Rail crack size measurement based on magneto-optical imaging in multi-layer lift-off

Abstract

In the current method of crack size measurement using magnetic flux leakage detection technology, it is difficult to quantitatively analyze the detected magnetic flux leakage data due to the unstable character of the sensor lift-off position. To solve this problem, based on the forward inference of the leakage magnetic field of crack, a method of magnetic field acquisition in multi-layer lift-Off is proposed, and a magneto-optical imaging detection system is designed and built, so as to achieve high-precision measurement of rail crack parameters. Firstly, in the forward analysis, by borrowing the classical magnetic dipole theory, the relationship between the crack parameters and the peak value of the magnetic leakage field as well as its position is analyzed. Secondly, a magneto-optical image acquisition system with a lift-off interval of 10 $\mu m$ is designed, which can effectively remove the influence of the light source system and magnetic domains on the imaging quality. Finally, a rail with standard processed 0.4 mm wide cracks is selected as experimental objects for crack measurement experiments. The crack width is obtained by fitting the peak gray value position curve under different lift-off, and the measurement accuracy reaches 25 $\mu m$, and the relative measurement error is found to be less than 7.5% in the comparison of several data sets. At the same time, in many groups of different crack depth measurement experiments, it is found that the measurement accuracy of crack depth measurement using the peak gray value is 0.5 mm.