Nuclear Engineering and Technology

Korean Nuclear Society (한국원자력학회)
권호 표지
DOI QR코드

DOI QR Code

Pipe thinning model development for direct current potential drop data with machine learning approach

  • Ryu, Kyungha (Department of Reliability Assessment, Mechanical System Safety Research Division, Korea Institute of Machinery and Materials) ;
  • Lee, Taehyun (Department of Reliability Assessment, Mechanical System Safety Research Division, Korea Institute of Machinery and Materials) ;
  • Baek, Dong-cheon (Department of Reliability Assessment, Mechanical System Safety Research Division, Korea Institute of Machinery and Materials) ;
  • Park, Jong-won (Department of Reliability Assessment, Mechanical System Safety Research Division, Korea Institute of Machinery and Materials)
  • Received : 2019.02.07
  • Accepted : 2019.10.02
  • Published : 2020.04.25
Download PDF
⟨ Previous Next ⟩

Abstract

The accelerated corrosion by Flow Accelerated Corrosion (FAC) has caused unexpected rupture of piping, hindering the safety of nuclear power plants (NPPs) and sometimes causing personal injury. For the safety, it may be necessary to select some pipes in terms of condition monitoring and to measure the change in thickness of pipes in real time. Direct current potential drop (DCPD) method has advantages in on-line monitoring of pipe wall thinning. However, it has a disadvantage in that it is difficult to quantify thinning due to various thinning shapes and thus there is a limitation in application. The machine learning approach has advantages in that it can be easily applied because the machine can learn the signals of various thinning shapes and can identify the thinning using these. In this paper, finite element analysis (FEA) was performed by applying direct current to a carbon steel pipe and measuring the potential drop. The fundamental machine learning was carried out and the piping thinning model was developed. In this process, the features of DCPD to thinning were proposed.

Keywords

References

  1. Electric Power Research Institute, Flow-Accelerated Corrosion in Power Plants: Technical Report (TR-106611), EPRI, 1996.
  2. Organization for Economic Cooperation and Development, Flow Accelerated Corrosion (FAC) of Carbon Steel & Low Alloy Steel Piping in Commercial Nuclear Power Plants: Topical Report (NEA/CSNI/R(2014)6), OECD-NEA, 2015.
  3. S.H. Lee, Y.S. Lee, S.K. Park, J.G. Lee, Corros. Sci. Technol. 14 (2015) 1. https://doi.org/10.14773/cst.2015.14.1.1
  4. H.J. Gwon, H.K. Ahn, C.H. Song, B.G. Park, J. Korea Acad. Ind. Coop. Soc. 12 (2011) 6.
  5. KAERI, Development of Carbon Steel with Superior Resistance to Wall Thinning and Fracture for Nuclear Piping System: Final Report (RR-3181), Korea Atomic Energy Research Institute, 2009.
  6. S.W. Han, J.S. Seo, J.H. Park, J. Korean Inst. Gas 21 (2017) 1. https://doi.org/10.7842/kigas.2017.21.1.1
  7. K.H. Ryu, I.S. Hwang, N.Y. Lee, Y.J. Oh, J.H. Kim, J.H. Park, C.H. Sohn, Nucl. Eng. Des. 238 (2008).
  8. K.H. Ryu, I.S. Hwang, J.H. Kim, Nucl. Eng. Des. (2013) 265.
  9. K.H. Ryu, T.H. Lee, J.H. Kim, I.S. Hwang, N.Y. Lee, J.H. Kim, J.H. Park, C.H. Sohn, Nucl. Eng. Des. (2010) 240.
  10. K.H. Ryu, I.S. Hwang, J.H. Kim, J. Korean Soc. Nondestruct. Test. 28 (2008) 2.
  11. K.H. Ryu, Development of Piping Wall Thinning Screening Technique Based on Equipotential Switching Direct Current Potential Drop Method, Ph.D. Thesis, Seoul National University, 2010.

Cited by

  1. A multi-layer approach to DN 50 electric valve fault diagnosis using shallow-deep intelligent models vol.53, pp.1, 2020, https://doi.org/10.1016/j.net.2020.07.001
  2. Multisensor Inspection of Laser-Brazed Joints in the Automotive Industry vol.21, pp.21, 2020, https://doi.org/10.3390/s21217335
  3. Non-intrusive Internal Corrosion Characterization using the Potential Drop Technique for Electrical Mapping and Machine Learning vol.33, pp.1, 2020, https://doi.org/10.1007/s40313-021-00823-9