Development of an AMR Eddy Current-Based Crack Detection Sensor for Live Inspection of Unpiggable Transmission Natural Gas Pipelines

The Explorer family of robotic pipeline inspection platforms have proven their ability to provide valuable access to and data on unpiggable pipeline segments. This currently includes high resolution Magnetic Flux Leakage (MFL) data for metal loss detection, mechanical damage data as well as internal video images. As a result of recent pipeline incidents though, there is a growing desire for the addition of crack data on pipelines. While crack data can be obtained for piggable pipelines, at the present time is no methodology to provide that data on unpiggable pipeline segments. As such, the feasibility and prototyping of an eddy current (EC) sensor-based method of crack detection using Anisotropic MagnetoResistive (AMR) sensors was performed by Radiation Monitoring Devices (RMD) of Watertown, MA. RMD's work (carried out in a previous phase that is not part of this work) concluded that an AMR EC sensor could be implemented aboard a Pipetel Explorer 8in robot to potentially detect axially aligned cracks in unpiggable pipelines. The objective of this project was to implement this validated sensor aboard the 8in Explorer platform and replicate test results from the earlier phase in a real pipeline test setting. This involved the conceptual development, design, manufacturing, system integration and testing of a full circumference AMR EC sensor in a module format that is compatible with Explorer's 8in modular robot. In addition, data analysis capability was developed. InvoDane Engineering (IE) of Toronto, Canada, the developer of the Explorer robots, and RMD cooperated to determine a mechanical arrangement that would allow the sensor suite to be mounted to the 8in Explorer robot. This resulted in 15 sensor units with 8 channels, each arrayed in a collapsible assembly that can provide both full coverage of the pipe wall during scanning as well as maneuverability matching the rest of the robot when collapsed. This allows the robot/sensor assembly to negotiate practically any feature/obstacle in a pipeline, except a plug valve. InvoDane developed the signal processing hardware and software to collect the eddy current data and validated it against laboratory testing.


  • English

Media Info

  • Media Type: Digital/other
  • Edition: Final Report
  • Features: Appendices; Figures; Photos; Tables;
  • Pagination: 59p

Subject/Index Terms

Filing Info

  • Accession Number: 01717770
  • Record Type: Publication
  • Contract Numbers: DTPH56-15-T-00013
  • Created Date: Sep 23 2019 6:16PM