The magnetic flux leakage (MFL) technique is extensively used for detection of flaws as well as for evaluation of their dimensions in ferromagnetic materials. However, proper analysis of the MFL signal is hindered by the MFL sensor velocity causing distortions of this signal. Traditionally measured components of the MFL signal are particularly sensitive to the scanning velocity. In this paper, an another signal – the gradient of the normal component of magnetic flux density – was proposed as it is less sensitive to the scanning velocity. Results obtained for scans of the steel plate with artificially manufactured flaws confirm this statement.
The change in the dislocation density, induced by plastic deformation, influences strongly the magnetic domain structure inside the material. Being so, classic parameters, like the coercivity or magnetic permeability, can be a good measure of the deformation level, yet their reliable determination in a non-destructive way in industrial environment is problematic. The magnetoacoustic emission (MAE) which results from the non-180° domain walls (DW) movement in materials with non-zero magnetostriction can be used as an alternative. The intensity of the MAE signal changes strongly as a result of plastic deformation for both tensile and compressive deformation. It is however possible to discern those cases by analysing the changes in the shape of the MAE signal envelopes. The set of the martensitic steel samples (P91) deformed up to 10% (for both tension and compression) was investigated. Due to geometrical limitations imposed by the special mounting system, enabling compression without buckling, the sample had the shape resulting in low signal to noise (S/N) ratio. Being so the optimization of FFT filtering and wavelet analysis was performed in order to improve sensitivity of the proposed method of deformation level determination.