pipe diagnostics carried out in 1997. It is quite obvious that a post-
overhaul sampling of defects is a reasonable basis for both qualitative and quantitative estimations of
changes during a certain operation period, which makes five years in this particular case. Apart from
increase in the number of defects, many times increase of the number of defects is observed only
within three sections (20-30 km, 80-90 km and 90-100 km) of the gas pipeline.
As far as the recorded defects are mostly of corrosive nature, it seems reasonable to express
data shown in Figure 1 through pipeline corrosion intensity. We imply that corrosion intensity means
the weight of metal lost by the pipeline due to corrosive damages. In order to estimate these weight
values, let’s represent each defect as a half of an oblate ellipsoid of rotation with axes corresponding
to the defect depth, length and width, respectively.
Figure 2 shows us that corrosion intensity is maximum for the same pipeline sections.

1400

1211
1200
1085

1000
Metal loss, gram

800
690

600
502
434
400

217
200
124 106 120
103 83
51 52
27 8
24 9
4 0 0
0
10 20 30 40 50 60 70 80 90 100
Pipeline length, km

Inspections of 2002 Inspections of 1997

Figure 2: Corrosion intensity over the pipeline length, grams

However, a certain peculiarity is also observed here: notwithstanding the minimum number of
defects within the 90-100 km section of the gas pipeline (see Fig.1), corrosion intensity appeared to
reach its maximum value within the same section (see Fig.2). This fact on its own is important for the
Gas Pipeline Operator with respect to the plan of pipeline repair and renovation works, mainly, for
purposes of additional examination and study of causes inducing origin and development of the
corrosion processes. On the other hand, said fact conveys the idea of a more generalised criterion of

5
 6

intensity of corrosion processes. There’s no doubt that the relative corrosion intensity may serve as
such criterion. The relative corrosion intensity shall be defined as the ratio of the weight of metal lost
within a special pipeline section to the weight of metal lost within the full length of the inspected
pipeline. Figure 3 shows dependence of the relative corrosion intensity versus the pipeline length.

0,60

0,52

0,50
Relative corrosion intensity

0,40

0,30
0,30
0,27

0,20
0,17
0,14
0,12
0,12 0,13
0,10
0,06 0,05
0,01 0,01
0,01
0,03 0,01
0,02 0,01
0,00 0,00 0,00 0,00
10 20 30 40 50 60 70 80 90 100
Pipeline distance, km

Inspections of 2002 Inspections of 1997

Figure 3: Relative corrosion intensity over the total of pipeline length, grams

The figure allows to distinguish two pipeline sections of 0-10 km and 30-40 km. As we see,
none of other previously mentioned pipeline sections attracts the attention of the Gas Pipeline
Operator by the relative corrosion intensity. The two distinguished sections of the pipeline are
characterised by the multiple decrease of relative pipeline corrosion intensity within five years. Such a
conclusion seems to be absurd at the first glance. In fact, the explanation may be quite simple:
decrease of the relative corrosion intensity is observed for some of the pipeline sections due to
substantial growth of the total level of corrosion metal losses of the pipeline.
Data provided by Figures 1-3 allow the Gas Pipeline Operator to isolate pipeline sections
where such indicators as density of defects, corrosion intensity and relative corrosion intensity have
maximum values. It’s quite obvious that similar graphical analysis may be extended to singling out
sections with the highest values of relative density of defects per length unit (meter, ten meters, etc.).
The gas pipeline sections distinguished in such a manner will become subject to more detailed study,
including other diagnostic methods, in order to estimate accuracy of the defect dimension
measurements with in-line inspection and possible causes of the enhanced defect development.
Such practical approach to the II data currently used together with other diagnostic methods
provides the required level of control over the gas pipeline conditions. Perhaps, it requires assessment
of the defects trends to change their dimensions, i.e., intensity of the corrosion processes. An attempt
to assess these trends comparing the defect dimensions inspected in 1997 and 2002 failed. The most

6
 7

representative in this respect is the section 31-40. The corrosion intensity at this section is shown
graphically at Figure 4.

140
120,5
120

100
Metal loss, gram

80

60

40

20
0,0 0,0 0,0 0,0 0,0 0,0 0,0 0,0 0,0
0
31 32 33 34 35 36 37 38 39 40
Pipeline section distance, km

8 7,3 7,3
7 6,5
Metal loss, gram

6
5
4
3 2,0
1,7
2 1,3 1,3
1 0,0 0,0 0,0
0
31 32 33 34 35 36 37 38 39 40
Pipeline section distance, km

Figure 4: Corrosion intensity at 31-40 km section of the pipeline

following inspection data of 1997 (upper) and 2002 (bottom)

As the said figure shows, the pipeline diagnostics carried out in 1997 records one defect at the
th
36 kilometer with metal loss of 120.5 gram. In other words, we have a defect of a considerable size
and, therefore, it would be interesting to see how its dimensions have changed within five years.
As we can see from Figure 4, the corrosion intensity level, according to inspection of 2002, of
is several times lower everywhere within the 10-km section than the one recorded in 1997 for one
defect. In other words, the defect recorded in 1997 seems to be absent. This case, referred as an
example, is not single. Therefore, without detailed explanation of the causes, we only emphasize that
data of the said two pipeline surveys, separated by with a 5-year period, give little information about
changes in dimensions of the defects. Thus, we encounter a problem of identification of the same
defects using data of twoinspections. The difficulties, caused by this problem, appeared to be
practically irreducible.

7
 8

3. COMPLEX EVALUATION OF TECHNICAL CONDITIONS OF A GAS PIPELINE

Said reasons brought us to the idea of a complex evaluation of a gas pipeline condition
according to inspection results. The method is based on definition of integral indicators of pipeline
conditions basing on the IP data.
The “integral evaluation” term is a generalized evaluation of the pipeline conditions, including
all pipeline defects, recorded by pipeline inspection, as terms of a statistical sampling. In such a
formulation, numerical characteristics of the relevant statistical sampling, such as mathematical
expectation, dispersion, standard deviation, mode, median, asymmetry, excess, etc., play the role of
integral indicators. Quite clear that these indicators will play an auxiliary role in preparation of the
plans for gas pipeline repair and renovation works. However, it is also obvious that the integral
indicators are convenient as an instrument for identification of the trends of changes of gas pipeline
condition both within the total pipeline length and its separate sections (elementary segments).
Let’s implement practically the above theory of the method of complex evaluation of the
conditions of Dolyna-Rososh 100 km section of DUSB-2 pipeline. As it was already mentioned, each
defect is characterised by five parameters. As far as the corrosive nature of defects is mainly
concerned, we’ll use only those parameters, which describe dimensions of defects. I.e., only three
defect characteristics will be used: depth, length and width. Thus, we arrive at more general, and,
undoubtedly, more representative method of evaluation of pipeline conditions basing on trends of
changes in numerical characteristics of a three-dimensional statistical sampling, as far as in our case
each defect will be characterised by three parameters: depth, length and width. Table 2 summarises
results of statistical treatment of the data obtained from the said two diagnostic examinations of the
pipeline.

Numerical characteristics of 1997 (post-overhaul 2002

samplings examination)
Total weight of the lost metal, in grams 836.72 4015.18

М 16.09 13.12

D 534.25 1395.41

Standard 23.11 37.36

Number of defects 51 306

Table 1: Results of inspection of 1997 and 2002

Corrosive pipeline damage within the five-year period makes 4015.18-836.72=3178.46 grams.
The average statistical defect mass decreased almost by three grams compared to 1997. This can be
explained by significant increase of the number of defects in comparison with 1997. Also, a substantial
growth of dispersion and root-mean-square deviation is observed.
As far as comparative characteristic of defects, recorded in 1997 and in 2002, appeared
impossible, and, besides, the inspection of 2002 failed to reveal a number of defects of considerable
size found in 1997, we can suggest a zero basis for comparison of results. In other words, we’ll asses
the gas pipeline condition, caused by its corrosive damages, according to relative corrosion intensity
along the total pipeline length (see Figure 3), basing only on results of the inspection of 2002.
The above allows us to prepare the plans of more extensive pipeline survey with direct
methods of its inspection. By the way, the plans of extensive pipeline survey should also account
problematic results of the inspection of 1997.

4. EVALUATION OF RESIDUAL LIFE OF A PIPELINE

As far as the corrosive defect depth is one of the basic parameters, used for evaluation of the
structural strength of pipelines, this parameter will be subject of our further consideration. The
following task is important practically: let’s determine the period of time T , during which, starting from

8
 9

a given time point T0 , the corrosive growth of the defect depth will not exceed the critical, or
admissible, defect depth. In other words, this is a task about the residual pipeline life.
We assume that the defect has the depth of d (T0 ) at the time T0 . Then we obtain that

within T years and for the rate of corrosion equal to υd , the defect depth becomes equal to d (T ) ,
which can be calculated from the following formula:
d ( T ) = d (T0 ) + Tυd (1)
Assuming that d (T ) = d kp , we obtain from formula (1), the length of the time period T ,
during which the defect depth will reach its critical level at the given rate of corrosion:
d kp − d (T0 )
T= . (2)
υd
In formula (2), dkp stands for the critical defect depth, i.e., such defect depth value, when
any of its further increase will result in pipe destruction.
Thus, formula (2) allows the Gas Pipeline Operator to estimate the length of the admissible
residual pipeline life basing on inspection data. As an example, let’s determine the integral value of
the residual pipeline life according to in-line inspection data obtained in 2002. Using parameter
R = 10.mm, which Rosen Technologies assumes as a minimum admissible residual thickness of the
pipeline wall, we calculate the critical defect depth as
d кр = δ − R , (3)

where δ is the rated value of the pipeline wall thickness, in mm.

The rate of corrosion can be calculated from the following formula:
υd = [d (T ) − d (T0 )]/ n . (4)
d (T ) − is the mathematical expectation of the defect depth, based on data
In formula (4):
of the pipeline examination during the period of T ; n − is the length of the period after the previous
pipeline examination, in years.
Using the formulas (3) and (4), we can rewrite formula (2) in the following manner:
n [δ − R − d (T0 )]
T= . (5)
d (T ) − d (T0 )
Table 2 contains results of calculations made according to formula (5) for the different values
of pipeline wall thickness. We assume in calculation that n = 5 (the length of the period passed from
the previous pipeline examination, in years); d (T ) = 2.00 mm (for the average defect depth value) and
d (T ) = 3.71 mm (for the defect depth recorded by the IP); d (T 0 ) = 0, because, as it was already
mentioned, a zero basis is assumed in calculations, as far as it was impossible to identify defects
using data of in-line inspection of both 1997 and 2002.

Rated pipeline wall thickness, Average defect Average defect depth =

mm depth = 2 mm 3.71 mm
16.5 14.6 7.9
17.5 17.1 9.2
Table 2: Residual life of the gas pipeline, years

As we can see from Table 2, the minimum value of the integral pipeline residual life makes
about eight years, which means that under existing conditions of corrosion there is no need to prepare
the plan of pipeline repair and renovation works. In practice, decision making requires similar analysis
for all pipeline areas with corrosion anomalies, which were detected by the II. According to Figure 2,
these are the following pipeline sections: 0-10 km, 20-30 km, 80-90 km, and 90-100 km. All the other
informational materials, received by the Gas Pipeline Operator during five years, should be also

9
 10

analyzed. Such an analysis should serve a basis for the final decision of a Gas Pipeline regarding
further pipeline operation, or additional pipeline survey, or pipeline repair and renovation works.
Since it’s impossible to identify pipeline defects following results of II of 1997 and 2002, the
above procedure of evaluation of gas pipeline condition is important rather from methodological than
any other standpoint. The pipeline surveys of 1997 and 2002 do not provide the Gas Pipeline Operator
with the most important thing: they say nothing about the trends of corrosion activity changes along
the total pipeline length during the last five years. Nevertheless, one should not be little importance of
the Table 2 data, as far as they are based on processing of a potent statistical sampling, which reflects
development of the pipeline defects during the five-year period.

5. PROBABILITY OF A PIPELINE FAILURE

As far as the residual pipeline life, presented in Table 2, is calculated on the basis of
mathematical expectation of the defect depth, or, in other words, on the basis of a certain average
value of the defect depth, it should be taken into account that technical standards of the safe pipeline
operation monitoring requires evaluation of probability of gas pipeline failure. The risk analysis
method, used for this purpose, involves all values of the relevant defect indicator in a form of a
statistical law describing its changes, rather than an averaged value of such defect indicator.
Figure 5 presents the relative density of defects, which may be also called by the engineering
science as a “polygon of densities of the studied random variable” (the defect depth in this particular
case).

1,2

1
0,89
0,82 1,00
Probability density and distribution function

0,94 0,99

0,8

0,61

0,74

0,6
0,46

0,4

0,22

0,21
0,2
0,15
0,13
0,26
0,08 0,07
0,05 0,06
0,04
0,04 0,01
0
2,16 2,62 3,09 3,55 4,01 4,48 4,94 5,41 5,87 6,33
Defect depth, in mm

Figure 5: Sampling from diagnostic data of 2002 (306

defects)

10
 11

The same figure presents also the accumulated relative density of defects, or density
distribution function. The experimentally obtained function of the random variable density distribution
allows to estimate the pipeline reliability depending on the depth of its corrosion defects; and its
variation with time depends on the rate of corrosion. In the case under consideration, F ( d ), will be
the function of the defect probability density distribution, which determines the probability of an event
for each value of d if the random variable X is less than d , i.e.:
F (d ) = P ( X p d ). (6)
In view of Formula (6), the probability of pipeline destruction I тр may be written in the
following form:
I тр = 1, 0 − P ( X p d ). (7)
If relative density of the defect depth probability is equal to 0.01, which corresponds to the
maximum defect depth value (see Figure 4), the accumulated density value will be equal to 0.99. Then
we can find from Formula (7):
I тр = 1, 0 − 0,99 = 0, 01, (8)
It means that the probability of pipeline destruction at a defect depth less than 6.1 mm, or the
probability of origin of a defect with 6.1 mm or larger depth, is 0.01.
The above considerations prove the need of auxiliary inspections examinations even in a case
when the residual pipeline life (the time of pipeline operation until the defect depth reaches its critical
value) makes about eight years, as in the case we have just considered.

CONCLUSIONS

1. The proposed method of complex assessment of a gas pipeline condition basing on the
data of its in-line inspection allows to estimate intensity of the corrosion processes by the number of
corrosion defects within specified pipeline sections.
2. Comparison of the values of complex valuation indicators of the pipeline condition, basing
on the data of two time-separated in-line inspections, is not less important from the practical
standpoint. In particular, it comes under our notice that the average defect depth within the pipeline
sections of 20-30 km and 80-90 km did not increase and even became lower during the five-year
period. The explanation seems to be simple from the first glance: it may be explained by substantial
growth of the small-sized defects during the five years. However, it would be reasonable for purposes
of more detailed analysis to single out defects found in 1997 from the sampling defects recorded in
2002, and then to make necessary analysis of their dimensions. Practically irreducible difficulties arise
on identification of defects using data of different pipeline examinations. In this concern, the 30-40 km
section of the pipeline is especially representative, and the total corrosion intensity value for this
pipeline section, basing on the inspection data of 2002, makes only 27.4 grams against 120.5 grams
of 1997.
3. The data presented in Tables 1 and 2 may serve a basis for planning and implementation of
the direct pipeline inspections and, also, they raise a problem regarding the defect recording accuracy
and possibility to identify defects basing on data of several time-separated in-line surveys.

11