Magnetic Barkhausen Noise Analysis of Stress and Texture in Steels

Abstract

Magnetic Barkhausen noise (MBN), the sudden changes in magnetisation as a ferromagnetic material is magnetised, is investigated in industrial steels important to New Zealand. Two fundamentally different ways of measuring MBN are investigated. Firstly, with a coil wrapped around the sample, the behaviour of MBN in a bolting steel under applied stress is found to be linear, saturating at high stresses. This is in line with behaviour published for many other steels. The coil MBN in a pressure vessel steel shows the same under compression and undergoes a Villari reversal in tension. Secondly, a portable probe, which measures MBN normal to the sample surface, is built and used to obtain a calibration of MBN in this steel under applied stress. The Villari reversal is not present in this case. This difference is explained physically in terms of the relationship between the magnetostriction and the measured magnetisation component. It is demonstrated that MBN can be used as a clear non-destructive test of textural anisotropy in steel that cannot be so easily determined by other methods. For two sheets of the same material welded together, the MBN as the applied field direction is rotated shows strong anisotropies that are mirror images of each other. This demonstrates that one sheet has been turned over before being welded. This history of the sample is not evident from optical observations, but is supported by X-ray diffraction (XRD) measurements of residual stress that show the same mirroring result. MBN results at the weld are found to be sensitive to residual stress, crystallographic texture, changes in microstructure, and probe lift-off. XRD measurements show large tensile residual stresses there. Optical and scanning electron microscopy reveal a finer microstructure at the weld. The effect of the slight air gaps due to the shape of the sample is found to be substantial. When preparing samples to eliminate this effect, care must be taken not to induce surface stress. MBN also shows remarkable sensitivity to the texture of the material, which is measured using electron backscattering diffraction (EBSD). As MBN is non-destructive and can probe deeper into the material than EBSD can, it has potential as a method of measuring texture. The sensitivity of MBN to all these phenomena means that it is not practical to use the applied stress calibration for direct residual stress measurements of the weld. However, in combination with XRD measurements of stress and EBSD measurements of texture, the calibration is an important tool in providing a complete description of the complex combination of effects that occur.