MFL Testing for Subsea Pipeline Corrosion — PatSnap Eureka
Magnetic Flux Leakage Testing for Sub-Surface Corrosion in Subsea Pipelines
MFL inspection is the most mature and widely patented method for direct sub-surface defect characterisation in ferromagnetic steel pipelines. Explore how it works, who leads the IP, and how machine learning is reshaping signal inversion.
How Magnetic Flux Leakage Detects Sub-Surface Corrosion
Magnetic flux leakage testing operates by saturating a ferromagnetic pipe wall with a strong magnetic field. At locations where corrosion has reduced wall thickness, the pipe's reluctance increases locally, causing magnetic flux to leak from the surface of the pipe. Hall-effect sensors or magnetometer arrays positioned close to the pipe wall capture this leakage signal, which is then processed to characterize the size, depth, and spatial extent of the defect.
Because the MFL signal is produced by volumetric wall loss rather than surface geometry alone, the technique is inherently sensitive to sub-surface metal loss — distinguishing it from purely optical or surface-contact methods. This makes MFL the primary in-line NDE method for sub-surface corrosion in steel subsea pipelines, as established by TDW Delaware's System and Method for Analyzing Anomalies in a Conduit (2021).
Multi-orientation sensing is particularly important because sub-surface corrosion in pipeline seams may present anisotropic flux perturbations. The TDW Delaware patent explicitly describes a probe that "detects magnetic flux leakage in at least two orientations," with anomaly identification and corrosion severity assessment based on the MFL signal combined with the depth of detected anomalies.
A four-step internal corrosion direct assessment protocol for CO2-bearing submarine pipelines — described by the School of Petroleum Engineering (2021) — situates MFL as one element within a broader integrity management programme. The study found that indirect evaluation results aligned consistently with direct measurement data, validating the predictive corrosion rate model used. For broader context on pipeline integrity standards, the DNV recommended practice framework provides industry-recognised benchmarks for in-line inspection tool selection.
Machine Learning Transforms MFL Signal Inversion
Recent patent activity reflects a decisive shift from empirical look-up tables toward machine learning-based inversion, enabling more accurate characterisation of irregular sub-surface corrosion geometries.
MFL Data → Image → ML Model for Corrosion Parameters
Halliburton's active 2024 patents describe a system that converts MFL data obtained from an in-pipe tool into image data, then submits that image data to a machine learning model trained to identify physical parameters associated with corrosion — including likely depth and areal extent of wall loss. The parallel WO filing confirms applicability in both subsea well and pipeline contexts, with the tool placed in a pipe within a wellbore.
Active US & WO patents, 2024GA-KELM Model for Stress Corrosion Defect Sizing
Research from Harbin University of Science and Technology proposes combining MFL with a genetic algorithm-optimised kernel extreme learning machine (GA-KELM). The approach first simulates defect dimension evolution over time and MFL signal distribution under various defect conditions, then uses the trained model to predict corrosion defect depth and length — enabling non-destructive characterisation of defect geometry that would otherwise require invasive inspection. The methodology directly transfers to the subsea context, where similar pipe steel grades and corrosion mechanisms are encountered.
Peer-reviewed, 2023Empirical Look-Up Tables and Analytical Approximations
Earlier MFL inspection relied on empirical look-up tables or analytical approximations to relate peak MFL amplitudes to defect depth and length. Earlier patents (pre-2019) focused on sensor geometry and probe orientation optimisation rather than learned inversion. This classical approach struggled with irregular corrosion geometries common in subsea seam welds and CO₂-exposed internal surfaces.
Superseded by ML approachesProgressive Hybridisation: MFL + Digital Signal Processing
Post-2022 filings and papers uniformly incorporate some form of learned inversion model. This shift implies that future MFL inspection tools for subsea pipelines will be increasingly autonomous in their defect characterisation capability. The PatSnap Analytics platform tracks this IP trend in real time across global patent offices. For academic context, NDT.net archives the peer-reviewed literature underpinning these advances.
Dominant post-2022 directionNDE Method Capabilities and IP Activity at a Glance
Visualising the relative sub-surface corrosion detection capability of the five dominant NDE approaches, and the publication timeline of key patent and literature milestones.
Sub-Surface Corrosion Detection Capability by NDE Method
Relative capability scores (out of 10) for five NDE methods in detecting sub-surface corrosion in subsea steel pipelines, synthesised from patent and literature evidence in the PatSnap Eureka dataset.
Key MFL & Subsea Corrosion Patent/Literature Milestones by Year
Timeline of significant patent grants and peer-reviewed publications in the MFL subsea pipeline inspection space, showing accelerating AI integration from 2022 onward.
NDE Approaches for Sub-Surface Corrosion Beyond MFL
When MFL deployment is impractical — non-piggable pipelines, non-ferromagnetic materials, or external ROV inspection — these complementary methods fill the gap.
| NDE Method | Lead Organisation | Key Capability | Limitation Addressed | Year |
|---|---|---|---|---|
| Gradient-Field Pulsed Eddy Current (GPEC) | Xi'an Jiaotong University | Quantitative sub-surface corrosion imaging; 3D analytical model; uniform field excitation resolves false signal contributions from pancake coils | Non-piggable lines; non-contact external inspection | 2017 |
| Transient Electromagnetic (TEM) | Yunnan Electric Power Research Institute | Induced EMF attenuation rate correlates monotonically with pipe wall thickness; indirect quantification of corrosion degree via ANSYS simulation | Remote sensing where contact sensors cannot be deployed | 2020 |
| Acoustic Emission (AE) & Ultrasonic | Universiti Kebangsaan Malaysia; Hihakai Kensa K.K. | Critical complementary sensor modality; surface-wave probes discriminate inner- from outer-surface corrosion via transmission attenuation | Inaccessible pipeline structures; under-coating corrosion | 2000, 2022 |
| Cathodic Protection (CP) Dual-Probe | CESCOR S.R.L. | Simultaneous potential gradient + seabed resistivity measurement; spatially resolved CP effectiveness along buried subsea pipelines | External corrosion risk zone identification before direct inspection | 2018 |
| Electrochemical Multi-Electrode Array | Deakin University | Detects and visualises localised corrosion including under disbonded coatings and stray-current corrosion; sensitivity of 10 µm/year | Corrosion under coatings; stray-current environments | 2021 |
Find patents for every NDE method in one search
PatSnap Eureka covers GPEC, TEM, AE, CP probes, and MFL in a single dataset with 50+ records.
Key Players and Innovation Trends in MFL Inspection
A clear trend across the dataset is the progressive hybridisation of MFL with digital signal processing and machine learning. Earlier patents focus on sensor geometry; post-2022 filings uniformly incorporate learned inversion models.
Halliburton Energy Services — Most Advanced ML Integration
Holds two active patents on MFL-based corrosion analysis with machine learning (US 2024 and WO 2024), plus an active EP patent on electromagnetic corrosion detection for multi-pipe configurations (2020). Halliburton's portfolio represents the most advanced integration of MFL with AI-driven analysis currently documented in the dataset.
TDW Delaware / T.D. Williamson / KPL South Texas — Largest MFL Patent Cluster
Collectively constitute the largest cluster of MFL-specific patent filings, with at least five records addressing multi-orientation MFL detection and selective remediation of seam weld corrosion in oil and gas conduits. Filings span US, CA, WO, and divisional formats, reflecting significant commercial interest in protecting this inspection and remediation workflow.
Internal CO₂ Corrosion and the Case for Multi-Modal Integration
Internal CO₂ corrosion in subsea multiphase pipelines presents unique challenges. Localised corrosion rates vary with flow velocity, partial pressure, and temperature — factors that must inform MFL tool selection and inspection intervals, as studied experimentally by Southwest Petroleum University (2015). The PatSnap chemicals and materials intelligence platform maps the corrosion mechanism literature alongside patent filings for a complete picture.
Multi-modal NDE integration is increasingly recognised as essential for reliable sub-surface corrosion characterisation. Acoustic emission, ultrasonic, electromagnetic, and electrochemical methods each capture complementary aspects of the corrosion state, as reviewed by Universiti Kebangsaan Malaysia (2022). Inaccessibility due to pipeline structure is identified as a primary challenge for all NDE methods.
Seabed pipeline integrity encompasses more than corrosion alone. Threats including anchor penetration — studied by Dalian Maritime University (2019) — and tsunami scouring — studied by Tokyo Gas Co. (2019) — must be considered in the design of comprehensive subsea inspection programmes, as physical damage can accelerate corrosion initiation in exposed pipe sections. The International Maritime Organization and IOGP publish guidelines that frame these multi-hazard inspection requirements for operators. For regulatory context, PatSnap's trust centre outlines how IP data is handled in compliance with enterprise security standards.
Magnetic Flux Leakage & Subsea Pipeline Inspection — key questions answered
MFL testing operates by saturating a ferromagnetic pipe wall with a strong magnetic field. At locations where corrosion has reduced wall thickness, the pipe's reluctance increases locally, causing magnetic flux to leak from the surface of the pipe. Hall-effect sensors or magnetometer arrays positioned close to the pipe wall capture this leakage signal, which is then processed to characterize the size, depth, and spatial extent of the defect. Because the MFL signal is produced by volumetric wall loss rather than surface geometry alone, the technique is inherently sensitive to sub-surface metal loss.
Signal inversion — converting raw MFL waveform data into quantitative defect dimensions — has historically been the dominant technical challenge in MFL-based inspection. Classically, empirical look-up tables or analytical approximations were used to relate peak MFL amplitudes to defect depth and length. Recent patent activity reflects a decisive shift toward machine learning-based inversion, enabling more accurate characterization of irregular sub-surface corrosion geometries.
TDW Delaware / T.D. Williamson / KPL South Texas, LLC collectively constitute the largest cluster of MFL-specific patent filings, with at least five records addressing multi-orientation MFL detection and selective remediation of seam weld corrosion in oil and gas conduits. Halliburton Energy Services holds two active patents on MFL-based corrosion analysis with machine learning (US 2024 and WO 2024), representing the most advanced integration of MFL with AI-driven analysis currently documented in the dataset.
Complementary NDE approaches documented alongside MFL include: gradient-field pulsed eddy current (GPEC) probes for quantitative sub-surface corrosion imaging in non-piggable lines; transient electromagnetic (TEM) methods for remote sensing where contact sensors cannot be deployed; acoustic emission and ultrasonic techniques; and cathodic protection and electrochemical probes for external corrosion risk assessment. Multi-modal sensor fusion is increasingly required for reliable corrosion detection under coatings.
Halliburton Energy Services has filed patents describing a system that converts MFL data obtained from an in-pipe tool into image data, then submits that image data to a machine learning model trained to identify physical parameters associated with corrosion — including likely depth and areal extent of wall loss. Research from Harbin University of Science and Technology proposes combining MFL with a genetic algorithm-optimised kernel extreme learning machine (GA-KELM) to predict corrosion defect depth and length, enabling non-destructive characterisation of defect geometry that would otherwise require invasive inspection.
Research from Deakin University confirms that electrochemical multi-electrode array probes can detect and visualise localised corrosion, including corrosion under disbonded coatings and stray-current corrosion, with sensitivity in the order of 10 µm/year.
Still have questions? Let PatSnap Eureka answer them for you.
Ask PatSnap Eureka About MFL InspectionAccelerate Your Subsea Pipeline NDE Research with AI-Powered Patent Intelligence
Join 18,000+ innovators already using PatSnap Eureka to accelerate their R&D — search 50+ MFL and subsea corrosion inspection records instantly.
References
- System and Method for Analyzing Anomalies in a Conduit — TDW Delaware, Inc., 2021
- System and Method for Detecting and Remediating Selective Seam Weld Corrosion in a Conduit — T.D. Williamson, 2019
- System and Method for Detecting and Remediating Selective Seam Weld Corrosion in a Conduit (WO) — T.D. Williamson, 2019
- System and Method for Detecting and Remediating Selective SEAM Weld Corrosion in a Conduit (KPL, CA) — KPL South Texas, LLC, 2019
- System and Method for Detecting and Remediating Selective SEAM Weld Corrosion in a Conduit (KPL, WO) — KPL South Texas, LLC, 2019
- System and Method for Detecting and Remediating Selective SEAM Weld Corrosion in a Conduit (KPL, US) — KPL South Texas, LLC, 2019
- Corrosion Analysis Using Magnetic Flux Leakage Measurements (US) — Halliburton Energy Services, Inc., 2024
- Corrosion Analysis Using Magnetic Flux Leakage Measurements (WO) — Halliburton Energy Services, Inc., 2024
- Magnetic Flux Leakage Testing Method for Pipelines with Stress Corrosion Defects Based on Improved Kernel Extreme Learning Machine — Harbin University of Science and Technology, 2023
- Imaging of Subsurface Corrosion Using Gradient-Field Pulsed Eddy Current Probes with Uniform Field Excitation — Xi'an Jiaotong University, 2017
- New Probes and Devices for Cathodic Protection Inspection of Subsea Pipelines — CESCOR S.R.L., 2018
- Field and Laboratory Assessment of Electrochemical Probes for Visualizing Localized Corrosion Under Buried Pipeline Conditions — Deakin University, 2021
- Recent Advances in Nondestructive Method and Assessment of Corrosion Undercoating in Carbon–Steel Pipelines — Universiti Kebangsaan Malaysia, 2022
- System and Method for Detecting Corrosion of a Pipe — Halliburton Energy Services, Inc., 2020
- Application of Internal Corrosion Direct Assessment in CO2 Slug Flow Submarine Pipelines — School of Petroleum Engineering, 2021
- An Experimental Study on the Internal Corrosion of a Subsea Multiphase Pipeline — Southwest Petroleum University, 2015
- ANSYS Simulation Model of Buried Metal Pipeline Corrosion Detection — Yunnan Electric Power Research Institute, 2020
- Method for Inspection of Outer-Surface Corrosion of Pipe — Hihakai Kensa K.K., 2000
- Development of Subsea Pipeline Buckling, Corrosion and Leakage Monitoring — Dalian Maritime University, 2023
- Buried Depth of a Submarine Pipeline Based on Anchor Penetration — Dalian Maritime University, 2019
- Evaluation of Tsunami Scouring on Subsea Pipelines — Tokyo Gas Co. Ltd., 2019
- DNV Recommended Practice — In-Line Inspection Tool Selection and Performance — DNV
- NDT.net — Peer-Reviewed NDE Literature Archive — NDT.net
- IOGP — Subsea Pipeline Integrity Management Guidelines — International Association of Oil & Gas Producers
- International Maritime Organization — Pipeline Safety Frameworks — IMO
All data and statistics on this page are sourced from the references above and from PatSnap's proprietary innovation intelligence platform.
PatSnap Eureka searches patents and research to answer instantly.