Magnetic NDT for Steel Microstructure Characterisation – Modelling the Effect of Ferrite Grain Size on Magnetic Properties

International audienceThe mechanical properties of steels are controlled by their microstructural parameters, such as grain size, phase balance and precipitates, which are developed during thermo-mechanical processing. It is desirable to be able to monitor microstructural changes during processing, allowing in-situ feedback control, or to characterize microstructure in steel products in a non-contact and non-destructive manner. Electromagnetic (EM) measurements are sensitive to changes in magnetic properties, which, in steels, vary with composition, microstructure and temperature. In order to interpret the EM signal for microstructure analysis, it is important to be able to predict the magnetic properties from the microstructural parameters.In this paper, an extra low carbon steel has been used to generate a single phase microstructure (ferrite) with a range of grain sizes of 14-78 μm. The grain structures were characterised by optical microscopy and EBSD. A Voronoi based algorithm (Multi Level Voronoi), which provides a parametric description of the microstructure, i.e. the boundaries, in a grid format, is used to generate 2D and 3D microstructure models with different grain sizes. Based on these microstructure models, multi scale 2D and 3D magnetic modelling approaches were applied to predict the magnetic properties. At the micro scale, a 3D micro-magnetic simulation code (EMicroM), which considers a finite volume (1003 μm3 cube blocks) of material at a resolution down to the thickness of the domain wall (0.2 μm), has been used to derive the dynamic magnetic behaviour (i.e. hysteresis curve). The predicted coercive fieldvalues have been shown to decrease when the grain size increases, in line with experimentalobservations. The modelled hysteresis loop has a similar coercive fieldvalue with the experimental measured one, although the slopes of the curves are different. At the meso level, a 2D finite element modelling approach using COMSOL Multiphysics was used tosimulate the EM response of an area of around 1000×1000μm2. The model predicts effective relative permeability by considering the ferrite grains and grain boundary regions as constituents with different relative permeability values. The modelled results agree well with the experimentally determined permeability values from EM sensor measurements