One corrosion feature can produce different valid depth interpretations. When those differences affect failure pressure calculations or intervention decisions, ROSEN's MFL Data Fusion combines complementary MFL-A and MFL-C signals into one continuous corrosion depth profile.
Assessment challenge
Axial and circumferential MFL technologies respond differently to corrosion morphology and orientation. Both responses may be technically valid, yet the resulting interpretation of depths can differ. This occurs because MFL-A and MFL-C are sensitive to different aspects of corrosion geometry.
In many cases, these differences do not materially affect the assessment. In complex corrosion cases where operation of the pipeline is critical, however, the reported depth can influence failure pressure estimations, feature prioritization, and intervention planning.
This creates a practical problem for decision-making:
- Which depth interpretation should be used?
- How should interacting corrosion be represented?
- Would another interpretation change the assessment result?
- Is the available input sufficiently consistent to support the decision?
ROSEN’s MFL Data Fusion addresses these challenges by combining complementary signal information before separate interpretations need to be reconciled. The result is one reconstructed 3D corrosion depth profile for further engineering assessments.
When to consider MFL Data Fusion
ROSEN’s MFL Data Fusion is a distinct service intended for complex corrosion for critical assets where sizing variability influences integrity decisions.
Consider MFL Data Fusion when:
Complementary MFL datasets provide different valid depth results for the same corrosion feature.
Axial, circumferential, or irregular corrosion profiles produce different signal responses depending on magnetization direction.
A change in the reported corrosion depth materially affects the estimated failure pressures or available safety margin.
Repeated inspections or separate evaluations produce inconsistent depth measurements that require additional technical reviews.
Excavation, repair, monitoring or reassessment priorities depend on which interpretation is selected.
ROSEN’s MFL Data Fusion is not intended as a standard upgrade for every MFL inspection. Its value is greatest where sizing variability has a meaningful effect on the assessment outcome.
Discuss whether MFL Data Fusion is relevant to your case.
How it works
Two Signals. One Profile.
MFL-A and MFL-C provide complementary information about corrosion. Each technology responds differently to feature morphology and orientation. Conventional assessment evaluates these responses separately.
ROSEN's MFL Data Fusion takes a different approach: it combines the complementary information at the magnetic signal level.
Machine learning supports the reconstruction of the corrosion profile. The resulting corrosion representation can be used as inputs for failure pressure estimations and efficient integrity decision-making.
ROSEN’s MFL Data Fusion uses an internally developed machine-learning model to reconstruct corrosion profiles from complementary MFL signals.
The model has been developed using relationships between MFL data and field laser scans. This enables the method to recognize how different signal responses relate to variations in actual corrosion depth and morphology.
The model does not:
- make integrity decisions
- replace engineering interpretation
- automatically determine repair requirements
- remove the need for technical validation
- apply equally to every inspection or corrosion case
The reconstructed profile is reviewed as part of ROSEN’s asset integrity services. The operator and responsible integrity expert retain ownership of the final integrity decision.
Why corrosion profiles matter
Corrosion does not form as a simple box. It is often represented using a limited set of dimensions such as length, width, and maximum depth. This simplified representation supports established and simplified assessment processes, but it does not always reflect how depth varies across a corrosion area.
This distinction becomes important when dealing with:
- complex corrosion morphology
- interacting features
- narrow and deep corrosion
- broad areas with variable depth
- corrosion with axial and circumferential components
- features close to an assessment or intervention threshold
A fused corrosion depth profile provides more information about how depth is distributed across the feature.
This can give the assessment a more representative input than selecting one depth value from several valid interpretations.
Conventional Evaluation vs. MFL Data Fusion
Conventional evaluation |
| MFL Data Fusion |
| Separate evaluation of MFL-A and MFL-C | → | Signal-level fusion of complementary data |
| Multiple valid depth interpretations | → | One continuous 3D corrosion profile |
| Interpretation reconciliation required | → | Unified corrosion representation |
| Assessment depends on selected interpretation | → | Reduced dependence on technology selection |
The outcome is case-specific. MFL Data Fusion can confirm an existing assessment, provide greater confidence in the assessment outcome, or identify situations where the selected corrosion representation materially affects the integrity decision.
Value for operators
Continuous corrosion profile
Reduced dependence on ILI technology selection
More stable failure pressure estimations
Higher confidence in intervention decisions
Integration into existing workflows
Every corrosion assessment is different. Our specialists can help determine whether MFL Data Fusion can add value to yours.
Gain deeper insights into our expertise
Recognition
Our paper on "Data Fusion of Complementary Axial and Circumferential Magnetic Flux Leakage In-line Inspections and Effects on Safe Remaining Life" received the Best Paper Award at the 2024 International Pipeline Conference.
The award recognized work on combining complementary axial and circumferential MFL inspections to improve corrosion assessment.
Discuss your case with our
MFL Data Fusion expert
Is corrosion depth variability affecting your assessment?
If different valid MFL interpretations are producing different failure pressure estimates or intervention priorities, MFL Data Fusion may provide a more consistent basis for the assessment.
Share available information, the corrosion concern, or the decision you need to support. A ROSEN specialist will review whether MFL Data Fusion is relevant to the case.
FAQs
No. MFL Data Fusion is excellent for complex corrosion situations. It supports cases where differences in corrosion depth sizing affect the assessment or integrity decisions. Conventional evaluation remains appropriate for many corrosion morphologies and inspection objectives.
Yes. MFL Data Fusion is based on complementary information from axial and circumferential MFL signals. Suitability depends on the availability and quality of the relevant datasets.
Potentially, yes. Where suitable MFL-A and MFL-C datasets already exist, they may provide the basis for a Data Fusion application. Data availability does not automatically mean the case is suitable, so the datasets and assessment objective must be reviewed first.
No. Machine learning supports reconstruction of the corrosion profile. The output is technically reviewed, and integrity decisions remain with the responsible engineering and operator teams.
Comparing inspection reports starts with separately interpreting results and then attempting to reconcile differences. MFL Data Fusion combines complementary information at the signal level before reconstructing one corrosion depth profile.
No. Field verification remains an important tool within pipeline integrity management where direct confirmation of corrosion conditions is required.
MFL Data Fusion is designed to improve the corrosion representation used for assessment by reconstructing a continuous 3D corrosion profile from complementary MFL datasets. In some applications, this can provide greater confidence in assessment inputs and decision-making before field activities are considered.
Yes. The reconstructed corrosion profile provides input for further engineering assessment rather than replacing the wider integrity management process.
Useful information includes:
- The type of corrosion threats of concern
- The asset and operating context
- Targeted reinspection dates
- Presence of past MFL-A and/or MFL-C data
- Existing assessment results, if available