Study on surface crack detection of ferromagnetic materials based on remanence

Abstract

Remanence detection is a technique of electromagnetic non-destructive testing (NDT). This paper studies a quantitative detection method for surface cracks of ferromagnetic materials based on remanence. The finite element analysis software COMSOL Multiphysics was used to establish the remanence detection model and the 'moving grid' function was used to realise the simulation of the remanence signal. The leakage magnetic field occurs due to the distortion of the magnetic induction lines near the surface cracks after ferromagnetic materials are magnetised. Remanence detection uses the leakage magnetic field to detect cracks. The relationship of the leakage magnetic field versus the crack depth and width was analysed using the magnetic dipole model. The relationship between the crack size and the remanence signal was verified by measuring the surface remanence signal of cracks of different sizes. The characteristic parameters related to the crack size were extracted and the regression model between the characteristic parameters and the crack size was established. For the remanence detection, the maximum error of width prediction was 16.25% and the maximum error of depth prediction was 18.48%. For the magnetic flux leakage (MFL) detection, the maximum error of width prediction was 12.1% and the maximum error of depth prediction was 12.32%. Under the same experimental conditions, the maximum error of crack width measurement of remanence detection was 4.15% larger than that of MFL detection and the maximum error of depth was 6.16% larger than that of MFL detection.