Vacuum arc remelting is a secondary melting process used to produce a variety of segregation sensitive and reactive metal alloys. The present day VAR practice for superalloys involves, typically, melting electrodes of 17'' into ingots of 20'' in diameter. Even larger diameter forging stock is desirable. However, beyond 20'' ingots of superalloys are increasingly prone to segregation defects if solidification is not adequately controlled. In the past years a new generation of model-based controllers was developed to prevent segregation in VAR by controlling melt rate, or the total amount of power flowing into the liquid pool. These controllers were seen as significant improvements in the industry of remelting processes, but these controllers were still focusing on the melting sub-process and ignoring ingot solidification. Accurate control of the liquid pool profile is expected to result in segregation-free ingots, but unfortunately a controller capable of stabilizing the solidification front in VAR is currently not available. The goal of the proposed research is to develop a cyber-enabled controller for VAR pool depth control that will enhance the capabilities of current technologies. More specifically, the objectives of this research are threefold. Firstly, a control-friendly model is proposed based on a high-fidelity ingot solidification model and is coupled to a thermal model of electrode melting. Secondly, sequential Monte Carlo estimators are proposed to replace the traditional Kalman filter, used in the previous VAR controllers. And finally, a model predictive controller (MPC) is designed based on the proposed reduced-order model. The time-critical characteristics of these methods are studied, and the feasibility of their real-time implementation is reported. / text

Vacuum Arc Remelting (VAR) is a secondary process used for homogenization of high-melting-point and oxygen-sensitive materials such as superalloys and titanium alloys. The VAR process is carried out with the aim of melting a large consumable electrode in such a way that the resulting ingot has improved homogeneity. The Specialty Metals Processing Consortium (SMPC) has spent the past 20 years developing technology to improve control over the final ingot remelting and solidification processes to alleviate conditions that lead to the formation of inclusions and segregation. Channel segregates are concentration defects arising during the solidification of large-diameter solute-rich alloys. As manufacturers for turbine engines and generators call for larger ingots, it becomes more difficult to produce them without these defects. If, however, liquid pool depth can be controlled precisely to stabilize the solidification zone in the ingot, we could, in principle, produce larger ingots that are defect free. A problem arises because measurements obtained from the VAR furnace do not give enough information to accurately estimate the liquid pool shape in dynamic melting situations. Also, the solidification process in VAR is extremely complex due to the multiple physical domains present and a high-fidelity model is required to give an accurate description of the dynamic process. The Basic Axisymmetric Remelting (BAR) code was initially developed by Lee Bertram at Sandia National Laboratories as a high-fidelity multi-energy model to describe ingot casting in this system. In this work we present a new strategy to improve the accuracy of the estimates used in the control system. This strategy consists of implementing BAR as a new set of measurements to be used by the estimator. This new strategy was used in tests jointly sponsored by SMPC and Los Alamos National Laboratory (LANL) in February 2011 using a laboratory-scale furnace and alloy 718 electrodes. / text

No description available.

Vacuum-Arc Remelting (VAR) is an important process for manufacturing Ti- and Ni-based superalloys. Currently, the sources and mechanisms behind microstructural anomalies produced in VAR superalloy ingots are not well understood. In order to help understand formation processes, a method of characterizing specific anomalies in VAR ingots is desired. This paper presents a method of characterizing the composition and morphology of anomalies in VAR alloy ingots using a combination of serial sectioning and X-ray fluorescence (XRF) energy dispersive spectroscopy (EDS) techniques. This process is demonstrated on a dirty white spot from an Alloy 718 sample. The white spot of interest was serial polished and 2-D XRF EDS maps were acquired at each polish depth. The EDS maps were then stacked to form a 3-D representation of the white spot. In addition, SEM and optical microscopy techniques were used to further characterize the composition and morphology of the dirty white spot. The dirty white spot is composed of both Ti-enriched and Nb-depleted regions. The 2-D EDS maps acquired with the XRF equipment provided adequate contrast for creating a 3-D representation of the Ti-rich region of the dirty white spot. However, contrast was not sufficient to create a 3-D representation of the Nb-depleted region. The XRF EDS equipment combined with SEM and optical microscopy techniques provided valuable information about the morphology and composition of the Alloy 718 dirty white spot. It is concluded that this dirty white spot was produced by fall-in from either the crown or shelf regions during the VAR process. / text

Dans le procédé de refusion à l’arc sous vide, la structure et la dynamique de l’arc électrique conditionnent les distributions spatiales d’énergie et de courant au sommet du lingot refondu. Ces distributions impactent fortement les champs de température et de vitesse du métal liquide, qui gouvernent les conditions de solidification du lingot et donc la qualité finale du produit. Une étude par mesures optiques, de la dynamique de l’arc aux longues échelles de temps lors de refusions industrielles d’alliages de titane a été entreprise. Cette analyse a été effectuée par une méthode qualitative, à l’aide d’enregistrements vidéo, et par une nouvelle méthode de diagnostic quantitative développée spécifiquement dans ce travail, à base de photodiodes. L’analyse de la dynamique de l’arc confirme la présence d’un mouvement d’ensemble de l’arc électrique lors de refusions d’alliages de titane et montre une forte corrélation entre la dynamique observée et l’intensité du brassage électromagnétique imposé. Nos résultats ont de plus permis de prédire la dynamique probable de l’arc dans des conditions non étudiées et de proposer de possibles origines à la dynamique observée. Enfin, une première approche de modélisation de cette dynamique dans le logiciel de simulation RAVEL est proposée afin d’étudier son impact sur la solidification du lingot. Ce travail présente également une modélisation détaillée des rayonnements thermiques émis au sommet du lingot, basée sur la méthode des radiosités / In the vacuum arc remelting process, the structure and dynamics of the electric arc are responsible for the spatial distributions of energy and current at the top of the remelted ingot. Those distributions strongly impact the ingot temperature field and the liquid metal velocity field, which govern the ingot solidification conditions, hence the final product quality. A study based on the optical measurement of the electric arc dynamics at a long time-scale during vacuum arc remelting of titanium alloys was undertaken. This analysis was performed either qualitatively using melt video recordings, or quantitatively by a new specifically developed diagnostic technique based on the use of photodiodes. The analysis of the electric arc dynamics confirmed the presence of an ensemble arc motion during vacuum arc remelting of titanium alloys and showed a strong correlation between the observed dynamics and the magnitude of the imposed electromagnetic stirring. Furthermore, our results allow to predict the arc dynamics in non-studied conditions and to propose possible origins for the observed behaviors. Finally, a first simulation of the influence of the arc behavior is proposed in order to study its impact on the ingot solidification. This work comprises also a detailed modelling of the thermal radiation at the ingot top, based on the radiosity method

Le procédé VAR (Vacuum Arc Remelting ou refusion à l'arc sous vide en français) est employé dans la production d'alliages à haute performance pour les industries aéronautique (aciers spéciaux, superalliages et alliages de titane) et nucléaire (alliage de zirconium). Comme pour tous les procédés de fonderie, la maîtrise de l'homogénéité chimique et de la structure métallurgique des lingots coulés par le procédé VAR constitue un enjeu industriel important. Les travaux présentés dans ce mémoire visent à identifier, pour les alliages de zirconium en particulier, les effets de la convection naturelle et de la convection forcée due au brassage électromagnétique sur la macroségrégation. Dans ce but, un modèle numérique a été développé. Il est basé sur la résolution couplée des équations de conservation d'énergie, de quantité de mouvement et de solutés, dans des conditions d'écoulement laminaire ou turbulent. La modélisation de la solidification tient compte du couplage fort entre le transport d'énergie et de solutés dans la zone pâteuse. Afin de décrire la microségrégation, la diffusion restreinte des solutés dans les phases liquides et solides peut être prise en compte. Parallèlement, deux électrodes chimiquement homogènes d'alliages Zircaloy-4 et M5® ont été spécialement refondues dans un four VAR industriel sur le site de CEZUS à Ugine (Savoie, France). La macroségrégation des lingots obtenus a été caractérisée.La comparaison entre les mesures expérimentales et les résultats de simulation a montré que pour un alliage dont l'intervalle de solidification est important (comme l'alliage Zircaloy-4), la convection solutale dans la zone pâteuse peut avoir une influence essentielle sur la macroségrégation de la région centrale du lingot. Par ailleurs, le mouvement de grains équiaxes lors de l'application d'un brassage électromagnétique de forte intensité semble accentuer significativement la macroségrégation dans la région externe du lingot. Pour un alliage dont l'intervalle de solidification est faible (comme l'alliage M5®), nous avons montré que la macroségrégation dépend plus spécifiquement de la convection forcée due au mode de brassage électromagnétique appliqué au cours de la refusion / Vacuum Arc Remelting (VAR) is used to produce high performance alloys for the aeronautic (special steels, superalloys, titanium alloys) and nuclear (zirconium alloys) industries. As for all casting processes, the control of the chemical homogeneity and the metallurgical structure in VAR ingots is an important industrial issue. The goal of this thesis is to identify, for zirconium alloys in particular, the effects of the natural convection and the forced convection due to the electromagnetic stirring on macrosegregation. To this purpose, a numerical model has been developed. It is based on the solution of the coupled transient energy, momentum and solute transport equations, under laminar or turbulent flow conditions. The solidification modeling accounts for a full coupling between energy and solute transport in the mushy zone. The finite diffusion of solutes in both solid and liquid phases can be taken into account to describe microsegregation. In addition, chemically homogeneous Zircaloy-4 and M5® electrodes have been specially remelted in an industrial VAR furnace at the CEZUS plant in Ugine (Savoie, France). The macrosegregation of the ingots has been measured. The comparison between the experimental measurements and the simulation results showed that for an alloy with a large solidification interval (like Zircaloy-4), the solutal convection in the mushy zone could have an essential influence on the macrosegregation in the inner part of the ingot. Furthermore, the motion of equiaxed grains caused by a strong stirring seems to seriously intensify macrosegregation in the outer part of the ingot. For an alloy with a small solidification interval (like M5®), we have shown that the macrosegregation depends more specifically on the forced convection due to the type of stirring applied during the remelting

L’apparition d’inclusions exogènes demeure un problème majeur pour les élaborateurs de titane. Afin d’améliorer la propreté inclusionnaire des lingots élaborés par le procédé de refusion à l’arc sous vide (Vacuum Arc Remelting), une étude numérique et expérimentale a été réalisée. La partie numérique de la thèse consiste à modéliser le comportement d’un défaut hard-α provenant de l’électrode consommable et tombant dans le puits liquide du lingot. Un modèle décrivant le processus de dissolution prédit l’évolution de la taille d’une inclusion durant son séjour dans le puits liquide. La trajectoire est déterminée à l’aide d’un modèle lagrangien tenant compte de la turbulence de l’écoulement en modifiant le coefficient de trainée. Les deux modèles ont été couplés et implémentés dans le logiciel SOLAR, qui simule la croissance d’un lingot VAR.Les résultats mettent en évidence la difficulté d’éliminer une inclusion hard-α avec une seule refusion, principalement à cause de la croissance d’une couche de phase β pendant les premiers moments de l’immersion. Le comportement global du défaut dépend fortement de l’hydrodynamique du puits et des caractéristiques de l’inclusion.Pour étudier la dissolution expérimentalement, des défautssynthétiques (hard-α et HDI) ont été immergés dans un bain de titane liquide chauffé dans un four à bombardement électronique. Les vitesses de dissolution ont été déterminées en mesurant les dimensions des défauts avant et après les expériences et ont été ensuite utilisées pour valider les modèles numériques. Par ailleurs, nous avons mis en évidence la grande influence de la température et de la vitesse de l’écoulement sur les cinétiques de dissolution / The presence of exogeneous inclusions has always been a major concern for the titanium industry. To help improve the inclusional cleanliness of VAR (Vacuum Arc Remelting) titanium ingots, a numerical and experimental study was undertaken.The numerical model is capable of predicting the motion and dissolution of a hard-α defect falling from the electrode tip into the ingot melt pool during vacuum arc remelting. It is implemented in SOLAR, a CFD code that simulates the ingot growth and solidification. The dissolution of the inclusion is governed by nitrogen diffusion from the defect towards the surrounding molten metal. A model describing this phenomenon predicts the particle size evolution and the nitrogen profile at each moment. The motion of the spherical particle is tracked using a Lagrangian model and the influence of turbulence is accounted for by a modification of the drag coefficient.Results show that inclusion removal is difficult with a single melt since the growth of a β-phase layer leads to an initial increase in the defect size. The inclusion behaviour is highly dependent on the pool hydrodynamics and on inclusion characteristics.In order to clarify dissolution aspects of these defects and to measure their dissolution kinetics, synthetically processed defects were introduced into molten titanium heated in an electron beam melting furnace. Dissolution rates were calculated by measuring the size of the defects before and after the experiments and the results were used to validate the numerical models. Furthermore, the experiments show that dissolution kinetics highly depend on fluid motion and temperature