Electromagnetic (EM) sensor signals are sensitive to changes in permeability and resistivity in steel, where both are determined by microstructure, chemical composition and temperature. EM sensors are currently being used in the steel industry to monitor the phase transformation of hot strip on the runout table. However, there are factors still not fully understood when it comes to the EM signal such as relationship between phase fraction transformed at high temperatures and the EM signal. The purpose of this research is to quantitatively relate the EM of zero crossing frequency to microstructure and more specifically phase fraction during transformation of steels at high temperature.

The EMspec速 industrial EM sensor which is currently used in industrial conditions on the run-out table of a hot strip mill has been installed into a run out table simulator that consists of a furnace and roller run out table. This system has been designed and constructed to allow the EMspec速 sensor to be able to monitor various steels through transformation during dynamic cooling on a lab-based scale.

A full 3D finite element model of the EMspec速 sensor has been developed in a parallel project and is able to use permeability and resistivity values to predict zero crossing frequency. This model has been used in this project to along with measured resistivities and extrapolated permeabilities from the literature to predict zero crossing frequency. The power law was used along with the permeability values, resistivities and phase fraction obtained from dilatometry to predict zero crossing frequency.

The EMspec速 industrial sensor has been used to monitor the full phase transformation of various carbon steels including mild steels of 3, 6 and 10mm in thickness, a high carbon steel and medium carbon steel. The EMspec速 sensor was able to distinguish between the mild steels of different thickness due to their different cooling rates in air and therefore transformation trajectory. The EMspec速 sensor was also used to successfully monitor a 2.25 Cr- 1 Mo steel to measure the full transformation to bainite below the Curie temperature. The magnetic transformation of the duplex stainless steel was also monitored by the EMspec速 as it cools through the Curie temperature. The EMspec速 sensor was shown to be sensitive to the formation of sigma phase as a result of heat treatment of the duplex stainless steel. The sample with higher amount of sigma phase had a lower zero crossing frequency.

When relating the ZCF to phase percentage transformed obtained from dilatometry, the 6mm and 10mm mild steels, the high carbon and the 2.25 Cr- 1Mo steel all needed to be approximately 30-40% transformed before an increase in ZCF would occur due to the need for ferromagnetic phase connectivity before permeability significantly increases. This agrees well with the literature for room temperature ferrite fractions however there is room for discrepancies when taking into the temperature difference between a dilatometry sample and a larger sample cooling on the ROT. For the mild steels, the ZCF would peak before dilatometry predicts the steels to be 100% transformed.

The modelling data agreed well with the experimental data for steels that consisted of lower permeability phases such as the 2.25Cr-1Mo and high carbon steels. For the 3mm, 6mm and 10mm thick mild steels, the model could not solve for the high permeabilities which may be due to a more refined mesh being required. The model was able to solve for the duplex stainless steel however the modelled data did not agree with the experimental data due to the nature of the calculation of effective permeability. Power law was used however this does not consider the microstructural parameters such as the connectivity of ferrite. The permeabilities at lower phase fractions is higher as a result of this.

The work in this project shows that the EMspec速 industrial sensor can monitor the full transformation of multiple steels below the Curie temperature and the EMspec速 signal can be related quantitatively to phase percentage transformed although the discrepancy between a dilatometry sample and a large sample with inhomogeneity in temperature cannot be ignored.