An Innovative Pipe Inspection Tool Utilizing Electromagnetic Resonance Coupling and Machine Learning

Mostafa T. M., Ooi G., ÖZAKIN M. B., Khater M., Zeghlache M. L., Bagci H., ...Daha Fazla

IEEE Transactions on Industrial Electronics, cilt.71, sa.5, ss.5338-5348, 2024 (SCI-Expanded) identifier

  • Yayın Türü: Makale / Tam Makale
  • Cilt numarası: 71 Sayı: 5
  • Basım Tarihi: 2024
  • Doi Numarası: 10.1109/tie.2023.3285970
  • Dergi Adı: IEEE Transactions on Industrial Electronics
  • Derginin Tarandığı İndeksler: Science Citation Index Expanded (SCI-EXPANDED), Scopus, Academic Search Premier, Aerospace Database, Applied Science & Technology Source, Business Source Elite, Business Source Premier, Communication Abstracts, Compendex, Computer & Applied Sciences, INSPEC, Metadex, Civil Engineering Abstracts
  • Sayfa Sayıları: ss.5338-5348
  • Anahtar Kelimeler: Eddy current, long short-term memory (LSTM), machine learning (ML), neural networks (NNs), pipeline inspection, resonance coupling, wellbore integrity
  • Yıldız Teknik Üniversitesi Adresli: Evet

Özet

—This article describes an advanced tool that uses electromagnetic resonance coupling and machine learning techniques to detect and characterize metal loss on the inner surface of a metallic pipe. The proposed tool uses a transmitter coil placed along the axis of the pipe and four sensor coils installed around the transmitter coil. Any defect on the pipe surface leads to changes in the impedance of the transmitter and sensor coils as well as in the mutual coupling between them, thus creating a detectable variation in the outputs of one or multiple sensor coils. An artificial neural network is developed to reconstruct two-dimensional pipe cross-sections and to completely characterize the defects using these variations. The proposed tool is tested and validated via simulations, and data are collected using an experimental prototype. Results show that the tool can fully characterize the size, location (azimuthal angle), and level (thickness) of metal loss.