Papers by Keyword: Phase Transformation Kinetics

Abstract: The effect of different volume fractions of pre-existing martensite on the low-temperature bainitic transformation and microstructures was quantitatively analyzed by dilatometer, optical microscope and scanning electron microscope. The results showed that pre-existing martensitic transformation accelerated the subsequent low-temperature bainitic transformation, and the incubation period and completion time of bainitic reaction were significantly shortened. This phenomenon was attributed to the increasing nucleation sites caused by the introduced dislocations in austenite due to the formation of pre-existing martensite. However, it was noteworthy that, because of the increased bainitic plates adjacent to the pre-existing martensitic plates, the probability of the impingement of bainitic plates during growth was increased, which resulted in a decrease in the maximum attainable volume fraction of bainite.

Abstract: Since the last few years several researches have focused in proposing new approaches to decrease the energy impact of the hot sheet stamping process. Among them, one possibility is to use steel grades characterized by a lower austenitization temperature in order to reduce the heating energy required for the process, but still maintaining the steel quenchability under the investigated process conditions. In the paper, the behaviour of two new steel grades developed by ArcelorMittal and devoted to hot stamping is investigated in terms of transformation kinetics and microstructural behaviour. An extensive dilatometric analysis, performed by using a modified axial extensometer, permitted to evaluate the steel kinetics of phase transformation as a function of the applied cooling rate, amount of strain and strain rate. The micro-hardness of water and die quenched samples of the investigated steel grades was measured and their microstructure analyzed by means of optical and scanning electron microscopy to evaluate the characteristics of the formed martensite and the possible phenomenon of autotempering.

Abstract: The critical nucleus size—above which nuclei grow, below dissolve—during diffusion controlled nucleation in binary solid-solid phase transformation process is calculated using kinetic Monte Carlo (KMC). If atomic jumps are slower in an A-rich nucleus than in the embedding B-rich matrix, the nucleus traps the A atoms approaching its surface. It doesn’t have enough time to eject A atoms before new ones arrive, even if it would be favourable thermodynamically. In this case the critical nucleus size can be even by an order of magnitude smaller than expected from equilibrium thermodynamics or without trapping. These results were published in [Z. Erdélyi et al., Acta Mater. 58 (2010) 5639]. In a recent paper M. Leitner [M. Leitner, Acta Mater. 60 (2012) 6709] has questioned our results based on the arguments that his simulations led to different results, but he could not point out the reason for the difference. In this paper we summarize our original results and on the basis of recent KMC and kinetic mean field (KMF) simulations we show that Leitner’s conclusions are not valid and we confirm again our original results.

Abstract: Using dilatometric technique, it is the purpose of present research to quantify the kinetics of austenite decomposition during cooling after isothermal hot compression. To reach this aim, based on lattice parameters and thermal expansion coefficients of different phases, a new model was proposed. In this model, the contributions of both isotropic and non-isotropic dilatations were considered. The model was applied to 22MnB5 steel to predict formation of bainite and martensite. Furthermore, the model was developed in order to determine the final fraction of deformation induced ferrite (DIF).

Abstract: The microstructure evolution of the supercooling austenite for the eutectoid carbon steel during deformation was analyzed. The experiment results show that during deformation part of the supercooling austenite transformed to pearlite. The pearlite transformation began around the grain boundary. The interlamellar spacing of the pearlite was fine. In addition, the kinetics of the austenite to pearlite transformation during deformation was discussed in detail. The analysis indicates that the kinetics of the pearlite transformation under deformation can be described by the Avrami equation: . Particularly, in the Avrami equation the value b varied obviously with the change of the deformation conditions. However, the influence of the deformation conditions on the value n was little.

Abstract: This work is part of a collaborative study between CEA-Saclay and LMT-Cachan on the numerical simulation of multi-pass GTA-Welding of thick specimens made of X10CrMoVNb9-1 (ASTM 387 or “T91”) steel. The final objective of this paper is to exhibit the prediction capabilities of an improved version of the Thermo-Metallurgical-Mechanical "TMM" model for X10CrMoVNb9-1 martensitic steel (initially developed by G.-M. Roux). In this paper, focus is made on the modelling of the martensite tempering due to the complex thermal loadings induced by the multi-pass process. Herein, it has been chosen to study the tempering kinetics via the evolution of the free carbon content or, conversely, the carbides precipitation overall fraction, growth and dissolution. Thermo-Electric Power (TEP) measurements as well as hardening measurements have been used to investigate the tempering phenomenon. Measurements are fast to perform and are not very sensitive to the geometry of the material (in opposite to resistivity measurements). A phenomenological tempering model was developed and identified from several tests at constant tempering temperatures ranging from 550°C up to 750°C. The improved TMM model including this tempering model was used to perform the 2D finite element analysis of the 16-pass GTA welding process of a narrow groove butt-weld. The predictions are favourably compared with the “real” microstructure.

Abstract: This work is part of a collaborative study between CEA-Saclay and LMT-Cachan on the numerical simulation of multi-pass GTA-Welding of thick specimens made of X10CrMoVNb9-1 (ASTM 387 or “T91”) steel. This material is considered as a candidate for some components of future Very High Temperature nuclear Reactors. Some parts of these components should be manufactured by assembling thick components (typically 200 mm) using narrow groove multi-pass GTA-Welding process. This welding process generates complex thermo-mechanical cycles in the HAZ (Heat Affected Zone) inducing complex microstructural transformations and residual stresses which should affect the integrity of the vessels behaviour. In a previous study, G.-M. Roux [1] developed a first version of a Thermo-Metallurgical-Mechanical "TMM" model for the X10CrMoVNb9-1 martensitic steel. This model was validated regarding residual stresses on simple mono-pass spot-welding tests. In this paper, focus is made on the modelling of the complex austenitisation process of the tempered martensitic steel as induced by the multi-pass process. Three different approaches are presented, viz. a model first proposed by Brachet et al., second a new model based on JMA approach and last, the simple differential Leblond model that is implemented in various finite element codes. These models are identified from standard dilatometry tests performed over a large range of heating rates, viz. [0.1°C/s, 100°C/s]. Finally, the response of these models, and therefore, their predicting capabilities, are compared to the experimental response of the material for different transients that have been designed to be representative of the temperature history in different points of a multi-pass welding HAZ.

Abstract: The precipitation of NbC in austenite is an important mechanism for improving the strength of steel because NbC-precipitates are known to decrease the ferrite grain size during the subsequent phase transformations upon cooling. The effect of the interaction between niobium (Nb) in solid solution and NbC-precipitates on the austenite-to-ferrite phase-transformation kinetics is not entirely clear. We study a high-purity Fe-C-Mn-Nb alloy cooled at different rates. Different annealing times at 850°C were applied to create different number densities and sizes of the NbC-precipitates in order to study the effect of NbC precipitation on the transformation kinetics. The alloy that is used in this study has an atomic ratio of Nb:C=1.3:1. The fraction of ferrite is measured as a function of temperature during cooling by means of dilatometry. The ferrite grain size is measured by means of optical microscopy. The results are interpreted with thermodynamic and kinetic models.

Abstract: A three-dimensional cellular automata (CA) model is developed for the kinetic and microstructural modelling of the relevant metallurgical mechanisms occurring in the annealing stage of low–alloy steels: recrystallisation, pearlite–to–austenite transformation and ferrite–to–austenite transformation on heating and austenite–to–ferrite transformation on cooling. In this model the austenite–to–ferrite transformation is described by a mixed–mode approach, which implies that the transformation kinetics is controlled by both the interface mobility and the diffusivity of the partitioning elements. This approach also allows incorporation of the ferrite nucleation occurring on structural defects. The developed CA algorithm, in which the transformation rules for the grain boundary and interface cells are controlled by the growth kinetics of the forming phase, allows three-dimensional systems to be treated within relatively short simulation times. The simulated microstructure reproduces quite well the microstructure observed in experimental samples. A good agreement is obtained between the experimental and simulated ferrite recrystallisation and ferrite and austenite transformation kinetics. The present approach also models the development of the carbon concentration profile in the austenite, which is, for instance, essential for subsequent martensite formation.

Abstract: The kinetics of phase transformations for which nucleation occurs on parent-micro-structure grain boundaries, and the resulting microstructures, were investigated by means ofgeometric simulations. The influences of parent microstructure grain-boundary area density,parent grain-size distribution and parent→product kinetics were analysed. Additionally, thesimulated kinetics were compared with predictions from two kinetic models, namely a modelproposed for spatially random nucleation and a model proposed for grain-boundary nucleation.It was found that the simulated transformed fraction as function of time lies in between the twomodel predictions for all investigated parent microstructures and parent→product kinetics.