Prediction of Inclusion Types From BSE Images: RF vs. CNN

The analysis of non-metallic inclusions is crucial for the assessment of steel properties. Scanning electron microscopy (SEM) coupled with energy dispersive spectroscopy (EDS) is one of the most prominent methods for inclusion analysis. This study utilizes the output generated from SEM/EDS analysis to predict inclusion types from BSE images. Prediction models were generated using two different algorithms, Random Forest (RF) and convolutional neural networks (CNN), for comparison. For each method, three separate models were developed. Starting with a simple binary model to differentiate between inclusions and non-inclusions, then developing to more complex four and five class models. For the 4-class model, inclusions were split into oxides, sulfides, and oxy-sulfides, in addition to the non-inclusion class. The 5-class model included specific types of inclusions only, namely alumina, calcium aluminates, calcium sulfides, complex calcium-manganese sulfides, and oxy-sulfide inclusions. CNN achieved better accuracy for the binary (92%) and 4-class (78%) models, compared to RF (binary 87%, 4-class 75%). For the 5-class model, the results were similar, 60% accuracy for RF and 59% for CNN.

Interaction of rail steel melt with refractory lining

The rail steel properties are adversely affected by rigid non-metallic inclusions, containing aluminum oxides. Therefore, aluminum content is limited to 0.004 % wt. in rail steel grades. Aluminum can get into steel from charge materials and refractory lining. In this work, we’ve analyzed how the chemical composition of refractories used in rail steel making influence steel quality on example of one domestic enterprise. To determine the main types of non-metallic inclusions created in E76F rail steels, we have performed fractional gas analysis of the samples taken in various process steps. It was found that the slag composition after degassing changes insignificantly, while the most part of non-metallic inclusions in rail steel is represented by aluminates.

630 ̊С±30 ̊С - Nodal (Critical) Temperature of Iron and Carbon Steel

The article deals with the problems of withstanding harsh temperatures by steel and iron. The authors of the work discuss iron denser high-temperature of γ modification and maximums and minimums of impact. In addition, the article analyses the transformations of iron and anomalies of properties: peak of heat capacity, acceleration of diffusion, etc. The authors take into account the consensus on the causes of polymorphism and the theoretical model of ferromagnetism. Besides, there is a consideration of "transformation" in interaction between Fe atoms that produce anomalies of steel properties. It is necessary to note the transformation detected by anomalies of any properties including mechanical. In the presented work the authors have made an attempt to prove transformations in iron at ~650 °C on the basis of extreme values of hardness and microhardness, metallographic structure, parameters of fine structure, precipitation resistance force depending on temperature. Therefore, the analysis of literature sources on physical and mechanical properties of iron and its derivatives has been made.

Non-metallic Inclusions in Different Ferroalloys and Their Effect on the Steel Quality: A Review

Ferroalloys have become increasingly important due to their indispensable role in steelmaking. In addition, the demand for improved steel qualities has increased considerably, which in turn highlights the quality of ferroalloys. This is due to the fact that the impurities in ferroalloys directly and significantly influence the quality of steel products. To gain a better understanding of the main trace elements and inclusions in ferroalloys (such as FeSi, FeMn, SiMn, FeTi, FeCr, FeMo, FeNb, FeV, FeB, some complex ferroalloys) and their behaviours in steel melt after the additions of these ferroalloys, information from a large number of previous results on this topic was extensively reviewed in this work. The applications of different ferroalloys and their production trends were discussed. In addition, the effects of some trace element impurities from ferroalloys on the inclusion characteristics in steel were also discussed. The possible harmful inclusions in different ferroalloys were identified. Overall, the results showed that the inclusions present in ferroalloys had the following influence on the final steel cleanliness: (1) MnO, MnS and MnO–SiO2–MnS inclusions from FeMn and SiMn alloys have a temporary influence on the steel quality; (2) the effect of large size SiO2 inclusions (up to 200 μm) in FeSi and FeMo alloys on the steel cleanliness is not fully understood. The effect of Al, Ca contents should be considered before the addition of FeSi alloys. In addition, Al2O3 inclusions and relatively high Al content are commonly found in FeTi, FeNb and FeV alloys due to their production process. This information should be paid more attention to when these ferroalloys are added to steel; (3) except for the existing inclusions in these alloys, the Ti-rich, Nb-rich, V-rich carbides and nitrides, which have important effects on the steel properties also should be studied further; and (4) specific alloys containing REM oxides, Cr–C–N, Cr–Mn–O, Al2O3, Al–Ti–O, TiS and Ti(C, N) have not been studied enough to enable a judgement on their influence on the steel cleanliness. Finally, some suggestions were given for further studies for the development of ferroalloy productions.

Transition in iron and carbon steels at ~ 450 ºC

Вокруг интервала температур 400-500 °С сложилась любопытная ситуация. М.В. Белоус с соавторами в своей монографии «Превращение при отпуске стали» его просто не замечают в классификации четырех превращений при отпуске, хотя еще в 1925 г. П. Обергоффер, основываясь на минимуме при 400-500 °С термоЭДС пары железо-платина, писал: «Имеем ли мы здесь дело с дальнейшим превращениями в чистом железе, должны показать новые подробные исследования». В настоящей работе на основании анализа многочисленных литературных данных, а также собственных экспериментальных (металлографического исследования, рентгеноструктурного анализа, сопротивления «горячей» осадке на образцах из практически чистого железа (0,008% С)) сделана попытка обосновать превращение в железе при ~ 450 ºС. Признание превращения при данной температуре дает возможность прогнозирования аномалий свойств сталей (производного железа) на том или ином этапе термической обработки. There is a curious situation around the temperature interval of 400-500 °C. M.V. Belous and co-authors in their monograph «Transformation during tempering of steel» simply do not notice it in the classification of the four transformations during tempering, although back in 1925. P. Oberhoffer, based on a minimum at 400-500 °C of the thermo-EDC of iron-platinum pair, wrote: «Whether we are dealing here with further transformations in pure iron, should show new detailed research». In the present work based on the analysis of numerous literature data and our own experimental data (metallographic research, X-ray analysis, resistance of «hot» precipitation on practically pure iron samples (0.008% C)) we have made an attempt to prove the iron transformation at ~ 450 ºC. The recognition of transformation at this temperature makes it possible to predict anomalies of steel properties (iron derivatives) at a certain stage of heat treatment.

Lift-off Invariant Inductance of Steels in Multi-Frequency Eddy-Current Testing

Eddy current testing can be used to interrogate steels but it is hampered by the lift-off distance of the sensor. Previously, the lift-off point of intersection (LOI) feature has been found for the pulsed eddy current (PEC) testing. In this paper, a lift-off invariant inductance (LII) feature is proposed for the multi-frequency eddy current (MEC) testing, which merely targets the ferromagnetic steels. That is, at a certain working frequency, the measured inductance signal is found nearly immune to the lift-off distance of the sensor. Such working frequency and inductance are termed as the lift-off invariant frequency (LIF) and LII. Through simulations and experimental measurements of different steels under the multi-frequency manner, the LII has been verified to be merely related to the sensor parameters and independent of different steels. By referring to the LIF of the test piece and using an iterative inverse solver, one of the steel properties (either the electrical conductivity or magnetic permeability) can be reconstructed with a high accuracy.

Lift-off Invariant Inductance of Steels in Multi-Frequency Eddy-Current Testing

Eddy current testing can be used to interrogate steels but it is hampered by the lift-off distance of the sensor. Previously, the lift-off point of intersection (LOI) feature has been found for the pulsed eddy current (PEC) testing. In this paper, a lift-off invariant inductance (LII) feature is proposed for the multi-frequency eddy current (MEC) testing, which merely targets the ferromagnetic steels. That is, at a certain working frequency, the measured inductance signal is found nearly immune to the lift-off distance of the sensor. Such working frequency and inductance are termed as the lift-off invariant frequency (LIF) and LII. Through simulations and experimental measurements of different steels under the multi-frequency manner, the LII has been verified to be merely related to the sensor parameters and independent of different steels. By referring to the LIF of the test piece and using an iterative inverse solver, one of the steel properties (either the electrical conductivity or magnetic permeability) can be reconstructed with a high accuracy.

Consideration of Magnetic Measurements for Characterisation of Ferrite–Martensite Commercial Dual-Phase (DP) Steel and Basis for Optimisation of the Operating Magnetic Field for Open Loop Deployable Sensors

The magnetic properties of commercial dual-phase (DP) steels (DP600, DP800 and DP1000 grades) were evaluated using initial permeability, incremental permeability and coercivity and correlated with the key microstructural differences between the grades. The ferrite grain sizes and ferrite fractions have been compared with the magnetic parameters obtained from minor and major magnetisation loops within each DP grade. It has been revealed that the incremental permeability increases with the applied magnetic field amplitude to reach a peak and then drops at a higher magnetic field, with the values being different for the three DP grades at a lower field and converging to a similar permeability value at the high field. The effects of ferrite grain size and phase fraction on the incremental permeability are considered, and it has been shown that the influence of ferrite grain boundaries on magnetic permeability is more dominant than the effect of ferrite fraction in commercial DP steel samples. An analysis of the correlation between coercivity and initial permeability with tensile strength shows that the initial permeability provides a slightly better prediction of strength for the steels examined, which is believed to be due to the fact that a combination of reversible and irreversible domain components affect the coercivity value, while the initial permeability is predominantly affected by reversible domain movements. Based on the trend between incremental permeability and applied magnetic field and the commercial EM sensor (EMspec) operating parameters, the effect of lift-off and hence magnetic field strength on the sensitivity to DP steel properties can be assessed.