Data Fusion
New level of ILI metal loss integrity assessments based on 3-dimensional corrosion profiles
Data Fusion integrates information from MFL-A and MFL-C tools using a neural network to comprehensively characterize all types of corrosion features. By utilizing both axial and circumferential magnetic field directions, Data Fusion harnesses complementary information from both to produce a detailed and accurate profile of pipeline corrosion. This approach ensures that all feature classes are properly identified and assessed, addressing the complexities inherent in pipeline integrity management.
Key advantages
More accurate burst pressure calculations
Comprehensive feature assessments across an entire line
Detailed 3D Anomaly Mapping
Mitigation of technology limitations
Through Data Fusion of axial and circumferential MFL signals, we turn raw data into high-resolution 3D profiles, replacing subjectivity with precision. This innovation enhances pipeline assessments, reducing costs and improving safety.
Best Paper Award for our innovation at IPC 2024
In a field of 345 technical papers, our paper on "Data Fusion of Complementary Axial and Circumferential Magnetic Flux Leakage In-line Inspections and Effects on Safe Remaining Life" was recognized with the Best Paper Award at IPC 2024.
The selection was based on originality, engineering significance, completeness, acknowledgment of work by others, organization, clarity, and graphic quality.
Missed the presentation? Join our experts at RIO Pipeline next year.
Traditionally, integrating findings from MFL tools A and C has been a manual task, requiring highly skilled analysts proficient in both MFL-A and MFL-C technologies. These experts must grasp the strengths and limitations of each inspection technology to extract the most relevant information and accurately characterize and size features. Such analyses only result in a basic 'box' representation of corrosion dimensions – length, width and depth.
In contrast, our proposed Data Fusion method involves combining the magnetic signals from MFL-A and MFL-C tools into a convolutional neural network (CNN), generating a comprehensive 3D depth profile.
This approach eliminates the need for the interpretation and manual merging of individual MFL signals, providing a complete 3D morphology description of the features and subsequently more accurate burst pressure calculations.