TY - JOUR
T1 - Magnetic resonance eddy current testing imaging and non-destructive evaluation of rebar in concrete
AU - Zhang, Jinming
AU - Liao, Leng
AU - Lan, Tian
AU - Chen, Haitao
AU - Zhang, Senhua
N1 - Publisher Copyright:
© 2025 Elsevier Ltd
PY - 2026/1/15
Y1 - 2026/1/15
N2 - This study, based on the principle of Magnetic Resonance Eddy Current Testing (MRECT), develops a small-diameter dual-coil resonant sensor for rebar detection in concrete. By increasing the aspect ratio of the magnetic core to optimize its geometry, high-resolution imaging of deeply embedded targets is achieved. Coupled with a multilayer perceptron (MLP) model, the system accurately predicts key geometric parameters. Experimental results demonstrate precise detection of rebars with diameters of 6–25 mm and embedded depths of 2–5 cm beneath a 5 cm concrete layer. The root mean square errors (RMSE) for diameter and depth prediction are 0.31 mm and 1.6 mm, respectively, with coefficients of determination (R2) reaching 0.99 and 0.98. Maximum relative errors are 7.68 % and 7.29 %, respectively. This approach provides an efficient and accurate solution for quantitative non-destructive testing of reinforcement in concrete structures, offering broad engineering prospects for construction quality assessment and structural health monitoring.
AB - This study, based on the principle of Magnetic Resonance Eddy Current Testing (MRECT), develops a small-diameter dual-coil resonant sensor for rebar detection in concrete. By increasing the aspect ratio of the magnetic core to optimize its geometry, high-resolution imaging of deeply embedded targets is achieved. Coupled with a multilayer perceptron (MLP) model, the system accurately predicts key geometric parameters. Experimental results demonstrate precise detection of rebars with diameters of 6–25 mm and embedded depths of 2–5 cm beneath a 5 cm concrete layer. The root mean square errors (RMSE) for diameter and depth prediction are 0.31 mm and 1.6 mm, respectively, with coefficients of determination (R2) reaching 0.99 and 0.98. Maximum relative errors are 7.68 % and 7.29 %, respectively. This approach provides an efficient and accurate solution for quantitative non-destructive testing of reinforcement in concrete structures, offering broad engineering prospects for construction quality assessment and structural health monitoring.
KW - Deep learning
KW - Magnetic coupled resonance
KW - Non-destructive testing
KW - Penetrating imaging
KW - Rebar diameter
KW - Rebar embedded depth
KW - Structural health monitoring
UR - https://www.scopus.com/pages/publications/105015149729
U2 - 10.1016/j.measurement.2025.118931
DO - 10.1016/j.measurement.2025.118931
M3 - Article
AN - SCOPUS:105015149729
SN - 0263-2241
VL - 257
JO - Measurement: Journal of the International Measurement Confederation
JF - Measurement: Journal of the International Measurement Confederation
M1 - 118931
ER -