Download as pdf or txt
Download as pdf or txt
You are on page 1of 17

Paper No.

05645
CORROSION 2005
Predicting Internal Pitting Corrosion of Oil and Gas Pipelines:
A Corrosion Engineering Approach
Sankara Papavinasam, Wally Friesen, R. Winston Revie, and Alex Doiron
Natural Resources Canada
CANMET Materials Technology Laboratory
568 Booth Street, Ottawa, Ontario K1A 0G1
Canada

ABSTRACT

In oil and gas production, one significant factor in the degradation of pipelines is internal pitting
corrosion. A corrosion engineer typically determines the risk caused by internal pitting corrosion based
on operating parameters, inline inspection, or online monitoring data. In this report time-series data from
oil fields were analyzed to determine if pipeline failures correlate with operating parameters,.

No significant differences in superficial oil, water, and gas velocities, and water-cut were observed
between lines that had and had not failed. Additionally, other operating parameters such as pipe
inclination, operating pressure (both maximum and average) and hydrogen sulfide (H2S), do not
correlate with pit growth rates. An online survey on the merits of monitoring techniques also indicates
that there is no one single technique that can be reliably used to measure the growth rates of internal
pitting corrosion.

INTRODUCTION

The oil and gas industry relies heavily on pipeline made from carbon and low-alloy steels. Corrosion of
these steels caused by CO2 and H2S materials has been, and remains, a major concern for the integrity of
pipelines. Localized corrosion (for example, isolated pitting) remains one of the more frequent
problems. Thus, when designing production equipment and transportation facilities, engineers need to
determine the risk due to internal pitting corrosion.1

Currently there is no single industry-standard approach to assess the risk of internal pitting corrosion..
There are a limited number of references in publications of the National Gasoline Association of
America, American Petroleum Institute, and Canadian Standards Association,2 but no professional body
or agency provides a standard guideline.

Copyright
©2005 by NACE International. Requests for permission to publish this manuscript in any form, in part or in whole must be in writing to NACE International,
Publications Division, 1440 South Creek Drive, Houston, Texas 77084 77084-4906. The material presented and the views expressed in this paper are solely
those of the author(s) and not necessarily endorsed by the Association. Printed in U.S.A.

1
Field operators require a tool to predict all stages of project development and subsequent operation.
Such a tool should answer the following questions:
• Does internal corrosion pose a significant risk?
• If so, when during operation will a failure occur?
• Where in the pipeline will the failure occur?
• What will be the failure mechanism?
• Which operating parameters should be monitored in order to correctly predict the failure?
• How can the predictions be validated and utilized in a user-friendly manner?

In attempts to answer these questions the approaches that have been developed can be classified broadly
into methods based on corrosion science, electrochemistry, and corrosion engineering. In this paper, the
merits and demerits of the corrosion engineering approach are reviewed. The corrosion science and
electrochemical models are discussed in other papers in this series.3, 4

CORROSION ENGINEERING APPROACH

A corrosion engineer typically determines the risk caused by internal pitting corrosion from field
operating conditions, data obtained from inline inspection and data obtained from online monitoring
techniques. The merits and demerits of this corrosion engineering approach are discussed below.

CORRELATION OF PIPELINE FAILURES WITH OPERATING CONDITIONS

Data provided by an oil production company from one of their fields were examined to determine if
there was any correlation between pipeline failures and operating conditions. For each pipeline, the
records in the dataset consisted of monthly values for the average daily oil, water, and gas production
and water cut. Time series of these values were constructed from the most recent and going back in time.
For those pipelines having gaps in the records, the time series was halted at the time of the first gap.

By including only those pipelines with records of at least 100 months, the data for 73 bare steel pipelines
were analyzed. These pipelines included 67 well lines (seven of which had failed), three satellite lines
(one of which had failed), and three transportation lines (one of which had failed). The outside diameter
and wall thickness of these pipelines ranged from 60.3 to 168.3 mm and 2.8 to 4.78 mm, respectively,
with most dimensions being 88.9 mm and 3.96 mm.

The simplest step in finding a correlation was to determine if a distinction could be made between failed
and non-failed lines with respect to simple flow variables. The explicit dependence of flow rate on the
(inside) pipe diameter, d, was removed by calculating superficial flow velocities according to:

4Qk
vk = (Eqn. 1)
πd2

where Qk is the volume flow rate of the fluid component k (and k = o,w,g for oil, water, and gas,
respectively). The gas velocity, vg, was calculated from the gas production rate, which was expressed as
m3/day at standard temperature and pressure. Since the actual temperatures and pressures of the fluids
were not given in the database, the true gas volume could not be determined and calculated, so vg is an

2
approximate value only. If the pipeline pressures and temperatures were comparable, then the relative
values of vg are approximately correct.

To facilitate the comparison between failed and non-failed lines, the mean value, Vk, and standard
deviation, σk, of vk were calculated. Subscripts n and f denote non-failed and failed lines, respectively.
For any difference between Vkf and Vkn to be significant, Vkf must be outside the limits defined by Vkn "
σkn.

The data for the individual lines and for the average values can be summarized using the following
factors: flow velocity, water cut, flow regime, pipe inclination, H2S concentration and pressure.

FLOW VELOCITY

Superficial oil velocity: Oil flow decreased with time for most failed lines but in all cases vof < 0.025 m/s
existed at the time of failure. For one line, this represented an order-of- magnitude decrease from a peak
velocity of 0.2 m/s. In Fig. 1, Vof is somewhat less than Von but during the entire period the difference
was always less than Fon.

Superficial water velocity: At the time of failure, vwf always varied by an order of magnitude of more
than 0.04 m/s to 0.9 m/s. Sudden substantial increases in vwf occurred in two lines about 65 months
before failure. These two lines are responsible for the jump in Vwf at -65 months as shown in Fig. 1. The
mean water velocity for the failed lines, Vwf, was somewhat higher than Vwn, but the difference was
consistently less than Fwn.

Superficial gas velocity: This quantity was calculated from the produced gas volume measured at STP
and is therefore very approximate. At failure, Vgs was < 4 m/s for all nine lines and one line exhibited
very large fluctuations. As was the case for both liquid components, Vgf and Vgn differed by less than Fgn
in ?

Water Cut
There was a wide variation in water cut among the failed lines. For two failed lines, the water cut
exceeded 90% for 200 months, but for another failed line it was greater than 60% for less than 50
months. As shown in Fig. 2, the average water cut was somewhat higher for the failed lines than for the
non-failed lines, but the difference was always less than the standard deviation.

Flow Regime
For multiphase flow, the aqueous-oil-gas interphases can take any one of an infinite number of possible
forms. These forms are delineated into certain classes of interfacial distribution called flow regimes. The
flow regimes depend on the inclination of the pipe (that is, vertical or horizontal), flow rate (based on
production rate), and flow direction (that is, upward or downward). For horizontal pipes, bubbly,
stratified, wavy, plug, slug, and annular flow regimes are possible.9

Figure 3 shows a plot of vl against vg for 69 lines (four non-failed lines with vg > 10 m/s are not
represented). The velocities were calculated from the average flow rates for the most recent 12-month
period. No real distinction between failed and non-failed lines was noted, although there were no
failures for lines with vg > 3.5 m/s. Furthermore, if the transition from stratified to slug flow occurs at vl
. 0.2 m/s then pipelines in both regimes failed.

3
Pipe Inclination
In Figures 4, 5 and 6, the two-dimensional pipeline side elevations derived from pipeline and
topographical surveys were compared with inline inspection data to predict where corrosion would
occur No correlation could be established between pipeline elevation and internal pitting corrosion.

H2s Concentration7
Using different data from failed pipelines, the pit growth rate was calculated as:

Pit growth rate = Pipe thickness/(time of first failure – time of installation).

As shown in Fig. 7, the growth rates were plotted against the partial pressure of H2S at the time of
failure Here again, no correlation between pit growth rate and thickness could be established.

Pressure7
In Fig.8, the pit growth rates (as determined in the H2S concentration section) were plotted against
normal operation pressure and in Fig.9 against maximum operation pressure. No correlation between pit
growth rates and pressure could be established.

IN-LINE INSPECTION (ILI) AND ONLINE MONITORING

Although many measurement techniques are available to investigate the reliability of an operating
pipeline, but each technique has its strengths and weaknesses. Some techniques can be used on-line,
whereas others are carried out off-line. Usually more than one technique is used so that the weaknesses
of one are compensated for by the strengths of another. In other cases, a combination of different
techniques can be synergistic.

Integrity management of pipeline infrastructure requires two components:


1. information on the current condition of the pipeline through sensing (monitoring and inspection)
and
2. interpretation of information for making maintenance and repair decisions.

Reliable sensors are requiredto monitor, inspect and detect internal pitting corrosion. Sensors are also
required for remote communication technologies and for the knowledge-based systems required to
process the information used to make decisions on maintenance and repairs.

An online survey was conducted to collect and analyze information on various techniques that are
currently being used around the globe to manage the risk of operating pipelines.8 Forty-seven people
responded to the survey. Respondents were from most sectors of the pipeline industry, including
pipeline companies, researchers, academics, and consultants. About 83% of the respondents were from
North America, while the remaining 17% were from Argentina, Australia, Belgium, India, Iran, South
Korea, South Africa, and Venezuela.

The majority of responses to the survey came from personnel interested in gas production (68%), oil and
gas production (59%), gas transmission (59%), oil transmission (58%), and oil production pipelines
(43%). One response each came from respondents interested in water, slurry, and gas/gas lift.

4
Approximately 60% of the respondents indicated they were interested both in internal and external
corrosion. The remaining people were interested either in internal corrosion (20%), or external corrosion
(10%) only, or skipped this question (10%).

The theory and application of these techniques are well documented.10 The techniques that pipeline
companies use to monitor internal pitting corrosion can be broadly grouped into the following six
classifications:

1. Physical techniques
2. Electrochemical techniques
3. Non-intrusive techniques
4. On-line monitoring techniques
5. Off-line monitoring techniques
6. New techniques

1. PHYSICAL TECHNIQUES

Figure 10 presents the physical techniques that companies prefer to use.

2. ELECTROCHEMICAL TECHNIQUES

As corrosion is electrochemical in nature, several electrochemical techniques can be used to monitor it.
These are shown in Fig. 11. Linear polarization resistance (LPR) is by far the most preferred and used
technique, followed by the electrochemical noise (ECN) technique. The industry is still not comfortable
using ECN, however, since this technique has been identified as less reliable than the LPR technique.
ECN also requires highly trained personnel to analyze and interpret the data.

3. NON-INTRUSIVE TECHNIQUES

As pipelines are typically operated under pressure and contain poisonous (H2S) and/or greenhouse (CO2)
gases, most companies prefer to use non-intrusive techniques to monitorpipeline corrosion. Figure 12
presents the non-intrusive techniques that companies use. Ultrasonic measurements and magnetic flux
methods are extensively used in the field. None of these techniques are preferred because they all are
considered unreliable.

4. DIRECT ONLINE MONITORING

During the regular operation of pipelines, measurements are frequently made, to determine their
susceptibility to corrosion. Figure 13 presents the on-line measurements that all companies use. The pH
is always measured because of its important influence on the corrosivity of fluids.

5. OFF-LINE MONITORING

To measure corrosion, the routine analysis of products, fluids, and other parameters is continually
carried out. Some of this data can be used indirectly to develop corrosion control strategies. Figure 14
presents the non-intrusive off-line techniques that companies use and their reliability.

5
6. NEW TECHNIQUES

About one quarter of the respondents indicated that they are aware of new corrosion monitoring
techniques on the horizon, including fibre optics (25%) and acoustics.

The factors that influence the companies’ decision for selecting monitoring techniques are: reliability
(75%), adaptability to operating conditions (53%), cost benefit (47%), and user-friendly operation
(25%). The responders emphasized that many operating factors will affect the performance of corrosion
measuring and monitoring techniques. These factors, which are of equal importance, are shown in Fig.
15 and include: temperature fluctuation, pressure fluctuation, environmental variation, and loss of
ruggedness during installation and operation.

DISCUSSION

The corrosion engineering approach to determine the risks posed by internal pitting corrosion is based
on the following six key questions.

1. Does internal corrosion pose a significant risk?


Vast experience has been accumulated in the industry over 50 years of operation. In theory, a corrosion
engineer could, based on operating conditions and data from monitoring techniques, predict if internal
corrosion is an issue. In practice, however, this happens after the event has occurred rather than before.

2. When during operation will a failure occur?


Industry has established threshold operating conditions when internal corrosion becomes an issue. This
rule-of-thumb approach has not always proved to be reliable.

3. Where in the pipeline will the failure occur?


Companies test, evaluate, and use many techniques to monitor pipeline reliability. Monitoring
techniques can be broadly classified into: physical techniques, electrochemical techniques, non-intrusive
techniques, on-line monitoring techniques, off-line monitoring techniques, and new techniques. If the
monitoring techniques are attached at appropriate locations or if the ILI technique can be used, the
location where the risk caused by internal corrosion has exceeded a pre-set limit can be determined.

4. What will be the failure mechanism?


Almost all failures are caused by localized corrosion in the form of isolated pitting. Depending on the
monitoring or inspection techniques used, the size and shape of the pits could be determined and
monitored.

5. Which operating parameters should be monitored in order to correctly predict the failure?
The merits and demerits of various techniques and operational parameters are very well understood
within the pipeline community. No single technique stands out to meet all the needs. On the other hand,
no standardized approach has been developed to determine the number of parameters that should be
analyzed to predict the failure.

The factors that influence decisions for selecting the appropriate monitoring technique are: the reliability
of the technique, its adaptability to operating conditions, cost benefit, and user-friendly operation.

6
It should also be emphasized that many operating factors will affect the performance of corrosion
measuring and monitoring techniques. The factors, which are of equal importance, include: temperature
fluctuation, pressure fluctuation, environmental variation, and deterioration of ruggedness after
installation and during operation.

6. How can the predictions be validated and utilized in a user-friendly manner?


Information on pipeline operating conditions is difficult to collect. Parameters are usually collected in
different databases. As a result, determining the best method to transfer data from one database to
another is not an easy task. In some cases, two or more pipelines may have identical location identifiers
and differ only with respect to licence or line number. The records in one database may or may not
contain licence or line numbers, so they cannot be matched to data in another database. A program often
needs to be written to read the data from a database and to transfer its contents in a format that can be
used in succeeding data manipulations. For these reasons, it is often difficult to validate the corrosion
engineering approach.

CONCLUSIONS
Based on this review of the corrosion engineering approach, the following conclusions can be made:

1. Field operating conditions are dynamic in nature. A sudden change in operating conditions can
have detrimental and long-term effects.
2. No single operational parameter (such as pipe flow, inclination, pressure, and H2S partial
pressure) can be correlated with pipeline failure. Although all these parameters are
interconnected, the current approach to measurement assumes no interactions among them,
which is clearly not the case. A possible alternate measurement approach would be to enter the
data into some sort of experimental design software to determine if any trend(s) can be observed.
3. Although considerable amounts of data on internal pitting corrosion and the operational
conditions of pipelines are available, the data has not been collected in a standardized and
consistent manner. For this reason data retrieval and its subsequent analysis is not user-friendly.
4. Several monitoring techniques are available, but no single technique can be universally used to
monitor all the parameters needed to predict internal pitting corrosion.

REFERENCES

1. M.B. Kermani, L.M. Smith, “CO2 Corrosion Control in Oil and Gas Production: Design
Considerations,” European Federation of Corrosion Publications, Number 23, EFC, 1997.
2. Canadian Standards Association (CSA) Standard Z662, 2003.
3. S. Papavinasam, R.W. Revie, and A. Doiron, “Predicting Internal Pitting Corrosion of Oil and Gas
Pipelines: Review of Corrosion Science Models,” NACE CORROSION 2005, Paper #5643,
Houston, Texas, 2005.
4. S. Papavinasam, W. Friesen, R.W. Revie, and A. Doiron, “Predicting Internal Pitting Corrosion of
Oil and Gas Pipelines: Review of Electrochemical Models,” NACE CORROSION CONFERENCE
2005, Paper #5644, Houston, Texas, 2005.
5. W.I. Friesen, “Failure in High Water-cut Pipelines,” Division Report WRC 96-27, July 1996.
6. A. Miller, “Pipeline Corrosion and Corrosion Control in Sour Devonian Multiphase Production,”
NACE Northern Area Eastern Conference and Exhibition, Paper No. 2A.2, Ottawa, Ontario, Canada,
Oct. 24-27, 1999.
7. Failure Data of Oil Production Pipelines, 1995.

7
8. S. Papavinasam and R.W. Revie, “Internet Survey on Currently Available Techniques for
Monitoring Internal Corrosion of Oil and Gas Pipelines,” NACE Northern Area Eastern Conference,
Ottawa, Ontario, Canada, Sept.15-17, 2003.
9. ASTM G170, “Standard Guide for Evaluating and Qualifying Oilfield and Refinery Corrosion
Inhibitors in the Laboratory,” 100 Barr Harbor Drive, West Conshohocken PA, USA, 19428.
10. NACE Report 3T199: “Techniques for Monitoring Corrosion Related Parameters in Field
Applications.”
11. J.Y. Sun, and W.P. Jepson, “Slug Flow Characteristics and their Effect on Corrosion in Horizontal
Oil and Gas Pipelines,” SPE 24787, 67th Annual Technical Conference Society of Petrol
Engineering, Washington, 1992.
12. J.M. Mandhane, , G.A. Gregory, and K. Aziz, “A Flow Pattern Map for Gas-liquid Flow in
Horizontal Pipes,” Int. J. Multiphase Flow 1, 537-53, 1974.
13. G. Hetsroni, “Handbook of Multiphase Systems,” Hemisphere, Washington, 237, 1982.

8
0.6
Water
0.4

0.2

0.0
Superficial Velocity (m/s)

0.2 Oil
0.1

0.0

20 Gas
10
0

-10
-250 -200 -150 -100 -50 0
Time (mo.)

Fig. 1 - Time dependence of the mean superficial oil, water, and gas velocities for failed (heavy line)
and non-failed (light line) pipelines. The dashed lines indicate the mean " standard deviation for
non-failed lines.

9
1.0

0.8

Water Cut 0.6

0.4

0.2

0.0

-250 -200 -150 -100 -50 0


Time (mo.)

Fig. 2 - Time dependence of the mean water cut for failed (heavy line) and non-failed (light line)
pipelines. The dashed lines indicate the mean " standard deviation for non-failed pipelines.

1.4
Superficial Liquid Velocity (m/s)

1.2

1.0

0.8

0.6

0.4

0.2

0.0
0 1 2 3 4 5 6
Superficial Gas Velocity (m/s)

Fig. 3 - Superficial liquid velocity plotted against superficial gas velocity for failed (squares) and non-
failed (triangles) pipelines.

10
Fig. 4 – Comparison of pit growth rates and pipe elevation.

Fig. 5 – Comparison of pit growth rates and pipe elevation.

11
Fig. 6 – Comparison of pit growth rates and pipe elevation.

12
25

Pit Growth Rate, mm/y 20

15

10

0
0 5 10 15 20 25 30
Partial Pressure of H2S (%)

Fig. 7 – % H2S versus pit growth rate.

25

20
Pit Growth rate, mm/y

15

10

0
0 3000 6000 9000 12000 15000
Normal Operation Pressure, kPa

Fig. 8 – Normal operation pressure versus pit growth rate.

13
25

20
Pit Growth Rate, mm/y

15

10

0
0 3000 6000 9000 12000 15000
Maximum Operating Pressure, kPa

Fig. 9 – Maximum operating pressure versus pit growth rate.

15

12
Number of Responses

0
X-ray
Mass-Loss

Destructive
Inspection

Inspection
Resistance

Measurement
Electrical
Coupons

Visual

Testing
Signature

(NDT)
Inline

(ILI)

(ER)

Non-
(FSM)
Field

Techniques

Fig. 10 – Physical techniques companies prefer to use for monitoring internal corrosion.

14
25
Note: The responders also mentioned other techniques including hydrogen permeation,
reference electrode testing and ER(2).
20
Number of Responses
Twelve responders had no answer.

15

10

Ammeter(ZRA) -

Ammeter(ZRA) -
Electrochemical

Electrochemical

Distortion
Harmonic
Polarization(LPR)

Zero-Resistance

Zero-Resistance
Potentiodynamic

Analysis
Spectroscopy
Noise (EN)

Impedance

Same Alloy
Dissimilar
(EIS)
Linear

Electrochemical Techniques

Fig. 11 – Electrochemical techniques the companies use for pipeline monitoring.

25
Note: The responders also mentioned other
techniques including close interval survey, guided
Number of Responses

20 wave ultrsonics, clipper tool, and chatodic


monitoring. Eight responders had no answer.
15

10

0
Emission

Activation and
Radiography

Electromagnetic

Electromagnetic
Ultrasonic

Electric Field

Acoustic
Remote Field
Eddy Current

Radiometry
Mapping

Technique

Gamma
Surface

Nonintrusive Measurement Techniques

Fig. 12 – Direct nonintrusive measurement techniques the companies use for pipeline monitoring.

15
Number of Responses

30
25 Note: 16 responders had no
20 answer.
15
10
5
0

pH re cit
y
in
g re hy nt e
tia
l
ge
n ty ... ...
ssu lo ul at u ap p oi gim n y ti vi M M
e er gr e t e x c
Pr Ve Fo w R Po O du n en
w m
p
rm
o De ow n ed on og
e
og
Fl
o Te e l i o l v C d r d r
Th F s o y y
o rro D iss (H (H
e
C em su
r
l .ch r es
E P
Direct On-Line Measurement Techniques

Fig. 13 – Direct on-line measurement techniques the companies use for pipeline monitoring.

25
Number of Responses

20
15
10
5
0
Gas Analysis

Analysis of

Alkalinity

Residual
Concentration
Metal Ion
Inhibitors

Oxidant
Residual

Analysis

of Dissolved
Chemical

Samples
Process

Solids

Indirect Off-Line Techniques

Fig. 14 – The indirect off-line measurement techniques the companies use.

16
20
Note: The responders also mentioned other factors including the survice
conditions, dirty pipeline corrosion geometry, sensitivity, application and
Number of Responses 16 location. Twenty-Three responders had no answer.

12

Environmental

Pressure Fluctuation
Installation/operation
Temperature

Loss of Ruggedness
Fluctuation

Variation

During
Factors

Fig. 15 – The companies opinions on what factors affect the performance of measuring/monitoring
techniques.

17

You might also like