Eddy Current Nondestructive Evaluation (NDE)
Automated Eddy Current Pipeline CUI Inspection System
- Externally applied testing solution without insulation removal
- Imperfections on aluminum or stainless jacketing raise no false positive.
- High sensitivity with up to 5 inch of lift-off
- Machine learning based signal processing.
Alternative Current Potential Drop System
- Accurate characterization of material electrical conductivity and linear magnetic permeability.
- Coating thickness measurement and crack sizing.
- Ultra high precision resistance measurement (rms error <20 nΩ).
Machine Learning on Edge Devices
- STM32 solutions for artificial neural networks.
- NVIDIA® Jetson™ and Raspberry Pi ecosystem.
- Xilinx® Vitis™ AI for FPGA accelerated model inference.
Yuan Ji
Research Scientist II
Expertise
- Eddy current system design and testing.
- Electromagnetic pipeline inspection.
- Radio frequency, analog, digital, power and motor control circuit design.
- Embedded system and machine learning FPGA acceleration.
Keywords
- Eddy current NDE, ACPD NDE, Embedded System, FPGA, Machine Learning
Selected Publications & Patents
- John R Bowler, Theodoros P Theodoulidis, Hui Xie, and Yuan Ji, “Evaluation of eddy-current probe signals due to cracks in fastener holes.” IEEE Transactions on Magnetics, vol. 48, no. 3, pp. 1159-1170, 2011
- US Patent 10816508 “Planar Array Pipeline Inspection Tool” July 18, 2019
- US Patent 10031108 “Multi-Frequency Eddy Current Pipeline Inspection Apparatus and Method” July 24, 2018