IBP1084_19

COMBINED INSPECTION (MFL-A, MFL-C AND

IEC) OF AN OIL TRANSMISSION PIPELINE
FOR AN OPERATOR IN BRAZIL
Vincent Brughuis1, Rafael Almeida2,
Kathrin Schroeer3 , Oliver Knupfer4, Phelippe Viana4

Copyright 2019, Brazilian Petroleum, Gas and Biofuels Institute - IBP

This Technical Paper was prepared for presentation at the Rio Pipeline Conference and Exhibition 2019, held
between 03 and 05 of September, in Rio de Janeiro. This Technical Paper was selected for presentation by the
Technical Committee of the event according to the information contained in the final paper submitted by the
author(s). The organizers are not supposed to translate or correct the submitted papers. The material as it is
presented, does not necessarily represent Brazilian Petroleum, Gas and Biofuels Institute’ opinion, or that of its
Members or Representatives. Authors consent to the publication of this Technical Paper in the Rio Pipeline
Conference and Exhibition 2019.

Abstract

Pipeline integrity is negatively affected by a multitude of anomalies, flaws and threats,

which sometimes may even interact. Trying to assess multiple threats with only a single
inspection technology often falls short, as each technology has its distinct strength.
To increase the reliability of integrity assessments, combining the evaluation of two or
more different in-line inspection (ILI) data sets is possible. While these data sets may stem from
consecutive runs with different ILI tools, there is also the possibility to merge different
inspection technologies on one in-line inspection tool. Performing several inspection tasks in
one ILI run significantly reduces the risks and efforts involved.
This paper describes a failure of one pipeline by an operator in Brazil caused by the
above-mentioned combination of complex anomalies, which could not be detected by one
single inspection technology alone. It describes the failure cause, ROSEN-proposed combined
technology to ensure threats affecting the safe operation of the pipeline are being detected and
reported, the findings of the combined inspection, and the benefits of implementing integrity
solutions in such a critical asset.

Keywords: Combined ILI. MFL-A. MFL-C. IEC. corrosion.

1. Introduction

The offshore oil pipeline in question has a long inspection history with ROSEN, dating
back to 1999. ILI inspections were performed regularly during this period and the main scopes
of these inspections were circumferential metal loss and internal geometry.
The tools used to detect metal loss during this period were based on the magnetic flux
leakage (MFL) principle. MFL tools induce magnetic field on the pipeline wall, which
generates different responses depending on the amount of metal in the inspected area. The
intensity of these responses is then recorded by sensors and can be used to infer pipeline wall
thickness and in addition, the dimensions (depth, length and width) of an existent anomaly.

______________________________
1
Bachelor of Mechanical Engineering with specialty in Drilling and Production Technology - ROSEN Brazil
2
Chemical Engineer – ROSEN Brazil
3
Ph.D., Geophysics – ROSEN Germany GmbH
4
Chemist – ROSEN Brazil
 Rio Pipeline Conference and Exhibition 2019

MFL tools are a good choice to detect anomalies transversally oriented to the applied
magnetic flux. Due to the axial direction of the MFL-A magnetic field, the flux is easily
disrupted for round and circumferential orientated indications (Figure 1, indications A and C),
but not easily disrupted by long narrow axial defect (Figure 1, indication B).

Figure 1. MFL-A Magnetic Field Orientation

ROSEN’s last inspection before the pipeline failure was an MFL-A inspection carried
out in April 2018.
The deepest metal loss anomaly found during the 2018 inspection was reported as being
66% deep. No anomaly was reported with such severity that could have resulted in this pipeline
to leak.

Table 1. Pipeline Inspection History with ROSEN

Year Tool
1999 MFL-A
2011 MFL-A/HD
2015 MFL-A
2017 XT
2018 MFL-A

After the final report delivery in September 2018, ROSEN was reported a leakage on
this pipeline by the contractor, which was generated by an internal metal loss on a girth weld.
As part of ROSEN’s standard procedure in such instance, a Root Cause Analysis (RCA) was
carried out internally to determine why no anomaly with such severity was included in the final
report.

2. Abbreviations

o EC – Eddy Current
o HD – Basic Geometry Tool
o IEC – Internal Eddy Current Tool
o ILI – In-line Inspection
o ID – Internal Diameter
o MFL-A – Axially oriented magnetic flux leakage tool
o MFL-C – Circumferentially oriented magnetic flux leakage tool
2
 Rio Pipeline Conference and Exhibition 2019

o POD – Probability of Detection

o POF – Pipeline Operators Forum
o POI – Probability of Identification
o RCA – Root Cause Analysis
o XT – Mechatronic Caliper High Resolution Tool

3. RCA – MFL-A 2018 Inspection

3.1. MFL-A Inspection

Initially, the scope of the work was revised. The metal loss detection tool used on this
pipeline was an MFL-A, which scope is to detect circumferentially oriented, volumetric wall
loss anomalies, like general metal loss, pitting, axial and circumferential grooving, and
circumferential slotting anomalies.

Figure 2. Anomaly Dimension Classes Defined by POF

There are two dimension classes not covered by the standard MFL-A: pinhole and axial
slotting. In addition, even though MFL-A is able to detect axial grooving, its sizing capabilities
in terms of length, width and depth is limited.
The dimension class of the anomaly that caused the leak could be a possible root cause
for the underestimation of these anomalies’ in the MFL-A final report.

3.2. Data Quality

The MFL-A run system check shows that the recorded inspection data was of good
quality, except for minor sensor loss. The acceptable operational velocity range for the used
tool was between 0.5 and 3 m/s. Velocity was within range over the entire pipeline length:
Over nearly the complete line length, the standard magnetization values of 10 - 30 kA/m
were achieved, except for some installation areas, where these values were slightly below the
pre-defined range.
The actual ILI conditions did not differ from the conditions as they were foreseen. Data
quality was good, and standard specifications could be met.

3.3. Processing Parameters and Sizing Algorithms

3
 Rio Pipeline Conference and Exhibition 2019

All processing parameters used on this dataset were re-checked and no inconsistencies
were found. The same goes for the sizing algorithm, which was adequate for the anomalies,
considering the MFL-A scope.

3.4. Signal Comparison

Since data quality and processing parameters were fine, a signal comparison was
performed. The signal comparison approach proposed by Almeida R. (2017) was applied
aiming on identifying if there was corrosion growth between the 2015 and 2018 inspections:

Figure 3. MFL-A Dataset 2015

Figure 4. MFL-A Dataset 2018

While comparing the same example area on both datasets, a clear corrosion growth was
observed. Anomalies visible on the 2018 dataset were barely visible on the inspection data from
2015. These new amplitudes were mainly recorded at the bottom part of the pipeline.
In addition, it was noticed, especially at girth welds, that the amplitudes generated by
metal losses were significantly higher in 2018 than they were in 2015.

Figure 5. 2015/2018 Datasets, Amplitudes Recorded at the Same Girth Weld

It was possible to notice unusual signal characteristics, especially at the girth weld where
the leakage occurred. The leaking indication showed a complex shape, being composed of
multiple metal loss peaks close to each other, resulting in long and wide clusters at the girth
weld area.

4
 Rio Pipeline Conference and Exhibition 2019

Figure 6. 2018 Dataset, Multiple Amplitudes Recorded Over the Same Girth Weld

The determination of anomaly area by the MFL technology has restrictions. The
magnetic field that leaks from the anomalies usually exaggerates the anomaly dimension
perpendicular to the tool’s magnetic field direction, that is width for MFL-A and length for
MFL-C. This limitation is not a problem for isolated anomalies, however, it is especially critical
for this type of signal, which makes it barely impossible to distinguish the limits of adjacent
anomalies, leading to multiple feasible interpretations for this scenario.
A total of 25 girth welds with data patterns similar to the leakage were found during the
data comparison routine. The MFL-A sizing algorithm was correctly used in all indications,
however, the reported anomaly dimensions suggested that it was unlikely that these indications
could have caused a leak to the pipeline.

3.5. RCA Conclusion

To date, no single pipeline inspection technology is capable of detecting and sizing all
possible kinds of anomalies and installations found in a pipeline system by itself.
Considering that some dimension classes are out of MFL-A detection capabilities, and
also that complex anomalies, most likely multidirectional, were found on this line, it was
concluded that MFL-A technology alone was not able to properly size the depth of some of the
anomalies existing on this line.

4. The Solution

ROSEN’s proposal was to re-inspect this line with MFL-C and IEC, combining the
results with the already acquired MFL-A data, creating a unique combined final report.

Figure 7. IEC, MFL-C and MFL-A Tools

The goal of combined reporting with multiple technologies is to ensure that all threats
of interest affecting a pipeline, are properly detected and reported.

4.1. MFL-C Inspection

5
 Rio Pipeline Conference and Exhibition 2019

The MFL-C tool has the same inspection principle as MFL-A, however, this tool is
mechanically arranged to induce the magnetic field in the circumferential direction. Its flux is
easily disrupted for round and axial orientated indications (indications B and C), but not easily
disrupted by wide circumferential defect (indication A).

Figure 8. MFL-A Magnetic Field Orientation

The scope of this technology is to detect axially oriented, volumetric wall loss
anomalies, like general metal loss, pitting, axial grooving, and axial slotting anomalies.
These characteristics naturally make MFL-A and MFL-C technologies complementary
to each other, as the combined evaluation would cover a good part of the types of anomalies
that may affect a pipeline.

4.2. IEC Inspection

Figure 9. IEC Caliper Arm and Sensor Carrier

As stated by Dr. Schroeer K. (2018), “following a well-proven high-resolution caliper

tool design, the internal eddy current (IEC) sensor carriers are mounted on an ILI tool by
spring-loaded arms for smooth guidance along the pipe’s inner surface. The EC measurement
method is essentially contactless. Mechanical displacements of the sensors in radial direction
due to geometry changes of the line are additionally monitored by angle measurements of the
individual suspension arms.
The metal-loss features are measured in absolute terms, which enables the tool to
function regardless of the wall thickness. Pits with diameters of at least 10 mm and minimum
depths of 1.0 mm can be detected. Internal shallow defects with a maximum depth of 10 mm are
sized with high accuracy, i.e. +/- 1.3mm”.
Due to using two (or more in case of multi-diameter lines) sensor planes, a 100%
coverage can be guaranteed.
The use of the mechanical caliper arm also helps identifying areas of general thinning

6
 Rio Pipeline Conference and Exhibition 2019

by monitoring the inner diameter in parallel to the Eddy Current measurements.

Due to the sensor combination the IEC technology can detect and size internal metal
anomalies and geometry anomalies/ID reductions at the same time.”
Unlike the MFLs, IEC is able to properly distinguish the area and depth of adjacent
anomalies. In addition, it detects internal metal loss anomalies regardless of orientation. IEC
data provides depth information and allows for a better interpretation of the MFL datasets.
The scope of the IEC technology in this specific case was to detect and size internal
volumetric metal loss anomalies. However, this tool is also able to detect and size geometric
anomalies and ID reductions.

4.3. Combined Inspection Scope

Considering the advantages of each tool run on this pipeline, it was expected that with
the combined evaluation it would be possible to detect and size internal corrosion, pitting
corrosion, closely adjacent features and also properly discriminate anomalies detected in any
direction.

5. 2019 Inspections and Results

ROSEN’s strategy was accepted by the contractor and two more ILI inspections were
carried out on this pipeline. IEC and MFL-C were mobilized in the second week of February
2019, being run on the week that followed. Both datasets have arrived to ROSEN’s office on
the same week of the run.

Figure 10. MFL-C Tool After Run

5.1. IEC and MFL-C Quality

ROSEN quality checks have shown that, in general, both inspection conditions did not
differ from the conditions as they were foreseen, except by minor sensor losses. Data quality
was good for evaluation and standard specifications could be met.

5.2. Combined Evaluation Strategy

Merging the datasets collected during the inspection is not a matter of gathering
redundant information and detecting new anomalies, but to use the measurements of all
inspection technologies to reach for a better sizing and coverage of the pipeline.
With the information collected by all three inspections, it was possible to apply different
evaluation strategies to reach for a better coverage.
Figure 9 shows an axially oriented anomaly that was recorded by the three tools. MFL-
A recorded a low amplitude, therefore, it would underestimate the anomalies’ depth. MFL-C
recorded a clear high amplitude, while IEC recorded the real area of the anomaly. It was
possible to reach the final anomaly sizing by combining MFL-C amplitude with IEC’s precise
length and width.
7
 Rio Pipeline Conference and Exhibition 2019

Figure 11. MFL-A Data / MFL-C Data / IEC Data

In general, the applied methodology determined in which inspections the anomaly was
better detected in terms of depth.
Circumferentially-oriented anomalies have not generated high enough amplitudes in the
MFL-C data for it to generate a reliable sizing, and the same goes for axially-oriented anomalies
in the MFL-A dataset. For rounded anomalies, such as pittings, both tools were able to give
similar depths.
Apart from MFL, the IEC tool was able to better describe the geometry of the anomaly
and return a sizing similar to at least one of the MFL tools. This methodology unites the great
capability of IEC to provide the real shape of the anomaly and also removes the outliers from
the MFL inspection.
Due to its characteristics, the dimensions of anomalies detected by IEC were often
applied into the MFL datasets in order to support the MFL sizing algorithm resulting in more
realistic depth values.
Complex corrosion anomalies/clusters on the other hand were often reported by
different technologies at different points, leading to a better surface discrimination, as can be
observed in the multi-directional complex corrosion shown in Figure 8:

Figure 12. MFL-A Data / MFL-C Data / IEC Data

5.3. Statistics

The combined evaluation has increased the total amount of reported indications by
around 50%. The number of anomaly clusters also has increased in the same ratio:

Table 2. Amount of Reported Indications

Detected Anomalies MFL-A MFL-A/MFL-C/IEC
Corrosion 172,435 343,534
Clusters 25,191 50,998

8
 Rio Pipeline Conference and Exhibition 2019

The deepest anomaly reported in 2018 and the leakage were repaired between the
MFL-A and the MFL-C/IEC runs, therefore, it was not possible to determine their sizing with
the combined evaluation.
A very different scenario in terms of depth distribution is observed now; there was a
huge increase of severe anomalies reported:

Table 3. Reported Metal Loss Indications by Depth

Anomaly Depth MFL-A MFL-A/MFL-C/IEC
60-100% 1 95
40-59% 44 395
20-39% 14,992 10,930
10-19% 112,129 186,445
Total 127,166 197,865

A total of 17 anomalies with 85% wall loss were detected with the combined sizing
approach. All of these indications were previously reported with less than50% of wall loss in
the stand-alone MFL-A report, however, they are all locations likely to result in leakages.

In addition to these points, 78 other critical locations were found, which also require
attention, with depths between 60% and 84% of the pipe wall.

6. Conclusion

To date, no single pipeline inspection technology is capable of detecting and sizing all
possible kinds of anomalies and installations found in a pipeline system by itself. In case of
complex corrosion, it is recommended to combine different tools to increase sizing accuracies
of the reported anomalies. These different technologies can be attached to the same tool body
or the collected data can also come from different runs.
The above mentioned combined inspection strategy has proven to be successful on the
described pipeline. This strategy may also be applicable to other pipelines, however, it may not
be suitable for all. Each pipeline requires a tool combination to accommodate the conditions of
each individual pipeline. It is all a matter of defining what the characteristics of pipeline being
inspected are and what the aim of the inspection is in order to define the most appropriate
combination of technologies to use.

7. Acknowledgements

We thank all ROSEN Brazil team members for all the commitment and effort made
during this project. The same goes for ROSEN Germany GmbH IEC evaluation team.

8. References

ALMEIDA, R. – IBP2059_17-Pipeline Inspection Data Cross-check and Pig Inspection

Findings Comparison, 2017.
Dr. SCHROEER K., ROSEN Germany GmbH, Improved Eddy Current Technology
for Inline Inspection of Heavy Wall Gas Pipelines, PPSA Newsletter February 2018.
Specifications and requirements for in-line inspection of pipelines – POF, 2016.
9