A novel broken wire evaluation method for bridge cable magnetic flux leakage testing under lift-off uncertainty

Magnetic flux leakage (MFL) is one of the most commonly used techniques for detecting broken wires of bridge cables. In practical testing, the sensor and magnetizer lift-off variations caused by mechanical vibration, sheath surface roughness and thickness nonuniformity significantly affect the accuracy of the estimated defect size. To address this problem, the theory of ratio of adjacent peaks (RAP) is developed based on the magnetic dipole model, which reveals the nonlinear relationship between RAP vector and broken-wire width. Based on the RAP theory, a radial sensor array (RSA) configuration and an evaluation method are further proposed in this paper. The RSA consists of four sensors along the bridge cable radial direction with a fixed interval and is employed to acquire RAP vector, which is identified as a spatial characteristic of MFL signal. An artificial neural network (ANN) algorithm is employed to develop the nonlinear defect width evaluation method based on the RAP vector. The effectiveness of the RSA configuration and the evaluation method is verified by finite element method (FEM). The results indicate that the broken-wire width can be accurately estimated utilizing RAP vector obtained by RSA without considering the lift-off variation.