Advances in In-line Inspection Technology for Pipeline Integrity

by: Jeff Sutherland, Hernan Paz

BJ Pipeline Inspection Services
V V A An nn nu ua al l I In nt te er rn na at ti io on na al l P Pi ip pe el li in ne e C Co on ng gr re es ss s
O Oc ct t 1 18 8- -2 20 0
t th h
2 20 00 00 0
M Mo or re el li ia a, , M Me ex xi ic co o
P Pa ag ge e 2 2 o of f 1 13 3
A Ab bs st tr ra ac ct t
In Magnetic Flux Leakage (MFL) inspection technology, corrosion anomalies are detected and identified via their leakage
field due to changes in wall thickness. With the increasing demand for higher accuracy and reliability in metal loss
corrosion sizing, we present some of the distinctive advances made in the MFL pipeline in-line inspection industry.
Designed and built with state of the art technology, a novel family of high-resolution MFL pipeline inspection tools is the
first to successfully incorporate these advances into proven operational tools. We explain the technological advances of
inertial mapping, circumferential sensor measurements, neural networks for metal loss sizing, and gas-bypass speed
control. Experiences from recent projects is discussed including quality assurance practices within a pipeline integrity
program.
I In nt tr ro od du uc ct ti io on n
Today pipeline operators have many options when choosing an internal inspection technology to investigate the
structural integrity of a pipeline. Two ILI techniques are available for wall thickness measurement, namely magnetic flux
leakage (MFL) and ultrasonic (UT). UT pigs require introduction of a liquid into the pipeline to couple the sensor signal
to the pipe wall, require a very "clean" pipeline and product and are generally excluded from inspecting gas pipelines. The
MFL method involves inducing a magnetic field into the pipe wall and sensing leakage of the field inside the pipe as the
wall thickness changes. The magnetic flux leakage technique is the most commonly used technique to inspect large
diameter gas transmission lines [1,2].
For gas transmission pipelines, the fact that the medium in which the inspection pig must travel is compressible
presents added challenges in maintaining velocity control in order to collect good quality data. Additionally the product
flow rate of a gas transmission pipeline must typically be reduced during an inspection run. For transmission lines
operating at or near capacity the economic impact of reduced throughput can be greater than the cost of the inspection
survey itself. Incorporating gas by pass and on-line speed control into an MFL pig improves data quality and reduces
throughput reduction costs.
The term "high resolution" has become somewhat confused with "high accuracy". The two may imply the same thing
but in fact are different. The term high resolution implies increasing the number of sensors and/or by increasing the
sampling rate. Such changes involve acquiring more data. It says nothing about the quality or accuracy of the data
collected. Accuracy includes the amount of uncertainty in a measurement and therefore reflects a system characteristic
since it is determined by factors beyond acquisition resolution [3].
The goal of incorporating advances in technology into a new inspection pig should be to provide better more reliable
information not just more information. Improvements in the miniaturisation and speed of electronic and computer
components have facilitated the incorporation of many more sensors into the latest inspection tools.
The following technological advances have been incorporated into the design and operation of the advanced inspection
tool shown in Figure 1. These advances have led to significant improvements in quality and accuracy of data and reliability
of tool performance.
New design, modelling and conceptualisation tools
3 Component magnetic flux vector sensor head with integrated Eddy Current ID/OD discriminator.
Digital signal processing horsepower.
Neural Network data analysis
Gas Bypass with speed control
Strapdown inertial navigation system (INS)
Modern software and information technology systems
P Pa ag ge e 3 3 o of f 1 13 3
GAS FLOW
SPEED CONTROL MECHANISM
INERTIAL NAVIGATION SYSTEM
3 AXIS HALL EFFECT SENSORS
Gas Bypass/Speed Control
BJ NPS 24 MFL Metal Loss Inspection Tool
GAS BYPASS
Figure 1. The Vectra MFL In-line Inspection Tool
T Th he e B Ba as si ic cs s
Magnets produce magnetic fields that exert a force of attraction on ferromagnetic materials. Lines of flux can represent
the strength and direction of the magnetic field. Magnetic flux tends to travel through steel rather than air or gas since it is
a ferromagnetic material. Magnetic saturation occurs when the amount of flux through an area starts to limit itself or
saturate Once the steel becomes saturated some of the flux actually starts to significantly leak out of the material. The
magnetisation system is the foundation upon which magnetic flux leakage measurement is built. Accurate determination
and interpretation of the length, width and depth of the defects relies upon a field that is not only strong but uniform and
consistent.
For pipeline inspection tools, a magnetic circuit is constructed with magnets, pliable steel brushes to couple the
magnetic flux into pipe wall and a "backing bar". This magnetic assembly is represented in Figure 2. Flux leakage is
measured by placing sensors adjacent to the pipe wall within the magnetic circuit. These magnetic assemblies are repeated
completely around an inspection vehicle to provide full circumferential coverage. Stronger magnets can help minimise the
effects of changes in wall thicknesses, stress and velocity since they produce strong leakage fields [4]. The Eddy Current
sensor was incorporated into the sensor head design of the tool in order to interpret and correlate its ID/OD signals with
the magnetic measurements. Elimination of a secondary sensor ring also affords a more robust mechanical tool design.
Advances in electronics, software, digital signal processing (DSP) and data storage technologies allow massive
amounts of data to be collected and stored within a self contained pipeline inspection vehicle. As these technological
N
S
S
N
Sensor
Magnetic
Flux
Pipewall Brushes
Magnets
F Fi ig gu ur re e 2 2. . M MF FL L P Pi ip pe el li in ne e I In ns sp pe ec ct ti io on n M Ma ag gn ne et ti ic c C Ci ir rc cu ui it t
P Pa ag ge e 4 4 o of f 1 13 3
advances improve metal loss detection and sizing accuracy, other inspection variables that could affect sizing accuracy
should also be examined in order to make appreciable improvements [4].
C Ci ir rc cu um mf fe er re en nt ti ia al l S Se en ns so or rs s
The magnetic field (H) and the flux density (B) indicated above exist as vector quantities (Figure 4). In order to
precisely measure the flux leakage, not only the magnitude of leakage must be measured but the direction of leakage must
be considered as well.
Hall Effect sensors collect flux data in each of the axial, radial and circumferential directions allowing for
measurement of the absolute flux leakage field vector. BJ was the first to utilise the circumferential sensor technology and
is the only inspection vendor using it in field operations.
Incorporating the circumferential sensor provides much needed information for improving defect length and width
accuracy. More accurate length and width measurements improve defect depth estimation. Since MFL tools use an
axially oriented magnetic field it is extremely difficult to detect axially oriented pipe wall defects such as long narrow axial
corrosion (LNAC or NAEC) and cracks using just the axial and radial sensors. The circumferential sensor provides data
F Fi ig gu ur re e 4 4 U Us se e o of f t th he e c ci ir rc cu um mf fe er re en nt ti ia al l s se en ns so or r t to o d di is st ti in ng gu ui is sh h a a l lo on ng g, , n na ar rr ro ow w c co or rr ro os si io on n d de ef fe ec ct t
F Fi ig gu ur re e 3 3. . M Ma ag gn ne et ti ic c F Fl lu ux x C Co om mp po on ne en nt ts s a as s r re ef fe er re en nc ce ed d f fo or r a a
p pi ip pe el li in ne e. .
P Pa ag ge e 5 5 o of f 1 13 3
that greatly enhances the probability of detecting such defects. [5] Because of the critical length dependence in most defect
assessment criteria (eg. B31G,NG-18,RSTRENG [9,10]), obtaining correct length and depth values is crucial.
D Da at ta a A An na al ly ys si is s a an nd d I In nt te er rp pr re et ta at ti io on n
Once defect signals have been collected they must finally be interpreted. It is at this analysis and interpretation stage
where the most significant difficulties can arise. Resolution of the data collected is a factor in the final analysis but as
stated previously high resolution should be considered in terms of better accuracy. And with the detailed information that a
high-resolution tool provides, more detailed severity criteria can be used with greater confidence. Of course more
sophisticated analysis and interpretation techniques must be used to process and manage the large amounts of data as well.
A combination of many analysis methods is the most reliable and accurate choice of in which artificial intelligence is
taking a significant role [6].
The most widespread use of artificial intelligence is neural network computational techniques. Neural networks attempt
to mimic the human brains past experience decision making process. The interaction of nodes in a neural network is
analogous to the interactions of neurons in the brain as shown in Figure 7. These neural networks can be useful in pattern
qualification and quantification where there are too many variables for effective traditional analysis or if the pattern is
vague and not clearly defined [7].
The accuracy of defect sizing obtained allow for confident calculations of failure pressures. Hence prioritisation of
defects can be done based on their predicted threat to the integrity of the pipeline. Having accurate data leads to fewer digs
and repairs and higher overall confidence in the integrity plan. However, field verification and assessment of any defect
should be performed for any engineering or repair decision. Traditionally, the verification of defect assessment has
become the determining task of the success of any project. It promotes that the quality and reliability of the inspection
information and company. Results from verification digs from a recent gas pipeline inspection project in South America
are shown in Figure 6.
Other recent results have been published from an extensive evaluation of a recent oil pipeline inspection where the
Vectra depth sizing accuracy was found to be reliably better than 6% WT. [8] But to do a true comparison, defects should
be assessed and considered for more than just peak depth. To do more advanced integrity and pressure based assessments
like RSTRENG [9,10], detailed depth profile information along the entire length of the defect or group of defects in
question should be compared (as illustrated in Figure 7). From such comparisons, the operator can then assess the
accuracy of predicted failure pressures from the inspection as well [8]
Input 2
Input 3
Input N
Input Layer
Output Layer
Hidden Layer
Bias nodes
Output 1
Output 2
...
MFL Signal MFL Signal
Parameters Parameters
Defect Size Defect Size
(L,W,D) (L,W,D)
F Fi ig gu ur re e 7 7. . S Sc ch he em ma at ti ic c l la ay yo ou ut t o of f a a N Ne eu ur ra al l N Ne et tw wo or rk k
P Pa ag ge e 6 6 o of f 1 13 3
Peak Depth Verification
0%
10%
20%
30%
40%
50%
60%
70%
80%
90%
100%
0% 10% 20% 30% 40% 50% 60% 70% 80% 90% 100%
BJ Reported Peak Depth (%WT)
F
i
e
l
d
P
e
a
k
D
e
p
t
h
(
%
W
T
)
F Fi ig gu ur re e 6 6. . F Fi ie el ld d c co om mp pa ar ri is so on n/ /v ve er ri if fi ic ca at ti io on n f fo or r p pe ea ak k d de ep pt th h p pr re ed di ic ct ti io on n a ac cc cu ur ra ac cy y o of f c cl lu us st te er re ed d d de ef fe ec ct ts s
f fr ro om m a a r re ec ce en nt t g ga as s p pi ip pe el li in ne e i in ns sp pe ec ct ti io on n . . T Th he er re e a ar re e o ov ve er r 1 10 00 0 i in nd di iv vi id du ua al l d de ef fe ec ct ts s r re ep pr re es se en nt te ed d h he er re e s so o
m mo or re e t th ha an n j ju us st t d de ep pt th h a ac cc cu ur ra ac cy y c ca an n b be e a as ss se es ss se ed d. .
P Pa ag ge e 7 7 o of f 1 13 3
G Ga as s B By yp pa as ss s & & S Sp pe ee ed d C Co on nt tr ro ol l
Velocity variations associated with pigging in gas pipelines affect the flux leakage signal because it is distorted under
high speed conditions resulting in unusable MFL data and incomplete coverage of the pipeline. The Vectra magnetic flux
leakage inspection tool incorporates a dynamic gas bypass and speed control system into its design. [19] (Figure 1). Gas
bypass decouples the pig speed from the pipeline gas flow. The velocity of the pig is then controlled by a valve or throttle
control mechanism. A hole through the centre of the tool's cupped front magnetiser module forces the gas flow away from
the sensor heads and arms, resulting in better quality of data collected. The velocity plots from a recent 30" gas line
inspection is shown in Figure 8, where the tool maintained itself well within its operating speed range while the line speed
at 8.3 m/s did not need to be altered to accommodate the ILI tool. By maintaining a more constant speed during an
inspection the flux leakage field has less variability which in turn improves defect characterisation.
F Fi ig gu ur re e 7 7 . . ( (a a) ) V Ve ec ct tr ra a p pr re ed di ic ct ti io on n o of f d de ef fe ec ct t l la ay yo ou ut t a an nd d d de ep pt th h
GRP 132
0-10 10-20 20-30 30-40 40-50 50-60 60-70 70-80 80-90 90-100
( (b b) ) F Fi ie el ld d m me ea as su ur re em me en nt t r re es su ul lt ts s o of f d de ef fe ec ct t l la ay yo ou ut t a an nd d d de ep pt th h
P Pa ag ge e 8 8 o of f 1 13 3
F Fi ig gu ur re e 8 8. . V Ve el lo oc ci it ty y p pl lo ot t f fr ro om m r re ec ce en nt t 3 30 0" " i in ns sp pe ec ct ti io on n. .
T Th he e t to oo ol l s sp pe ee ed d w wa as s 3 3 m m/ /s s a an nd d w we el ll l u un nd de er r t th he e a av ve er ra ag ge e g ga as s s sp pe ee ed d o of f 8 8. .3 3 m m/ /s s
Another design consideration in speed control is that of incorporation of a backup fail safe or override mechanism
should the throttle control fail open thereby preventing the tool from moving in the pipeline. The BJ gas bypass/speed
control pig incorporates an autonomous bypass override system (B.O.S.). This system consists of a flapper valve operating
via independent power source and control electronics. The valve actuates in a positive close manner and is time based.
Todays new gas transmission pipelines are operating at higher and higher pressures while demands are also being placed
on existing lines to maintain maximum allowable operating pressure (MAOP). During an MFL inspection the gas
throughput of the line typically must be reduced substantially with conventional ILI tools in order to accommodate data
acquisition speeds (1 - 4 m/s). If a transmission line is operating at maximum throughput there is an additional cost of
loss revenue from the "lost product" associated with the reduced throughput during inspection.
As the demand for gas grows, more and more large gas transmission pipelines may be required to operate at capacity
year round. Inspecting such pipelines may only be economically feasible with gas bypass capable inspection pigs.
Another example is that of a NPS 36 pipeline at 1000 psi and running at 9.5 m/s. Using the bypass technology, the
inspection tool ran at 3 m/s while allowing the full gas throughput at 9.5 m/s. The difference in gas delivered over that
time compared to reducing the gas speed to 3 m/s and running a conventional tool was 28 million cubic meters which was
worth $1,000,000 in revenue to the pipeline company [12] That product delivery revenue significantly overshadowed the
cost of running the inspection tool, making it a "profitable" inspection project over a conventional tool inspection.
D De ef fe ec ct t L Lo oc ca at ti io on n U Us si in ng g I In ne er rt ti ia al l N Na av vi ig ga at ti io on n a an nd d G GP PS S
Once a severe or critical defect has been characterised it must be located in the field. The BJ MFL tool incorporates a
strapdown inertial navigation system (INS) to determine position and attitude of the pig along its trajectory within the pipe.
Three axis gyros and accelerometers combined with odometer information are used to determine three-dimensional
changes in the pigs position on earth as it travels through the pipeline. Such a system has been well proven via the
Geopig
TM
[13-17].
Actual geographic co-ordinates are calculated by establishing Global Positioning System (GPS) control points along
the pipeline and then tying the inertial data to these points. By transforming the pigs trajectory into the tie points,
accurate "real world" GPS co-ordinates are obtained for any point along the pipe [18].
0
1
2
3
4
5
6
7
8
9
10
0 20000 40000 60000 80000 100000
Odo (m)
V
e
l
(
m
/
s
)
Avg Tool Speed Avg Gas Speed
P Pa ag ge e 9 9 o of f 1 13 3
From knowing all points along the pipeline, we can generate a "virtual pipeline", created from the processed inertial
data directly. All pipeline features are given a co-ordinate and ancillary information such as bends and the welds in the
pipeline are all identified and provided which aids in location in the field (Figure 9).
The sensitivity of the inertial devices provides information regarding dents and other physical anomalies that affect the
tool's motion the pipeline. Inertial deflection information can then be used to correlate the location of dents with specific
metal loss regions that are often the result of third party damage.
Since the final processed inertial data (plan and profile) is in geographic co-ordinate form all the data collected by the
MFL tool can be readily be presented and incorporated into a Geographic Information System (GIS) or an automated
mapping/facilities management (AM/FM) system.
P Pa ag ge e 1 10 0 o of f 1 13 3
D Da at ta a P Pr re es se en nt ta at ti io on n
In today's digital world it's impractical to attempt to manage a pipeline integrity program without some formal
information technology especially with the amounts of data that ILI tools can provide. This can be as simple as a desktop
spreadsheet/database to a company wide integrated GIS system.
In either case, one of the time-consuming and critical issues is getting the new data into the system. Although provided
with each inspection, paper copies of the report are usually only good as a reference. The useful data comes in a digital
"softcopy" form such as in a spreadsheet or even a pre-determined formatted file agreed to by the pipeline company and
the inspection company. Figure 10 demonstrates some options from recent examples of preferred reporting formats for
pipeline operators. With the flexibility of commercially available databases, data can be reported in many different ways.
The biggest benefit comes from the flexibility to sort and filter information on demand according to different criteria, eg.
F Fi ig gu ur re e 9 9. . ( (a a) ) T Th he e " "V Vi ir rt tu ua al l P Pi ip pe el li in ne e" " g ge en ne er ra at te ed d f fr ro om m i in ne er rt ti ia al l d da at ta a
( (b b) ) " "V Vi ir rt tu ua al l P Pi ip pe el li in ne e" " s sh ho ow wi in ng g a a 3 3D D b be en nd d, , a a m ma ar rk ke er r b bo ox x l lo oc ca at ti io on n ( (g gr re ey y b bo ox x j ju us st t b be ef fo or re e t th he e b be en nd d) )
a an nd d a a h he ea av vy y w wa al ll l s se ec ct ti io on n ( (p pu ur rp pl le e) ). . T Th he e b bl lu ue e a ar rr ro ow w i in nd di ic ca at te es s N No or rt th h. .
P Pa ag ge e 1 11 1 o of f 1 13 3
all defects > 20%WT between 100 and 2000 meters (odometer) in order of decreasing predicted failure pressure. Hence
prioritisation of the worst defects has already been done. Understanding the way that a pipeline company prefers the data
format has resulted in less time taken in manipulating and reformatting data for input into an integrity database and less
confusion and potential mistakes from what data means. This allows more time to be spent reviewing and prioritising for
upcoming maintenance planning and scheduling.
C Co on nc cl lu us si io on ns s
The goal of the inspection vendor should be to provide as accurate information as possible regarding the metal loss
condition of a pipeline so that the pipeline operator can make the best decisions with respect to integrity assessment. The
accuracy of magnetic flux leakage as a technique to detect and size metal loss defects in gas pipelines will continue to
improve as the latest technological advancements are incorporated into both the ILI tools and the interpretation of the data
collected. Ongoing improvements in electronics, computing power, software and data analysis techniques should readily
facilitate this.
Gas transmission pipelines previously not inspected due to reduced gas throughput considerations now have the option
of running gas bypass MFL inspection pigs. Many pipeline operators can now benefit from flexibility in scheduling of ILI
Figure 10. (a) Example of formatted file for input to pipeline operator's
integrity database. [19]
(b) Example of prioritisation report based on predicted failure pressure and
depth.
P Pa ag ge e 1 12 2 o of f 1 13 3
runs while greatly reducing or eliminating gas throughput costs. Onboard pig speed control will also provide more reliable
MFL data and reduce the chances of reruns caused by speed excursions.
MFL data provided with inertial navigation data will allow ease of incorporation into a pipeline GIS. The MFL data
can then be accurately cross referenced to other pipeline information at that location, thereby, providing more complete
pipeline integrity information. Information technology has advanced and can be readily utilised for easier pipeline data
reporting and management.
Other novel solutions to ongoing gas transmission inspection and integrity problems should be sought via the
application of the latest technologies as they emerge.
P Pa ag ge e 1 13 3 o of f 1 13 3
R Re ef fe er re en nc ce es s
1. Crouch A. E., In-Line Inspection of Natural Gas Pipelines, Gas Research Institute Topical Report GRI -91/0365,
May 1993.
2. G. J. Posakony and V. L. Hill, Assuring the Integrity of Natural Gas Transmission Pipelines, Gas Research Institute
Topical Report GRI-91/0366, 1992.
3. M. Siebert, Differences Between Low Resolution and High Resolution Pigs, Nowsco Pipeline Services, 1995.
4. T. A. Bubenik, J.B. Nestleroth, R.J. Eiber and B.F. Saffell, Magnetic Flux Leakage (MFL) Technology for Natural
Gas Pipeline Inspection, Gas Research Institute Topical Report GRI-91/0367, November 1992.
5. J. Sutherland, M. Siebert, "Application of the Circumferential Component of Magnetic Flux Leakage Measurement for
In-Line Inspection of Pipelines", NACE Corrosion Expo 99, San Antonio Tx, USA, April 1999.
6. J. Sutherland, M. Siebert, "Development and Operational use of Artificial Intelligence in In-line MFL Pipeline
Inspection"", Review of Progress in QNDE, Vol 18. (eds) D.O. Thompson, D.E. Chimenti, (Plenum Press, NY, 1998).
7. C.H. Chen, "IEEE Fuzzy Logic and Neural Network Handbook ", McGraw-Hill, Inc., 1996.
8. T. Morrison, A Bhatia, N.S. Mangat, G Desjardins, "Validation of an In-Line Inspection Metal Loss Tool", ASME
paper IPC00-0050, Proceedings of IPC , International Pipeline Conference 2000, Calgary, Alberta, Canada, October
2000
9. ASME B31 Code for Pressure Pipeline, Section 4 (B31.4) "Liquid Petroleum Transportation Pipelines", and Section 8
(B31.8) " Gas Transmission and Distribution Piping Systems", 1995
10. J.F. Kiefner, P.H. Vieth, "A Modified Criterion for Evaluating the Remaining Strength of Corroded Pipe", American
Gas Association Catalog No. L51609, Dec 1989.
11. B. Long, M Siebert, J Sutherland, R. Wade and T. Sawyer, "Field Trial Experiences with a State of the Art Magnetic
Flux Leakage Pig", Pipeline Week and Exhibition, Houston Tx, USA, Sept. 1997.
12. J. Sutherland, S. Clouston. "Advances in Magnetic Flux Leakage Measurement and Inspection for Metal Loss in
Pipelines", 3
rd
International Conference on Pipeline Rehabilitation and Maintenance, Abu Dhabi, UAE, May 1999.
13. G. Cataford and D. Joerissen, Pipeline mapping using an inertial inspection tool, First National Oil & Gas
Technical Conference, Calgary, March 1995.
14. J. A. Czyz and J. R. Adams, 1994. Computation of pipeline bending strains based on Geopig measurements. Pipeline
Pigging and Integrity Monitoring Conference, Houston, February 1994.
15. J. A. Czyz, C. Fraccaroli, A. P. Sergeant, Measuring Pipeline Movement in Geotechnically Unstable Areas Using An
Inertial Geometry Pipeline Inspection Pig, ASME International Pipeline Conference, Calgary, June 1996.
16. M. Cox, A. Garrigus, W. Walker and R. Wade, Pipeline monitoring and remedial action from inertial geometry
surveys in buried pipelines, The Pipeline Pigging Conference, Houston, Texas. February, 1995.
17. H. A. Anderson, C. T. McDougall and J. R. Adams, Pipeline evaluation using the geometry pipeline internal gauge
(Geopig), Pipeline Pigging and Integrity Montioring Conference, Houston, February 1993.
18. R. L. Wade and J. R. Adams, An Integrated Approach For Pipeline Fitness For Purpose Determination Using
Corrosion and Geometry Pipeline Pig Inspection Systems, Pipetech, Pattaya, September 1995.
19. Morrison Scientific Inc. Format for Corrosion Data Ver 1.1 June 1997.