Quantitative assessment of back-surface flaws in cladded conductors via near- and remote-field integrated transient eddy current testing

• A novel design of the compact testing probe for integrated sensing is proposed. • The testing and sensing mechanisms are revealed via spectral and temporal analysis. • Fused signal feature for high-quality flaw image is extracted via a neural network. • Assessment accuracy of the sizing parameter for each back-surface flaw exceeds 90%. In a bid to non-intrusively characterise and evaluate back-surface flaws in cladded conductors, in this paper a novel design of the compact testing probe of Near- and Remote-field integrated Transient Eddy Current testing (NRiTEC) is proposed for fused visualisation-based flaw testing and assessment. A numerical model is established to analyse signal responses to back-surface corrosions in a cladded conductor and particularly the remote-field phenomena. It is implied from the simulations that the near- and remote-field sensing can be simultaneously realised with the designed probe, enabling the NRiTEC mechanism for retrieval of more comprehensive information of the back-surface flaw. With the testing system alongside the fabricated probe, the fused signal feature is extracted from the integrated testing signal using AlexNet convolutional neural network for producing the flaw image with the enhanced quality. An image-processing algorithm is proposed for corrosion-boundary identification and opening-area approximation. The corrosion depth is inversely approximated based on the established correlation of the maximum pixel variation with the flaw depth. Flaw-evaluation results indicate that the relative error of the estimated flaw sizing parameter is less than 10%. This implies the designed NRiTEC probe alongside the fused visualisation is capable of high-accuracy assessment of back-surface flaws in cladded conductors.