• Fagan, Patrick

Abstract

The Barkhausen noise is a magnetic phenomenon employed for the non-destructive evaluation of ferromagnetic samples, mainly for microstructure evaluation and detection of surface defects. Its interpretation is still qualitative because the Barkhausen noise is sensitive to a great number of parameters, like the grain size. A simulation tool, allowing testing a great number of configurations, is sorely needed to improve quantitative analysis of Barkhausen noise measurements. The chosen simulation tool is the combination of the multi-scale model (MSM), giving anhysteretic functions as outputs, with a classical hysteresis model. The hysteresis loop of the Barkhausen noise has been obtained with the MBNE loop and it has been simulated by removing the magnetization rotation contribution in the MSM. This approach allows to compare the MBNE to the classical hysteresis loop, which has been tested without and with an uniaxial mechanical stress. Measurements have found a qualitative agreement with simulated loops, a better precision requiring very detailed studies on the texture of the material. 2D simulations show that the sensitivity of MBNE magnetic indicators heavily depend on the direction of both mechanical and magnetic excitation. A feedback algorithm has been implemented to obtain the sinusoidal magnetic induction required by the Statistical Theory of Losses (STL). First results show that the relationship between MBNE surface and excitation loop is similar to the classic STL law, which could be exploited to standardize the MBNE normalization coefficient. The association between MSM and hysteresis model gives qualitative results allowing to anticipate some trends of the chosen magnetic indicators, mainlywhen the mechanical stress is variable.

Topics

• impedance spectroscopy
• surface
• grain
• grain size
• theory
• simulation
• steel
• texture
• defect
• magnetization
• quantitative determination method