FR3074072A1 - ALUMINUM CASTING PROCESS AT LOW SPEED AND LOW FREQUENCY - Google Patents

Abstract

The invention is a method for forming an aluminum alloy ingot (1). During the casting of the aluminum alloy, a sliding magnetic field is applied to the marsh and the amplitude of the magnetic field is periodically varied according to a frequency (f), so as to apply a Lorentz force (F) at different points. of the marsh, the average Lorentz force, during a period (P), being inclined relative to the vertical axis (Z) according to an angle of inclination (?). The method is characterized in that the casting velocity (V) and the frequency (f) are such that throughout the middle zone of the marsh, at the forehead, interface between the swamp and the solidified alloy, the angle d The inclination of the force is strictly less than the angle of inclination of the forehead.

Description

LOW-SPEED, LOW-FREQUENCY ALUMINUM CASTING PROCESS DESCRIPTION

TECHNICAL AREA

The technical field of the invention is the manufacture of ingots following a casting of a liquid alloy.

PRIOR ART

During a vertical casting, aimed at forming an ingot, the solidification of a metal or a metal alloy is affected by phenomena known as macroscopic segregation. During the cooling of the metal, convection currents are formed, generating recirculation vortices, the latter being at the origin of macroscopic segregations when their lifetime is of the same order of magnitude as the characteristic durations of solidification. These phenomena lead, in the solidified ingot, to a local impoverishment or to a local enrichment in chemical species. These macroscopic segregations, or macrosegregations, are at the origin of heterogeneities in the composition of the ingot.

A macrosegregation well known to those skilled in the art is the negative central macrosegregation, resulting from a depletion of eutectic alloy elements, along a vertical central axis of the ingot. These macro-aggregations have been described in the work of John Wiley et al "Direct-Chili Casting of light alloys", Editor Wiley, September 2013, pp 158-172.

The main mechanisms at the origin of the central macro-segregation described in this work are: • Thermosolutal convection in the marsh caused by temperature and concentration gradients, and the penetration of these convective flows in the pasty zone;

• The transport of grains in the supercooled zone under the effect of gravity, Archimedes' force and natural or forced convection;

• The flow in the pasty zone caused by the volumetric solidification shrinkage, which can be assisted by metallostatic pressure;

• The flow of liquid in the pasty area caused by mechanical deformations;

• Forced flows which may result from lodging, injection or release of gas, stirring, vibration, etc. which enter the supercooled zone and the pasty zone and modify the direction of the convection movements.

It is a continuous macro-aggregation, this term designating the fact that the macro-aggregation takes place continuously over all or part of the height of the ingot, in other words that it is essentially uniform along the axis of casting. .

The phenomenon of intermittent macro-aggregation has been less often described in the literature and results in the formation of V-shaped bands on either side of the negative central macro-aggregation. These V-shaped bands are alternately enriched and depleted in eutectic and peritectic alloy elements. These bands can be observed by X-ray radiographs of vertical slices of ingots, typically in the L / TC plane at half-width, when the segregated elements absorb X-rays in a differentiated way from the atoms of the metal composing the ingot. Other means make it possible to visualize this phenomenon, for example ultrasound or observation with the naked eye of anodized vertical slices, due to the difference in optical reflectivity between the zones enriched or depleted in alloying elements. Generally, intermittent macro-aggregation is most marked in the vicinity of the T / 2.5 region, and typically between T / 2.3 and T / 3.3, the T / 2 region corresponding to the central axis of the ingot. According to a nomenclature known to those skilled in the art, the term T / n, where n is a positive number, designates a region located at a distance T / n from an edge of the ingot, where T designates a thickness of the ingot.

Periodic intermittent macro-aggregations appear very soon after the start of casting, as soon as an inclined front is formed between a solid zone and a liquid zone. They are observed in all cases of casting of aluminum alloys typically loaded according to formats of thickness greater than 300mm, this thickness threshold itself depending on the speed of casting.

The publication R.C Dorward et al. "Banded segregation patterns in DC cast AlZnMgCu alloy ingots and their effect on plate properties" Aluminum, 1996, 72. Jahrgang, 4, p.251-259 describes the formation of bands of intermittent segregation in a 7000 alloy. According to these authors, this phenomenon is due to avalanches of grains periodically triggered by convective oscillations of the marsh, that is to say the liquid phase of the metal, in connection with a mechanism of emission of vortices. This article shows in particular that intermittent macro-segregation can be the source of variations in mechanical properties, for example toughness, on sheets obtained from raw casting products. It is therefore advantageous to find a casting process which would eliminate these intermittent macro-aggregations.

The reduction or elimination of continuous macro-aggregations, for example central macro-aggregation, has already been described. In particular, it has been shown that the application of a magnetic field, for the purpose of stirring or braking the flows, makes it possible to limit the appearance of continuous macro-aggregations. The document US5375647 describes, for example, a process for reducing central macro-segregation occurring during the casting of a metal alloy ingot. This method includes the application, during cooling, of a static magnetic field generated by at least one coil traversed by a direct current.

The inventors have considered that the methods described above do not make it possible to effectively reduce the occurrence of intermittent macrosegregations. They propose a process making it possible to limit the formation of such segregations, or even to eliminate them, so as to better control the mechanical properties of the products resulting from the casting.

STATEMENT OF THE INVENTION

A first object of the invention is a method for forming an aluminum alloy ingot in an ingot mold, the ingot mold defining a parallelepiped, such that the ingot formed extends parallel to a longitudinal axis, along a width and parallel to a transverse axis, along a thickness, the thickness being less than the width, the ingot having a median plane extending along the mid-thickness, parallel to the longitudinal axis, the method comprising the following steps:

preparation of the liquid aluminum alloy;

pouring of the liquid aluminum alloy into the mold, along a vertical axis, the alloy being cooled, during casting, by a runoff of a coolant liquid in contact with the mold and / or the alloy , so as to form a solid alloy, extending around a liquid alloy, called swamp, the solid alloy forming a front at the interface with the swamp, the front being inclined at an angle of inclination, relative to the vertical axis, the angle of inclination of the front varying along the transverse axis;

displacement of the solid alloy, along the vertical axis, according to a casting speed;

during casting, application of a sliding magnetic field whose amplitude is varied periodically according to a frequency, the magnetic field being generated by at least one magnetic field generator disposed at the periphery of the mold, so as to apply a Lorentz force at different points in the marsh, the Lorentz force averaged over a period of the magnetic field, being inclined relative to the vertical axis at an angle of inclination, called the inclination angle of the Lorentz force, the latter varying along the transverse axis;

the sliding magnetic field propagating along an axis of propagation parallel to the vertical axis;

the marsh comprising a median zone, extending symmetrically on either side of the median plane, the thickness of which corresponds to half the thickness of the ingot;

the method being characterized in that:

the frequency is less than 5 Hz;

and in that the casting speed and the frequency are adapted so that in the median zone of the marsh, at the interface between the liquid alloy and the solid alloy, at the level of the front, the angle of inclination force is strictly less than the angle of inclination of the forehead;

so that throughout the middle area, the Lorentz force acts on the liquid alloy, at the interface, pressing it against the forehead.

In other words, if T / 2 designates half the thickness of the ingot, the technical plating effect is obtained between T / 2-T / 4 and T / 2 + T / 4.

The term Lorentz force averaged over a period means a Lorentz force determined according to a time interval corresponding to the inverse of the frequency.

The method makes it possible to obtain a significant reduction in intermittent macro-segregations in the middle zone.

In the middle zone, the angle of inclination of the forehead is between 0 ° and 90 °. The lower it is, the more the front is oriented parallel to the vertical axis. The same is true of the angle of the Lorentz force. An angle value of 0 ° corresponds to a vertical orientation.

Preferably, the angle of inclination of the average Lorentz force is less, by at least 4 °, than the angle of inclination of the front, so that the average Lorentz force is more inclined, towards the vertical, than the forehead.

Preferably, the frequency is less than 2 Hz or less than 1 Hz. Preferably, the casting speed is less than 45 mm / minute or 40 mm / minute.

According to one embodiment, the casting speed and the frequency are adapted so that in the middle zone of the marsh, in an interface layer between the liquid alloy and the front, the angle of inclination of the force Lorentz average is strictly less than the angle of inclination of the front, the interface layer having a thickness, in a direction perpendicular to the front, is less than 2cm or 1 cm or 5 mm.

Preferably, the magnetic field is modulated, at the frequency, between a minimum value and a maximum value, the maximum value is constant during the casting. Thus, the average value of the Lorentz force is constant during several successive periods, for example during at least 10 successive periods.

The aluminum alloy can in particular be chosen from alloys of the 2XXX, 6XXX or 7XXX type.

Advantageously, the thickness of the mold is greater than 300 mm.

The method may include, prior to casting, a modeling of the Lorentz force applying at at least one point on the front, so as to define, taking into account the thickness of the mold, a frequency value and / or a speed of casting value enabling an average Lorentz force to be obtained, the angle of inclination of which relative to the vertical, is less than the angle, at said point, formed by the front relative to the vertical. Preferably, this modeling is carried out at different points, along the front, along the transverse axis. The modeling can make it possible to define a value of frequency and / or a value of speed of casting making it possible to obtain an average Lorentz force whose angle of inclination, relative to the vertical, is less than 4 ° to the angle of inclination formed by the front relative to the vertical.

A second object of the invention is an ingot, in particular an aluminum alloy ingot, obtained by a process according to the first object of the invention.

Other advantages and characteristics will emerge more clearly from the description which follows of particular embodiments of the invention, given by way of nonlimiting examples, and represented in the figures listed below.

FIGURES

Figures IA to 1D illustrate an example of a device allowing an implementation of a method according to the invention. Figures IA and IB show the main components of the device. Figures IC and 1D respectively represent a spatial and temporal distribution of the amplitude of a sliding magnetic field, according to an exemplary embodiment.

FIG. 2 illustrates a Lorentz force, applying to part of the marsh of a flow, implementing the invention.

FIGS. 3A, 3B, 3C and 3D show results of simulations making it possible to obtain the average orientation of a Lorentz force in the marsh, at the level of the front, respectively at different frequencies, for a casting speed of 55 mm. per minute.

Figures 3E, 3F, 3G and 3H show results of simulations to obtain the average orientation of a Lorentz force in the marsh, at the front, respectively at different frequencies, for a casting speed of 35 mm per minute.

FIGS. 4A and 4B are curves established by considering respectively different frequencies, and representing an evolution of an angle, called differential angle, along the front, the differential angle representing a difference between the angles, relative to the vertical, respectively formed by the Lorentz force and the front. In FIG. 4A, a casting speed of 55 mm per minute has been considered. In FIG. 4B, a casting speed of 35 mm per minute has been considered.

FIG. 4C shows an abacus making it possible to define an operating domain as a function of the casting speed (abscissa axis) and of the thickness of the casting (ordinate axis), in order to obtain an orientation of the Lorentz force according to l invention, in a median zone extending between T / 2 ± T / 4.

FIG. 5A represents a horizontal profile, parallel to an axis defining the thickness of an ingot, representative of a concentration of an alloying element in an ingot, as well as smoothed profiles, obtained by different smoothings of the horizontal profile.

FIG. 5B shows a profile representative of the intermittent macro-segregation affecting an ingot obtained by casting.

Figures 5C and 5D are representative profiles of intermittent macro-segregation, represented in a Fourier space. The abscissa axis corresponds to the spatial period, expressed in mm. The profiles were obtained from ingots respectively formed with and without implementation of the invention.

EXPLANATION OF PARTICULAR EMBODIMENTS

Figures IA and IB illustrate an ingot mold allowing an implementation of the invention. In this example, an aluminum alloy 1 flows in an ingot mold 2, through an opening 2i. The casting takes place along a vertical Z axis, through the ingot mold. The ingot mold is delimited by a peripheral enclosure the section of which, in a horizontal XY plane, is parallelepiped. The ingot mold defines a frame, parallel to a longitudinal axis Y, along a width W, and, parallel to a transverse axis X, by defining a thickness T. The width W is greater than the thickness T. The thickness T corresponds to a distance between two vertical walls 2p delimiting the ingot mold 2. The casting forms a parallelepipedal ingot. A plane, called median plane M, extends at mid-thickness (T / 2), parallel to the vertical axis Z and the longitudinal axis Y of the ingot. The thickness T is preferably between 300 mm and 750 mm. It has been found that intermittent macro-aggregations appear markedly when the thickness T exceeds 300 mm. In the examples or simulations mentioned in this description, a thickness T equal to 525 mm has been considered. The width W is equal to 1650 mm. The length, depending on the vertical axis, can reach several meters, for example between 3 and 10 meters.

A cooling fluid 3, for example water, flows against the wall of the solidified ingot. This process is known as semi-continuous casting by direct cooling ("DirectChili Casting"). A false bottom 4 is translated so as to move away from the opening 2i during the casting. The translation speed of the false bottom corresponds to a so-called casting speed V.

Under the effect of cooling, a zone of solid alloy ls is formed, near the cooled enclosure, around a zone of liquid alloy 1 £, designated by the term "marsh". The interface between the swamp 1 £ and the solid zone ls forms a front 10. At the end of the cooling, the ingot, also designated by the term "product", is formed. The front 10 has a slope, relative to the vertical, which varies according to the thickness. The angle of the front is called an angle a between the tangent to the front, at a point, and the vertical, that is to say the axis Z. The smaller the angle of the front a, the more the tangent to forehead is oriented vertically. The angle of the front is shown in FIG. 2. The angle of the front varies along the transverse axis X.

In the example shown in Figure IA, the front 10 is stationary: it remains in substantially the same position, while the material moves vertically, at the casting speed.

In the methods according to the prior art, intermittent macro-segregations 11 are formed in the ingot, and in particular in a thickness range between T / 2.3 and T / 3.3 on either side of the median plane M.

The alloy is an aluminum alloy from the 1XXX, 2XXX, 3XXX, 4XXX, 5XXX, 6XXX, 7XXX or 8XXX series. Alloys whose mass fraction of alloying elements is greater than 1%, even greater than 3% or even 5% are particularly suitable for a method according to the invention, because the more this mass fraction of these alloying elements is important, the more marked the intermittent segregations. The invention is particularly advantageous for alloy products of type 2XXX, 5XXX, 6XXX or 7XXX.

A magnetic field generator 5 has been shown, capable of generating a magnetic field B intended to be applied to the liquid alloy 1 £. Such a generator can be a mobile permanent magnet or an electromagnetic inductor, the latter generating a magnetic field when it is traversed by an electric current, called induction current.

The magnetic field B applied to the liquid alloy 1 £ is an alternating field, of amplitude B _o and of frequency f. The effect of this magnetic field is to apply a mixing of the marsh, under the effect of Lorentz forces applying to the liquid alloy 1 £. Indeed, the application of a magnetic field B generates, in the alloy, the formation of an electric current J resulting, within the liquid alloy 1 £ subjected to the magnetic field, in the appearance of a Lorentz force F such that F oc J x B where x denotes the vector product operator, and oc denotes a relation of proportionality. The Lorentz force has an oscillating component at a frequency twice the frequency f of the magnetic field.

Due to the thickness of the mold, the frequency f is chosen so as to allow sufficient penetration of the magnetic field B into the swamp, so as to obtain efficient mixing of the liquid. The frequency f is lower the higher the thickness of the product. In the case of an aluminum alloy with a thickness greater than 300 mm, the frequency is preferably less than 5 Hz, and even more advantageously less than 2 Hz or 1Hz.

The generator 5 is able to generate a sliding magnetic field. The term sliding magnetic field designates an alternating magnetic field, the amplitude B _{o of which} is not constant, and varies between a minimum value and a maximum amplitude B ™ ^ax , the maximum amplitude Β ™ ^αχ propagating along an axis of propagation Δ, preferably rectilinear and oriented along the vertical axis Z. By amplitude is meant the maximum value taken by a periodic quantity. The application of a sliding magnetic field results, at a point in the swamp, by a periodic variation in its amplitude. Thus, the amplitude of the magnetic field at a point in the marsh varies as a function of time, between a minimum amplitude B ™ ⁱⁿ and a maximum amplitude B ™ ^ax .

The sliding magnetic field generator 5 can be constituted by several electromagnetic inductors arranged around the peripheral enclosure. In FIG. 1B, three pairs 5i, 5 ₂ and 5 _{3 are shown} of electromagnetic inductors. The upper part 5s of the inductors is positioned at the level of the free surface l _sup of the liquid alloy. Each inductor has a 90 ° phase shift between the upper part 5s and the lower part 5i. In the examples described below, a device as described in the application was used

WO2014 / 155357, and more precisely according to the configuration described in connection with FIGS. 19 and 20A, in which three inductors, oriented along the vertical axis Z, are arranged facing each large face of the ingot. By large face of an ingot is meant a face extending along the longitudinal axis Y and the vertical axis Z. Each inductor comprises one or more coils. In this example, each coil is arranged at a distance of 185 mm from the mold. In general, the distance between a coil of an inductor and the ingot mold can be between 130 mm and 200 mm.

The sliding magnetic field can also be generated from one or more permanent magnets placed on the periphery of the mold and set in motion relative to the latter. For example, it is possible to generate a sliding magnetic field by rotating a permanent magnet.

The distance λ separating two maxima of amplitude of the magnetic field is the wavelength of the sliding magnetic field. FIG. 1C represents an example of the distribution of the amplitude B _o of a magnetic field sliding along a propagation axis Δ at an instant t (solid line), and at an instant t + At (dotted line). On the axis of propagation, there is shown a coordinate r corresponding to the position of a point of the marsh. FIG. 1D illustrates a temporal evolution of an alternating magnetic field sliding at this point. This evolution is periodic, and takes place according to a period P. The application of a sliding magnetic field results, at a point in the swamp, by a periodic variation in its amplitude. Thus, the amplitude of the magnetic field at a point in the marsh varies as a function of time, between a minimum amplitude B ™ ⁱⁿ and a maximum amplitude B ™ ^ax .

The Lorentz force, at a point of coordinates r of the marsh, has an oscillating component, modulated at a frequency 2 / double the frequency of the magnetic field. The amplitude F _o of the oscillating Lorentz force density can be explained by the expression:

F ₀ (r) = -σ / ΛΒθζΓ) (1), where σ denotes the electrical conductivity.

The amplitude of the Lorentz force, at a point r of the marsh depends on the square of the amplitude of the magnetic field applied at this point.

The inventors have found that the appearance of intermittent macro-segregation 11 can be limited by adjusting the electromagnetic stirring when the average Lorentz force applied to the liquid alloy 1 £ flowing at the edge 10, has a certain orientation, and this in a median zone of the marsh, extending symmetrically on either side of the median plane M, between T / 2 - T / 4 and T / 2 + T / 4. The thickness of the middle zone M corresponds to half the thickness of the ingot. By mean Lorentz force is meant an average of the Lorentz force during a period P of the magnetic field. The period P of the magnetic field corresponds to the time interval separating two successive maxima or minima of the magnetic field, as shown in FIG. 1D. The period P corresponds to the inverse of the frequency f. The inventors observed that in the middle zone, at the interface of the swamp 1 £ and the solid alloy ls, at the level of the front 10, the angle β formed by the mean Lorentz force F, with respect to the vertical , must advantageously be less than the angle a of the front, mentioned above, corresponding to the angle between the tangent to the front and the vertical, the angles a and β being oriented in the same direction. In other words, it is advantageous that the direction of the average Lorentz force F is more vertical than the direction of the tangent to the front. Thus, in the middle zone, at the interface between the marsh and the front, the average Lorentz force F is oriented towards the solid alloy ls, and not towards the liquid alloy 1 £. This condition is illustrated in Figure 2. In this figure, there is shown a section of a flow along an XZ plane. The position of the median plane M corresponds to the thickness T / 2.

The plating effect of the liquid alloy 1 £ against the front 10 is obtained at the interface between the liquid alloy and the front 10. Preferably, this effect is obtained in a layer, called the interface layer, adjacent of the forehead, the thickness of which is less than 2cm, or 1cm or 5mm. The thickness is defined in a direction perpendicular to the front. It is indeed in such a layer that the liquid alloy, in contact with the cold isotherm formed by the forehead, becomes locally denser. A fluid layer is then formed along the front, in which the flow of the liquid alloy is accelerated, and can move away from the front, leading to the appearance of vortices. It is mainly in this layer that it is necessary to apply a Lorentz force pressing the liquid alloy against the forehead, in order to maintain the liquid alloy against the forehead, so as to limit the formation of intermittent macro-segregations .

Under these particular conditions, the Lorentz force F tends to press the liquid alloy 1 £ of the marsh against the front 10, which limits the formation of intermittent macro-segregations. The Lorentz force is said to be tackling. It allows the formation of a convective laminar flow along all or part of the front 10, limiting the appearance of intermittent macro-segregations.

As described below, the phenomenon of plating of the liquid alloy by the Lorentz force against the front 10 is all the more marked when the casting speed V and the frequency f are low.

A person skilled in the art knows how to model the orientation of an average Lorentz force F, exerted over a period, in the marsh. Calculation codes, for example the AC / DC module of the COMSOL code, allow such modeling, based in particular on the characteristics of the inductors (dimensions, number of ampere-turns, pole pitch, positioning relative to the ingot mold) , the geometry of the ingot mold and operational parameters such as the casting speed or the frequency of the magnetic field. The simulations make it possible to model the electromagnetic stirring of the liquid alloy and to estimate a temporal evolution of the force of Lorentz F, at any point of the marsh, during a period. By evolution, we mean both the evolution of the intensity and the evolution of the direction. It is then possible to determine the orientation and intensity of the average Lorentz force applying at a point in the swamp, during a period P of the magnetic field.

FIGS. 3A, 3B, 3C and 3D show the orientation of the average Lorentz force, obtained by simulation, at different points of a front 10. In these figures, part of a front 10 is shown, according to a plane XZ, extending between the median plane M (abscissa x = 0) and a wall of the ingot mold (abscissa x = 0.26). The abscissa axis represents a position along the transverse X axis and the ordinate axis represents a position along the vertical Z axis. The frequencies considered are equal to 5Hz (Figure 3A), 1 Hz, 0.5 Hz and 0.2 Hz (Figure 3D), the casting speed being 55 mm / min. We observe that by considering the same position on the front 10, the lower the frequency f, the smaller the angle β of the average Lorentz force F. Thus, in the same position on the forehead, the average Lorentz force F tends to tilt vertically as the frequency decreases.

Furthermore, as previously described, the Lorentz force is tackling when the angle β of the average Lorentz force F is less than the angle a of the front. There is shown, in each figure, a range of thickness Δχ, extending from the median plane M, in which the plating force effect is obtained. This width range is indicated by a double arrow. It is observed that the more the frequency decreases, the more the thickness range Δχ extends, from the median plane M (x = 0), corresponding to the thickness T / 2, towards the wall of the mold. The technical effect of minimizing intermittent macro-segregation appears in this thickness range Δχ, and it is preferable for it to be as wide as possible, preferably including the thickness range T / 2.3 -T / 3.3, the latter being generally conducive to the formation of intermittent macro-aggregations. In these figures, the thickness range T / 2.3 - T / 3.3 corresponds to the interval between x = 0.03 m and 0.1 m. The T / 4 coordinate corresponds to x = 0.13 m.

Figures 3E, 3F, 3G and 3H respectively show the average orientation of the Lorentz force, obtained by simulation, at different points of a front 10, the frequencies being respectively equal to 5Hz, 1 Hz, 0.5 Hz and 0.2 Hz .

In FIGS. 3A to 3H, so-called normalized forces have been shown, each force being normalized by its intensity, so as to better show the evolution of the orientation of the average Lorentz force on the front 10, as a function of position along the forehead.

The comparison of FIGS. 3A to 3H shows that the lower the frequency, the greater the range of thickness Δχ according to which the Lorentz force becomes tackling is large. The thickness range Δχ extends from the median plane M towards the wall of the mold, along the transverse axis X. It widens as the frequency decreases. Furthermore, at the same frequency, the lower the casting speed, the greater the thickness range according to which the Lorentz force is tacky. It is therefore advantageous to favor both a low frequency f, preferably less than 2 Hz, or even 1 Hz, and a low casting speed, preferably less than 45 mm / min, or even 40 mm / min.

On the basis of simulations such as illustrated in FIGS. 3A to 3H, the inventors determined an evolution, along the transverse axis X, of an angle Θ, called differential angle, representing a difference between, at the same point, the angle of the front a and the angle of the Lorentz force β, that is θ = α - β. Recall that the angles a and β are oriented in the same direction. When θ> 0, a> β: the Lorentz force is more inclined than the tangent to the front. It is therefore tackling.

FIGS. 4A and 4B show the evolution of the differential angle θ as a function of a position x on the front 10, along the transverse axis X. The abscissa axis represents the position x, expressed in meters, on the forehead along the transverse axis. As in FIGS. 3A to 3H, the coordinate x = 0 corresponds to the position T / 2, the coordinate x = 0.26 corresponding to the wall 2p of the ingot mold. Figures 4A and 4B were obtained by considering respectively a casting speed of 55 mm / min and 35 mm / min. In each figure, the simulations of the orientation of the average Lorentz force F were carried out by successively considering several frequencies f, between 5 Hz and 0.2 Hz. In each figure, we have shown, in mixed horizontal lines, lines corresponding to the values θ = 0 ° and θ = 4 ° when the frequency is equal to 1 Hz. The Lorentz force is tackling when Θ> 0, but the inventors consider that it is advantageous that Θ> 4 °. We can thus define, on each configuration, a thickness range, in which the Lorentz force is tackling, from the median plane M (thickness T / 2).

FIGS. 4A and 4B show the thickness ranges Δχ (θ = 0 °) and Δχ (θ = 4 °) for f = 1 Hz. Similarly to FIGS. 3A to 3H, it is observed that the thickness ranges are all the more important the lower the frequency and the lower the casting speed. Optimal results are obtained for f <2 Hz, even f <1 Hz, and when the casting speed is 35 mm. The lower the frequency, the greater the intensity of the Lorentz force applied to the liquid alloy bordering the front, in the thickness range T / 2 - T / 4. This strengthens the intensity of the Lorentz force and increases the desired technical effect. In other words, to obtain a significant reduction in intermittent macro-segregation, an orientation of the Lorentz force as described above is necessary. However, its intensity must be sufficient to obtain a plating of the liquid alloy 1 £ against the front 10. This is why it is preferable to modulate the magnetic field according to a relatively low frequency f.

Using simulations taking into account different casting thicknesses, the inventors established an abacus, represented in FIG. 4C, making it possible to define an operating range for which the Lorentz force is considered to be sufficiently plating, that is to say ie when the differential angle θ is greater than or equal to 4 °. This abacus is the subject of FIG. 4C. The abscissa and ordinate axis of the abacus correspond respectively to the casting speed V and to the thickness T of the ingot. The thickness being determined, the abacus makes it possible to define the casting speed and the maximum frequency making it possible to place oneself in the conditions of implementation of the invention. On this chart, we have materialized, by crosses, the experimental test conditions described below, in connection with Table 1.

This chart depends on the number and characteristics of the inductors, their positioning relative to the mold, the dimensions of the latter and the operational parameters of the installation, in particular relating to the magnetic field applied. A person skilled in the art, knowing the characteristics of the installation, can carry out simulations aiming to obtain the orientation of the average Lorentz force F at different points along the front 10, along the transverse axis X. He can then determine a frequency range and a casting speed range for which we obtain θ> 0 °, or advantageously θ> 4 °, so as to implement the invention and obtain the desired technical effect, that is that is to say a limitation of the intermittent macro-aggregation between T / 2 and T / 4, and more particularly between T / 2.3 and T / 3.3.

Experimental tests have been carried out, using alloys of type 7010 and 7035, the thickness T of the ingot being equal to 525 mm. Each pour was made using a fixed frequency. Between the different castings, the frequency f was varied between 0.250 Hz and 0.850 Hz. Each casting was carried out using a fixed casting speed. Between the different castings, the casting speed was varied between 35 mm / min and 55 mm / min.

Table 1 summarizes the experimental test conditions, five of which implement the invention. Tests 1 to 3 were carried out using an alloy 7010, while tests 4 to 10 6 were carried out using an alloy 7035. The parameters of each test are the frequency f and the casting speed V During each test, an ingot of format 1650 mm x 525 mm was produced.

The Ref 7010 and Ref 7035 tests are reference tests, carried out without electromagnetic stirring.

Trial Alloy V (Mm / min) f (Hz) ACmeso (W / 4) Fourier (W / 4) ACmeso (W / 2) Fourier (W / 2) Ref 7010 7010 45 0.82 0012 0.71 1 7010 40 0475 0.44 0005 2 7010 40 0850 0.50 0005 3 7010 45 0250 0.56 0006 Ref 7035 7035 35 0.59 0012 0.75 0014 4 7035 35 0475 0.32 0005 0.41 0006 5 7035 35 0.270 0.46 0007 6 7035 55 0475 0.59 0013 0.51 0008

Table 1

Each of these tests is represented by a cross on the abacus of FIG. 4C. It can be seen that tests 1, 2, 3, 4 and 5 are considered to implement the invention if one refers to the abacus shown in FIG. 4C. Under these conditions, a plating Lorentz force is obtained between T / 2 and T / 4. Test 6 is outside the invention, because given the thickness of the ingot and the casting speed, the abacus indicates a maximum frequency of 0.1 Hz, while the frequency used during this test is raised to 0.475 Hz.

The ingots formed were characterized by analyzing horizontal profiles (along the TC axis) of an X-ray taken at half-width along a vertical L / TC plane, these profiles being calibrated to obtain the spatial distribution of alloying elements. heavy Zn and / or Cu type. For alloy 7010, the alloying elements considered are Zn and Cu. For alloy 7035, the alloy element considered is Zn. The terms L and TC, known to those skilled in the art, correspond to the dimension of the ingot along the vertical axis and along the axis known as "short cross". An intermittent macro-segregation can be characterized by a maximum deviation in mass of an alloying element in the zone most marked by intermittent macro-segregation, that is to say between T / 2.3 or T / 3, 3. As previously indicated, T / n denotes a distance from an edge of the ingot, along a horizontal axis, T / 2 corresponding to the center of the ingot.

To quantify the intermittent macro-segregation, the concentration profiles were treated as illustrated in FIG. 5A. The profile obtained with a resolution of 0.1 mm is the raw profile referenced profile A, and shown in dotted lines in FIG. 5A. A first smoothing is carried out, according to a sliding average over 2 mm, the smoothed profile obtained being referenced profile B, represented by a solid line in FIG. 5A. This smoothing makes it possible to attenuate the effect of micro-segregations, which correspond to local fluctuations in concentrations.

A second smoothing of the raw profile is carried out, according to a sliding average of 50 mm, to overcome intermittent macro-segregation, and obtain a central continuous segregation profile, or basic profile, referenced profile C. This profile is represented by dashes in Figure 5A. The basic profile C is subtracted from the smoothed profile B to obtain a so-called corrected profile D, representative of the intermittent macro-segregation. The latter is shown in Figure 5B. As can be seen in FIG. 5B, the corrected profile D is mainly representative of the intermittent macro-segregation, and is not or little affected by the central continuous macro-segregation and by the micro-segregations. Intermittent macro-segregation is characterized by determining a concentration deviation AC _meso = C _max - C _min . C _max and C _min respectively denote the maximum and minimum concentrations measured on the corrected profile. It is considered that a value of & C _meso less than 75% of a reference value, obtained on a cast ingot without stirring, constitutes a satisfactory result, that is to say a significant reduction in intermittent macro-segregation.

The concentration deviation kC _meso was measured at W / 4 and / or at W / 2, that is to say along planes perpendicular to the median plane, parallel to the transverse axis X and whose coordinate, according to l longitudinal axis Y, is respectively equal to W / 4 and W / 2.

The corrected profile D was normalized by the nominal concentration of the alloying element considered (Zn and Cu for the alloy 7010, Zn for the alloy 7035). The profile thus normalized was analyzed by Fourier transform, so as to identify the spatial period characterizing the intermittent macro-aggregation. When the intermittent macro-aggregation is large, one or more amplitude peaks are generally observed in the range 8 mm - 25 mm. We determine an adimensional criterion of spectral intensity ζ which corresponds to the maximum amplitude of the Fourier components in a range of spatial period between 8 and 25 mm. The products obtained by the process according to the invention preferably have a criterion ζ less than 0.01, preferably less than 0.007 and preferably less than 0.005. The criterion ζ was measured at W / 4 and / or at W / 2. FIGS. 5C and 5D respectively show an example of a distribution of spatial periods between 0 and 30 mm for Examples 4 and 6, over several profiles. In these figures, the ordinate axis represents the spectral intensity, while the abscissa axis represents the spatial period, expressed in mm.

On ingots cast without stirring, corresponding to the reference measurements, the criterion ζ is greater than 0.01, with typical values of 0.012 to W / 4 or 0.014 to W / 2.

Tests 1, 2, and 3 were obtained using an alloy 7010. The casting speed was 40 mm / min for tests 1 and 2 and 45 mm / min for test 3. The frequency f was respectively equal to 0.475 Hz, 0.850 Hz and 0.250 Hz. On each of these tests, a difference in concentration kC _meso was obtained which was significantly lower than the reference measurements:

for tests 1 and 2, kC _meso at W / 2 is 0.44 and 0.50 respectively, the reference value being 0.71;

for test 3, kC _meso at W / 4 is 0.56, the reference value being 0.82.

On each of these tests, the spectral intensity criterion ζ in the spatial period range between 8 and 25 mm is systematically less than 0.001.

Tests 4, 5, and 6 were obtained using an alloy 7035. The casting speed was 35 mm / min for tests 4 and 5 and 55 mm / min for test 6. The frequency f was respectively equal to 0.475 Hz, 0.270 Hz and 0.475 Hz. On tests 4 and 5, a difference in concentration kC _meso was obtained which was significantly lower than the reference measurements:

for tests 4 and 5, AC _meso at W / 4 is 0.41 and 0.46 respectively, the reference value being 0.75;

for test 4, AC _meso at W / 2 is 0.32, the reference value being 0.59.

For tests 4 and 5, the spectral intensity criterion ζ in the spatial period range 5 between 8 and 25 mm is less than 0.001.

Test 6, being outside the invention, does not make it possible to obtain a reduction in intermittent macro-aggregations. The values obtained are comparable to the reference values.

The invention may be implemented for the production of ingots intended for components for which the quality requirements are high, for example components related to applications in the aeronautical field.

Claims (11)

1. Method for forming an aluminum alloy ingot (1) in an ingot mold (2), the ingot mold defining a parallelepiped, such that the ingot formed extends parallel to a longitudinal axis (Y), according to a width (W) and parallel to a transverse axis (X), along a thickness (T), the thickness being less than the width, the ingot defining a median plane (M) extending along the mid-thickness (T / 2), parallel to the longitudinal axis (Y), the method comprising the following steps: preparation of the liquid aluminum alloy; pouring the liquid aluminum alloy into the ingot mold, along a vertical axis (Z), the alloy being cooled, during casting, by a trickle of coolant (3), so as to form a solid alloy (ls), extending around a liquid alloy (1 €), called swamp, the solid alloy forming a front (10) at the interface with the swamp (IC), the front being inclined according to a angle of inclination (a), with respect to the vertical axis, the angle of inclination of the front varying along the transverse axis (X); displacement of the solid alloy (ls), along the vertical axis (Z), according to a casting speed (V); during casting, application of a sliding magnetic field whose amplitude is periodically varied according to a frequency (/), the magnetic field being generated by at least one magnetic field generator arranged at the periphery of the mold, so applying a Lorentz force (F) at different points in the swamp, the average Lorentz force (F), during a period (P) of the magnetic field, being inclined relative to the vertical axis (Z) at an angle d 'tilt (β), said tilt angle of the Lorentz force, the latter varying along the transverse axis; the sliding magnetic field propagating along an axis of propagation parallel to the vertical axis; the marsh comprising a median zone, extending symmetrically on either side of the median plane (M), the thickness (T) of which corresponds to half the thickness of the ingot; the method being characterized in that: the frequency (/) is less than 5 Hz; and in that the casting speed (V) and the frequency (f) are adapted so that throughout the median zone of the marsh, at the interface between the liquid alloy and the solid alloy, at the front (10), the angle of inclination (β) of the force is strictly less than the angle of inclination (a) of the front; so that in the middle zone, the Lorentz force (F) acts on the liquid alloy (1 €), at the interface, by pressing it against the forehead (10). 2. Method according to claim 1, in which the angle of inclination (β) of the average Lorentz force is less, by at least 4 °, than the angle of inclination (a) of the front, such so that the average Lorentz force (F) is more inclined than the front (10). 3. Method according to any one of the preceding claims, in which the frequency (f) is less than 2 Hz or less than 1 Hz. 4. Method according to any one of the preceding claims, in which the casting speed (Vj is less than 45 mm / minute. 5. Method according to any one of the preceding claims, in which the casting speed (7) and the frequency (/) are adapted so that in the median zone of the marsh, in an interface layer between the alloy liquid (1 €) and the front (10), the angle of inclination of the Lorentz force (β) is strictly less than the angle of inclination (a) of the front, the interface layer having a thickness , in a direction perpendicular to the front, is less than 2 cm. 6. Method according to any one of the preceding claims, in which the mean Lorentz force value (F) is constant during several successive periods. 7 Method according to any one of the preceding claims, in which the aluminum alloy is chosen from alloys of type 2XXX, 6XXX or 7ΧΧΧ. 8. Method according to any one of the preceding claims, in which the thickness of the casting is greater than 300 mm. 9. Method according to any one of the preceding claims, comprising, prior to casting, a modeling of the Lorentz force (F) applying at at least one point on the front (10), so as to define, taking into account the thickness (T) of the ingot mold, a frequency value (/) and / or a casting speed value (V) allowing an average Lorentz force (F) to be obtained, the angle of which d 'inclination (β) relative to the vertical, is less than the angle, at said point, formed by the front (10) relative to the vertical. 10. The method of claim 9, wherein the modeling defines a frequency value and / or a casting speed value for obtaining an average Lorentz force (F) whose angle of inclination (β) , relative to the vertical, is 4 ° lower than the angle of inclination (a) formed by the front (10) relative to the vertical. 5 11. An aluminum alloy ingot obtained by a process which is the subject of any one of claims 1 to 10.