Australia’s oil and gas sector is facing a pivotal moment.
Ageing pipeline infrastructure, evolving regulation, and increasing replacement costs are creating unprecedented pressure to maximise asset lifespans while upholding the highest safety and compliance standards.
Magnetic flux leakage (MFL) inspection is the industry’s most trusted method for detecting metal loss in pipelines. Its two primary models – MFL-A (axial) and MFL-C (circumferential) – use differently oriented magnetic fields to detect anomalies aligned in different directions. Each offers distinct advantages. But both rely on a unidirectional magnetic signal, which limits their ability to fully characterise complex corrosion morphologies – such as pitting within general corrosion – where accurate sizing of the deepest points is essential.
As pipeline integrity teams are pushed to deliver less conservative and more precise assessments, these limitations matter. Excess uncertainty in inspection results can force operators into worst-case assumptions, driving unnecessary digs, inflating costs, and sometimes misallocating resources.
ROSEN has developed a new approach to support operators in overcoming these limitations. MFL Data Fusion merges the strengths of MFL-A and MFL-C into a single, highly detailed 3D depth map, giving operators a new level of precision in detecting, sizing, and characterising metal loss features – regardless of their shape, orientation, or complexity.
The severity of potential pipeline incidents – combined with higher throughput, commercial pressures, and a need to extend asset life – has made the usual tolerance for MFL inspection uncertainty marginally adequate.
Traditionally, when data from MFL-A and MFL-C runs were available, different combined reporting approaches were possible. While this aims to improve accuracy, it is labour-intensive, subjective, and prone to interpretation differences.
ROSEN’s MFL Data Fusion approach solves these challenges. By aligning and processing both datasets with advanced algorithms and a pre-trained convolutional neural network (CNN), the approach produces a single, objective output – one that maximises each tool’s strengths and eliminates much of the ambiguity.
MFL Data Fusion enhances integrity decision-making by avoiding unnecessary digs, identifying complex anomalies, reducing integrity management uncertainty, and optimising inspection planning.
The process begins by combining axial and circumferential magnetic field data to enhance feature characterisation across all POF anomaly classes.
Data alignment: MFL-A and MFL-C datasets are precisely matched.
Machine learning analysis: A CNN model processes the aligned data to identify and profile corrosion anomalies.
3D anomaly profiling: Output is delivered as a River Bottom Profile (RBP): a continuous, high-resolution 3D map of each defect’s geometry.
Advanced pressure modelling: From these depth maps, burst pressure and models like Psqr can be calculated with confidence approaching that of an in-ditch laser scan.
The result is UT-like performance in metal loss sizing – across all anomaly types – without the cost or disruption of excavation.
This brings numerous benefits to operators, including improved sizing accuracy, optimised dig programs, comprehensive feature assessment, 3D anomaly mapping, overcoming MFL limitations, better failure pressure calculations, and added value to historical data.
A practical, flexible service
ROSEN’s MFL Data Fusion delivers fused MFL-A and MFL-C data as high-resolution 3D depth maps, River Bottom Profiles (RBPs), and multiple failure pressure models for every anomaly or pipeline segment.
The two datasets can be collected in separate ILI runs – ideally with minimal time between to prevent corrosion morphology changes – or in some instances, for larger diameters (over 20 inches), even in a single run. No special ILI tool setup or inspection conditions are required, and previously collected data can be used. Existing laser (or AUT) field scan data may also be incorporated to fine-tune algorithms for a pipeline-specific model.
In an industry where every decision matters, MFL Data Fusion isn’t just a new offer – it’s a step change.
With MFL Data Fusion, operators no longer have to make critical decisions based on partial information or broad statistical assumptions. Instead, they gain precise, objective, and comprehensive insight into every corrosion anomaly.
ROSEN’s MFL Data Fusion service is already delivering validated, high-resolution 3D corrosion profiles and pressure models with engineering-grade accuracy. While the service is in use with selected operators, ongoing development focuses on expanding model adaptability and enhancing pipeline-specific calibration – reflecting ROSEN’s commitment to continuous innovation.
For more information, visit the website.
This feature also appears in the September edition of The Australian Pipeliner.