WO2017207886A1 - Method for producing sheet ingots by vertical casting of an aluminium alloy - Google Patents

Abstract

The invention is a method for casting a metal alloy, and in particular an aluminium alloy, in an ingot mould extending along a vertical axis, the horizontal section of the ingot mould being parallelepiped in shape. During casting, a travelling alternating magnetic field is applied to a liquid phase of the alloy, the magnetic field having a maximum amplitude propagating along an axis of propagation. Under the effect of the magnetic field, a Lorentz force is applied to said liquid phase of the alloy, such that a Lorentz force of maximum intensity propagates along the axis of propagation. The method according to the invention comprises modulating said maximum intensity of the Lorentz force propagating along the axis of propagation. This modulation is obtained by varying, over time, one or more parameters, referred to as force parameters, governing the Lorentz force. The invention also concerns an ingot obtained by the method. The invention makes it possible to obtain a significant reduction in instances of macrosegregation, referred to as intermittent macrosegregations, the latter giving rise to a non-uniform spatial distribution of alloy elements in the ingot. The ingot obtained is therefore more uniform.

Description

 METHOD FOR MANUFACTURING VERTICAL CASTING ROLLS OF A

 ALUMINUM ALLOY

Technical area

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

Presentation of the invention

 During a vertical casting, aimed at forming an ingot, the solidification of a metal or a metal alloy is affected by so-called macroscopic segregation phenomena. During the cooling of the metal, convection currents form, 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 solidification times. These phenomena lead, in the solidified ingot, to local depletion or local enrichment of 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 alloying elements, along a vertical central axis of the ingot. These macrosegregations have been described in the work of John Wiley and the "Direct-Chili Casting of light alloys", Publisher Wiley, September 2013, pp 158 - 172. The main mechanisms at the origin of the central macrosegregation described in this book are

Thermosolutal convection in the marsh caused by gradients of temperature and concentration, 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 solidification volumetric shrinkage, which may be assisted by the metallostatic pressure;

The flow of liquid in the pasty zone caused by mechanical deformations; • The forced flows that may result from lodging, injection or gas evolution, mixing, vibration, etc. which enter the supercooled zone and the pasty zone and change the direction of the convection movements. It is a continuous macrosegregation, this term designating the fact that the macrosegregation takes place continuously on 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 macrosegregation has been less often described in the literature and results in the formation of V-shaped bands on both sides of the negative central macrosegregation. These V-shaped bands are alternately enriched and depleted in eutectic and peritectic alloying elements. These bands are observable by X-ray radiographs of vertical slices of ingots, typically in the L / TC plane at mid-width, when the segregated elements absorb X-rays in a manner differentiated from the atoms of the metal composing the ingot. Other means make it possible to visualize this phenomenon, for example ultrasonography or observation with the naked eye of anodized vertical slices, because of the difference in optical reflectivity between the zones enriched or depleted of alloying elements. Generally, the intermittent macrosegregation is most marked at the T / 2.5 region of the thickness, with 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, denotes a region located at a distance T / n from an edge of the ingot, where T denotes a thickness of the ingot.

Periodic intermittent macrosegregations appear very early 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 aluminum alloys loaded with aluminum alloys, typically cast in thicknesses greater than 300mm, this thickness threshold itself depends on the casting speed.

The publication C. Dorward et al. Aluminum, 1996, 72. Jahrgang, 4, p. 251-259 describes the formation of intermittent segregation bands in an alloy 7000. According to these authors, This phenomenon is due to grain avalanches triggered periodically by convective oscillations of the marsh, that is to say the liquid phase of the metal, in connection with a vortex emission mechanism. In particular, this article shows that intermittent macrosegregation can cause variations in mechanical properties, such as toughness, on sheets obtained from raw casting products. It is therefore advantageous to find a casting process that would eliminate these intermittent macrosegregations.

The reduction or elimination of continuous macrosegregations, for example central macrosegregation, has already been described. In particular it has been shown that the application of a magnetic field, for purposes of stirring or braking of the flows, made it possible to limit the appearance of continuous macrosegregations. The document US5375647 describes for example a central macrosegregation reduction process occurring during the casting of a metal alloy ingot. This method comprises the application, during cooling, of a static magnetic field generated by at least one coil traversed by a direct current. Document F 2530510 describes a process for the electromagnetic casting of metals in which a stationary magnetic field and a variable frequency magnetic field are simultaneously made to produce both radial vibrations within the metal that has not yet solidified and to limit mixing. .

B. Zhang et al. "Effect of low-frequency magnetic field on macrosegregation of continuous casting aluminum alloys" Materials Letters 57 (2003) ppl707-1711 applied a low-frequency variable magnetic field (between 10 and 100 Hz) to a billet of 200 mm AA7075 alloy and found a beneficial effect on the decrease of central macrosegregation, mainly for a frequency of 30 Hz.

EP 2682201 describes a method of electromagnetic stirring using two inductors mounted symmetrically with respect to each other with respect to the vertical plane of symmetry of an ingot mold. These inductors generate two electromagnetic fields of different frequencies propagating in opposite directions along a vertical axis. At least one of the inductors generates a magnetic field at a resonant frequency of the liquid metal.

WO 2014/155357 relates to methods and apparatus for moving a molten metal, the electromagnetic inductor comprising at least two pairs of electromagnetic poles and a first magnetic field component being generated between a pole in a first pair of electromagnetic poles and a second pole in a different pair of electromagnetic poles, and a second magnetic field component being generated between the two poles in one or more pairs of electromagnetic poles, the second magnetic field component thereby generating one or more eddy currents in the molten metal . The inventors have considered that the processes described above do not make it possible to effectively reduce the appearance of intermittent macrosegregations. They propose a method making it possible to limit the formation of such macrosegregations, or even to eliminate them, so as to better control the mechanical properties of the products resulting from casting. Presentation of the invention

An object of the invention is a method for casting an ingot of aluminum alloy in a substantially rectangular mold comprising the following steps:

 preparation of the aluminum alloy;

 pouring the aluminum alloy into the mold, along a vertical axis of flow, the alloy being cooled, during casting, by a flow of a coolant liquid in contact with the solidified metal;

 during the casting, applying a magnetic field whose amplitude is periodically varied according to a frequency, said magnetic field being generated by at least one magnetic field generator disposed at the periphery of the mold, so as to apply a force of Lorentz at different points of a liquid part of the alloy being solidified;

 the magnetic field applied being a sliding magnetic field, propagating along a propagation axis, such that a maximum amplitude of the magnetic field propagates along said propagation axis, by defining a propagation wavelength, said magnetic field sliding causing propagation, according to said axis of propagation, a Lorentz force of maximum intensity;

the method being characterized in that a magnetic force parameter, governing a Lorentz force value of maximum intensity, is variable in a predetermined time interval, said parameter being:

 Said maximum amplitude of the magnetic field;

 and / or said frequency of the magnetic field;

 and / or the propagation wavelength of the magnetic field;

so as to obtain a modulation, in said time interval, of said Lorentz force of maximum intensity propagating along the axis of propagation. The method may include any of the following features, taken alone or in combination: the section of the mold, in a horizontal plane, defines a thickness and a length, the thickness being less than or equal to the length, the thickness being greater than 300 mm and preferably at least 400 mm;

 the frequency of the magnetic field is less than 5 Hz, or 2 Hz or 1 Hz;

the Lorentz force of maximum intensity, propagating along the axis of propagation, varies by at least 30 Nm ^"3 in a time interval of between 20 seconds and 10 minutes;

the magnetic field is such that the absolute value of the change in the density of the maximum Lorentz force is greater than or equal to 0.05 Nm ^-3 s ¹ during said time interval;

 the axis of propagation of the maximum amplitude of the magnetic field belongs to a plane parallel to the direction of casting;

 during casting, the variation of the force parameter is periodic, the period being between 20s and 20 minutes, or between 1 minute and 15 minutes, or between 2 minutes and 10 minutes;

 during the casting, the Lorentz force of maximum intensity is not equal to zero, during the casting, the variation of the force parameter is not obtained by a periodic interruption of the sliding field.

 the adimensional number of Hartmann, in at least one point of the liquid part of the alloy, varies by at least a factor of 3, or even a factor of 5, in said time interval;

 the aluminum alloy is chosen from alloys of types 2XXX, 6XXX or 7XXX, the thickness being at least 400 mm or 450 mm.

According to one embodiment, the generators are electromagnetic inductors, each electromagnetic inductor being traversed by a current called induction current. The method comprises, during said time interval:

 a variation of an intensity of the induction current;

 and / or a variation of a frequency of the induction current;

 and / or a variation of a distance between an electromagnetic inductor and the mold.

According to this embodiment, the method may comprise a variation of the intensity or frequency of the induction current flowing through an inductor, the method then comprising: a preliminary step of defining at least one critical value of the intensity and the frequency of the induction current generating, at a free surface of the aluminum alloy flowing in the mold, a wave resonance;

 determining a range of variation of the intensity or frequency of the induction current as a function of said previously defined critical value.

The method may include a definition of a plurality of critical values of the intensity and frequency of the induction current, so as to define a resonance curve, representing the critical intensity and frequency values generating a resonance of said free surface, the method comprising determining a range of variation of the intensity or frequency of the induction current in a range delimited by said resonance curve.

Preferably, the method comprises a variation of the frequency of the induction current flowing through an inductor.

 According to one embodiment, at least one generator is a permanent magnet, the method comprising:

 a variation of a distance between the permanent magnet and the mold;

 and / or a rotation of the permanent magnet, and a variation of the rotational speed of the magnet;

 and / or a rotation of two permanent magnets. Another object of the invention is an aluminum alloy ingot obtained by the method as described above and in the description which follows.

The ingot may have, for an element of the alloy, whose weight content is greater than

0.5%, or the sum of two elements of the alloy whose individual content is greater than

0.5%, a dispersion criterion of less than 3.3, preferably less than 3, more preferably less than 2.5, still more preferably less than 2 and most preferably less than

1.5, said dispersion criterion being defined according to the following expressions:

e = C _ZA / AC _Z "(6)

ACZA = max (C _ZA ) - min (C _ZA ) (4),

ACZR = max (C _ZR ) - min (C _ZR ) (5),

or :

max (C _ZA ) and min (C _ZA ) denote respectively the maximum and minimum concentrations of the element considered or the sum of the two elements considered measured in an analysis zone, exhibiting intermittent macrosegregations, for example between T / 2.3 and T / 3.3;

max (C _Z R) and min (C _Z R) denote respectively the maximum and minimum concentrations of the element considered or the sum of the two elements considered in a reference zone considered as being little affected by intermittent macrosegregations, for example between T / 6 and T / 12;

said concentrations being measured on at least one profile established at half width in a vertical plane L / TC and in the direction TC, said profile being representative of said intermittent macrosegregations in said direction TC. The ingot may have a spectral intensity criterion of less than 0.01, preferably less than 0.007 and preferably less than 0.005, said spectral intensity criterion being calculated in:

 determining a maximum amplitude of a Fourier transform of a profile representative of an intermittent macrosegregation of an element whose weight content is greater than 0.5% or the sum of two elements of the alloy whose individual content is greater at 0.5%, the profile being established in said direction TC, said maximum amplitude being determined in a range of spatial periods between 8 and 25 mm,

 normalizing said maximum amplitude by a nominal concentration Co of said element or by the sum of the nominal concentrations of the two elements considered.

Other advantages and features will emerge more clearly from the following description of particular embodiments of the invention, given by way of non-limiting examples, and shown in the figures listed below.

figures

FIGS. 1A to 1E illustrate an exemplary device and method according to the prior art and according to the invention. FIG. 1A shows the main components of the device whereas FIGS. 1B and 1C respectively represent a spatial and temporal distribution of the amplitude of a sliding magnetic field according to the prior art. FIGS. 1D and 1E respectively show a spatial and temporal distribution of the amplitude of an unsteady sliding magnetic field according to embodiments of the invention.

FIG. 2 represents a so-called free-surface resonance curve of the marsh, representing values, called critical, of the intensity and the frequency of an induction current to which a resonance of the free surface of the marsh appears, this by implementing an electromagnetic stirring process.

FIG. 3 is a radiography of a vertical slice of a product obtained by implementing a first exemplary method, representative of the prior art, according to a first example, said example 1, representative of the prior art.

Figure 4 shows an example of a Zn concentration profile along a horizontal line of the vertical slice shown in Figure 3 and the analysis and reference areas.

FIG. 5A shows the digital processes successively carried out on each profile obtained with a resolution of 0.1 mm. Figure 5B shows a profile resulting from the treatments performed.

FIGS. 6A and 6B illustrate characterization profiles of a product obtained by implementing a method according to example 1. FIG. 6A shows Zn concentration profiles along several horizontal lines of the vertical slice represented on FIG. Figure 3. Figure 6B shows the profiles resulting from the digital processing performed. Figure 7 shows Fourier transforms of the profiles shown in Figure 6B.

FIG. 8 represents a so-called free-surface resonance curve of the marsh, obtained by implementing a method of a second example, said example 2, according to the invention.

FIGS. 9, 10A, 10B and 11 illustrate a characterization of a product obtained by implementing a method according to this second example. Figure 9 is an X-ray of a vertical slice of the product. Fig. 10A shows Zn concentration profiles along several horizontal lines of the vertical slice shown in Fig. 9. Fig. 10B shows the profiles resulting from digital processing performed on the profiles shown in Fig. 9. Fig. 11 shows Fourier transforms of these different profiles.

DETAILED DESCRIPTION OF THE INVENTION Unless stated otherwise, all the information concerning the chemical composition of the alloys is expressed as a percentage by weight based on the total weight of the alloy. The expression 1.4 Cu means that the copper content expressed in% by weight is multiplied by 1.4. The designation of alloys is in accordance with the regulations of The Aluminum Association, known to those skilled in the art. FIG. 1A illustrates an exemplary casting method known from the prior art. In this example, an aluminum alloy 1 flows into an ingot mold 2, through an opening 2i. The mold 2 extends along a vertical axis Z. It is delimited by a peripheral enclosure whose section, in a horizontal plane XY, is parallelepipedic. A cooling fluid 3, for example water, flows against the wall of the solidified product. This process is known as semi-continuous direct cooling casting ("Direct-Chili Casting"). A false bottom 4 can be translated so as to move away from the opening 2i during the casting. The mold 2 extends, parallel to a first horizontal axis X, with a thickness e and, parallel to a second horizontal axis Y, perpendicular to the axis X, along a length β. The thickness e is for example greater than 300 mm. It is beyond such a thickness that the intermittent macrosegregations 11 appear markedly.

Under the effect of cooling, a solid zone ls forms, near the cooled enclosure, around a liquid zone 1β, designated by the term "swamp". The interface between the liquid zone 1β and the solid zone ls is a front 10, the latter progressing towards the center of the mold as the solidification of the alloy takes place. After cooling, a parallelepiped ingot, also referred to as "product", is formed.

The alloy is an aluminum alloy of the 1XXX, 2XXX, 3XXX, 4XXX, 5XXX, 6XXX, 7XXX or 8XXX series. Alloys whose mass fraction of alloying elements is greater than 1%, or even greater than 3% or even 5%, are particularly suitable for a process according to the invention, because the higher the mass fraction of these alloying elements. important, the more intermittent macrosegregations are marked. The invention is particularly advantageous for products in alloy 2XXX, 5XXX, 6XXX or 7XXX whose thickness is at least equal to 400 mm or 450 mm.

There is shown a magnetic field generator 5, capable of generating a magnetic field B intended to be applied to the liquid zone 1β of the alloy. Such a generator may be a permanent magnet or an electromagnetic inductor, the latter generating a magnetic field when traversed by an electric current, said induction current.

The magnetic field B applied to the liquid zone 1β is an alternating field of amplitude B ₀ and frequency /. The effect of this magnetic field is to apply a mixing of the marsh, under the effect of Lorentz forces applying on the liquid metal zone 1β. Indeed, the application of a magnetic field B generates, in the alloy, the formation of an electric current / resulting, within the liquid zone of the alloy subjected to the magnetic field, in the appearance a Lorentz force F such that F oc / x B where x denotes the vector product operator, and oc denotes a proportionality relation. This Lorentz force has an oscillating component at a frequency twice the frequency / magnetic field.

Because of the thickness of the mold, the frequency / is chosen so as to allow sufficient penetration of the magnetic field B in the marsh, so as to obtain an efficient stirring of the liquid. The frequency / is all the lower as the thickness of the product is high. 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 1 Hz.

The generator 5 is able to generate a sliding magnetic field. The term "sliding magnetic field" designates an alternating magnetic field, whose amplitude B ₀ is not constant, and varies between a minimum value and a maximum amplitude B ™ ^ax , the maximum amplitude gm ^a x _is p _r0 p _a g _ean t _se | on a propagation axis Δ, preferably rectilinear. By amplitude, we mean the maximum value taken by a periodic quantity. Preferably, the propagation axis belongs to a plane parallel to the direction of casting. The distance λ separating two maxima of amplitude of the magnetic field is the wavelength of the sliding magnetic field. FIG. 1B represents an example of distribution of the amplitude B ₀ of a magnetic field sliding along a propagation axis Δ at a time t (solid line), and at a time t + At (dashed line). On the axis of propagation, there is shown a coordinate r corresponding to the position of a marsh point. Figure 1C illustrates a time evolution of an alternating magnetic field sliding at this point. Due to the propagation of the maximum amplitude value B ™ ^ax , the amplitude of the magnetic field at this point varies between a minimum value B ™ ⁱⁿ and the value B ™ ^ax, the latter not changing over time. .

A sliding magnetic field generator 5 may consist of several electromagnetic inductors arranged around the peripheral enclosure. The Lorentz force, at a coordinate point r of the marsh, has an oscillating component, modulated at a frequency 2f double the frequency of the magnetic field. The amplitude F ₀ of the oscillating Lorentz force density can be explained according to the expression: 0 (r) = - afAB (r) (1), where σ denotes the electrical conductivity. It is possible to define a sliding speed V _G of the magnetic field V _G = fX (2) in which case the expression (1) can be expressed as follows: ₀ (r) = - aV _G B (r) (3) Thus, 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 application of a sliding magnetic field is translated, at a point of the marsh, by a modulation of its amplitude. Thus, the amplitude of the magnetic field at a point in the marsh varies with time, between a minimum amplitude B ™ ⁱⁿ and a maximum amplitude B ™ ^x . It is the same for the Lorentz force density, the latter having, at a point r of the marsh, a maximum value when the amplitude of the magnetic field, at this point, is maximum. By placing itself in the reference XYZ, linked to the mold 2, the propagation of a maximum value of the amplitude of the magnetic field B ™ ^ax , along an axis of propagation, causes, simultaneously, the propagation of a Lorentz force of maximum intensity F _max along the axis of propagation Δ. The combination of forces propagating along the axis of propagation establishes a movement of the liquid along this axis constituting an electromagnetic pump element.

The inventors have found that by modulating, in time, the maximum amplitude of the Lorentz force F _max propagating in the marsh, the intermittent macrosegregations are attenuated or even disappear, and this particularly on ingots whose thickness is greater than 300 mm.

This temporal modulation can be obtained by a variation of a parameter, called magnetic force parameter, controlling the amplitude of the Lorentz force density explained in equations (1) and (3), for example:

the value of the maximum amplitude β ™ ^χ of the magnetic field;

 the frequency / magnetic field;

 the wavelength λ of the sliding magnetic field.

When the sliding magnetic field is generated by a plurality of electromagnetic inductors arranged at the periphery of the mold, the temporal modulation of the Lorentz force density can be obtained by modifying the polar pitch, that is to say the phase shift between the induction currents flowing in each inductor. Such a modification makes it possible to vary the wavelength λ of the sliding magnetic field, that is to say the distance between two maxima propagating along the axis of propagation. The frequency of the induction current flowing in the inductors can be variable, which modifies the frequency / magnetic field. The amplitude of the induction current can also be variable, which modifies the value of the maximum amplitude B ™ ^x of the magnetic field. In FIG. 1D, an embodiment is shown in which the value of the maximum amplitude τ ^η αχ _{c u nam nam nam} et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et et Thus, a spatial distribution of the amplitude B ₀ (i) in the marsh, at a time t (continuous line), as well as a spatial distribution of the amplitude B ₀ (t + At), is represented at a moment t + At (dashed line). During the time interval At, the maximum amplitude B ^ ^ax varies between B ^ ^ax t) and B ^ ^ax (t + At). Similarly, the wavelength λ has been changed from (t) to (t + Δί). FIG. 1E, which represents a temporal evolution of an alternating magnetic field sliding at a point, shows an embodiment in which the value of the maximum amplitude B ™ ^ax of the magnetic field varies, over time, for a constant frequency / wavelength λ. As a result, in the examples shown in FIGS. 1D and 1E, the maximum amplitude of the Lorentz force, propagating in the marsh, varies between t and t + At, between the values F _max (t) and F _max (t + At).

 The temporal modulation of a force parameter is implemented during casting, during a significant period of time, preferably greater than 50% or even 80% of the duration of the casting. This temporal modulation may for example be applied for at least 30 minutes, or even at least 1 hour.

A sliding magnetic field B can in particular be generated from two inductors disposed on the same face of the ingot. The inductors are preferably arranged facing a large face of the ingot, that is to say one of the two faces of the ingot having the largest vertical section. The inductors can be superimposed on one another, so as to generate a so-called vertical phase shift, or arranged side by side, so as to generate a horizontal phase shift. In the examples described below, a device described in the application WO2014 / 155357, and more precisely according to the configuration described with reference to FIGS. 19 and 20A, in which three inductors oriented along the vertical axis Z, are arranged facing each large face of the ingot.

The sliding magnetic field can also be generated from one or more permanent magnets disposed at the periphery of the mold and set in motion relative thereto. For example, it is possible to generate a sliding magnetic field by rotating a permanent magnet. A variation of the sliding magnetic field parameters, be it its amplitude, its frequency or its wavelength makes it possible to apply a non-stationary Lorentz force in the marsh. The inventors have found that this helps to mitigate the appearance intermittent macrosegregations or even make them disappear. Such conditions probably affect spontaneous recirculations in the marsh and reduce their consequences.

Preferably, in the marsh, the rate of variation of the maximum Lorentz force density is greater than 0.05 Nm ^-3 s ^-1 , and preferably greater than 0.1 Nm ^-3 s ^-1 , and preferably greater than 0.2 Nm ^"3 .s ^_1. in one embodiment the maximum rate of change of the maximum density of Lorentz force during casting is at least 1 ^{Nm" 3} .s ¹ and preferably at less than 2 Nm ^"3 s ^"1 .

Preferably, the variation of one or more force parameters takes place in a time interval that is less than or equal to the characteristic durations of the recirculations generated by natural convection. These times vary according to the thickness of the ingot and the casting speed. Considering thicknesses e between 300 mm and 700 mm, and casting speeds of between 30 mm / min and 80 mm / min, the characteristic durations of the recirculations extend between 20 seconds (thickness of 300 mm, casting speed 30 mm / min) and 10 minutes (thickness of 700 mm, casting speed of 80 mm / min). Thus, the force parameters vary in a time interval At determined according to these characteristic times. By variation is meant a significant variation, of at least 10% of the force parameter considered, and preferably of at least 20% or even 30% of the force parameter.

The variation of a force parameter may be periodic, the time period of variation may be of the order of a characteristic recirculation time, that is to say be between 20 seconds and 10 minutes depending on the conditions of the dimensions and speed of the casting. Preferably, in the marsh, during the time period of variation, the maximum density of Lorentz force varies by at least 30 Nm ^"3 , and advantageously by at least 40 Nm ^{" 3} , and preferably by at least 50 Nm ^"3 , and even more preferably at least 60 Nm ^{" 3} . The variation of a force parameter can also be monotonous during the casting, for example according to an increasing or decreasing function between the beginning and the end of the casting, the value of the force parameter varying continuously or in successive increments .

Advantageously, during the casting, the Lorentz force of maximum intensity is not equal to zero. Typically, it is zero when the current in the inductors or coils is zero. Therefore, advantageously, the variation of the force parameter is not obtained by a periodic interruption of the sliding field. Advantageously, during the casting, the maximum intensity Lorentz force is greater than 80 N / m ³ , preferably greater than 100 N / m ³ , preferably greater than 120 N / m ³ , even more preferably greater than 140 N / m ³ . The inventors have indeed found that the suppression of intermittent macrosegregations was not optimum when the force was too weak as shown in Example 5 (Fig 20 a to d). The minimum value from which the suppression of intermittent macrosegregations is improved depends on the set of casting parameters, in particular the mixing mode, the position of the inductors with respect to the plate and the composition of the alloy.

According to one embodiment, the frequency / and / or the maximum amplitude B ™ ^x of the magnetic field are modified respectively by varying the frequency and amplitude of the induction current flowing in inductors. For this, the method may comprise a preliminary step of defining an operating range, that is to say a range of variation of the frequency and / or the intensity of the induction current. This preliminary step comprises the determination of one or more values of frequency / intensity pairs, called critical values, generating, on the free surface of the _sup of the marsh, a resonance, the resonance resulting in the appearance of significant oscillations of said free surface l _sup , the latter being shown in Figure 1A. These significant oscillations are generally observed with the naked eye. By significant oscillation is meant for example an oscillation whose amplitude is greater than or equal to 5 mm along the vertical axis Z. For example, the frequency of the current is fixed and the intensity of the induction current is increased up to a significant oscillation is observed.

Considering different critical values of frequency (or intensity), it is possible to determine experimentally a resonance curve R, in a frequency / intensity plane corresponding to the different pairs (frequency / intensity) at which a resonance is observed at the free surface marsh. From this curve R, a range of variation of the intensity and / or the frequency is determined so as to avoid or limit the appearance of a resonance of the free surface of the marsh. Indeed, the resonance curve defines a zone of stability and a zone of instability, in which the casting can become dangerous. However, the fact of modulating the frequency or the intensity of the induction current, and therefore the frequency / or the maximum amplitude B ™ ^x of the sliding magnetic field, makes it possible to approach temporarily the resonance curve R, by example periodically, while remaining in the stability zone. This maximizes the intensity of the Lorentz force, and therefore the brewing the marsh, while remaining in acceptable safety configurations. Indeed, in the vicinity of the resonance curve, the brewing effect is particularly important.

Such a resonance curve depends on the casting conditions, that is to say the dimensions of the mold, the casting speed, the configuration of the applied magnetic field, the latter depending on the magnetic field generator, it is that is to say inductors or permanent magnet (s) used. A resonance curve R is shown in FIG. 2, this curve having been obtained by casting a ingot of thickness 525 mm × 1650 mm, with a casting speed of 45 mm / min, a magnetic stirring being carried out by the application a magnetic field by three inductors arranged in front of each large face of the ingot and phase shifted by 90 ° to form a horizontal electromagnetic pump element, as previously mentioned. This figure also shows abacuses representing a percentage of the intensity of a Lorentz force, called nominal, 100% corresponding to the intensity of the maximum induction current usable in the installation when the frequency is equal at 0.2 Hz. This intensity corresponds to the appearance of a resonance at the frequency of 0.2 Hz. Preferably, the intensity and the frequency of the induction current are in a space delimited by the curve representing a certain percentage of the intensity of the nominal Lorentz force, for example 10% of this intensity, and the resonance curve.

Preferably, the method comprises a variation of the frequency of the induction current flowing through an inductor. The inventors have found that it is advantageous to vary the frequency because the resulting change in field penetration allows the force gradient to be varied more effectively in the thickness and depth of the liquid well. On the other hand, the power electronics cause the frequency variation to be faster than the intensity variation; which gives an additional degree of freedom to the weaker periods of unsteady forcing. It is indeed advantageous to decouple the hydrodynamic characteristic times characteristic times of solidification to avoid intermittent macrosegregations.

Another example of a curve is shown in FIG. 8 and will be discussed later in connection with the examples. FIGS. 2 and 8 show the experimentally determined resonance curve R as well as the curve representing a Lorentz force whose intensity is equal to 10% of the previously defined nominal Lorentz force. The variation of one or more force parameters can notably make it possible to alternate periods during which the adimensional number of Hartmann Ha is respectively small, typically less than 1, and high, typically greater than 3 or even 5. The non-dimensional number of Hartmann Ha is a number commonly used in the field of magnetohydrodynamics. It represents a ratio between the magnetic viscosity and the viscosity of a charged liquid flowing in a magnetic field. The greater the number, the greater the contribution of the Lorentz forces. Preferably the adimensional number of Hartmann Ha alternates with a ratio between low and high values of at least 3 or at least 5. Such a configuration is preferred because it allows to alternate periods during which the kinetic energy applied by the magnetic field opposes the natural convection of the liquid metal, and periods during which natural convection predominates.

As described in connection with the examples presented below, the products obtained by a method according to the invention have an intermittent macrosegregation limited compared to methods of the prior art, or even not perceptible. In the following examples, product characterization was carried out by analyzing horizontal profiles (along the TC axis) of a radiography made at mid-width along a vertical L / TC plane, these profiles being calibrated to obtain the distribution. space of Zn or Cu type heavy alloy elements. Areas enriched with such heavy elements, more absorbent, appear as dark spots on the negative radiographs performed and therefore clear spots on radiographs presented. An example of obtaining the Zn concentration profile from a radiograph of an Al-Zn alloy is shown in FIG. 4.

The terms L, TL and TC, known to those skilled in the art, respectively correspond to the dimension of the ingot along the vertical axis, the so-called "long axis" along the so-called "short cross" axis. Complementarily or alternatively, it is possible to perform chemical analyzes in horizontal profiles, so as to quantify the spatial distribution of said chemical elements along the axis TC. An intermittent macrosegregation can be characterized by a maximum mass deviation of an alloying element, in this case Zn, in the zone most marked by intermittent macrosegregations, that is to say in the vicinity of T / 2.5. . To quantify the intermittent macrosegregation, the concentration profiles, obtained by radiography or by any other method, with a spatial resolution of 0.1 mm were treated as illustrated in FIG. 5A. The profile obtained with the resolution of 0.1 mm is the raw profile referenced profile A. A sliding average over 2 mm makes it possible to overcome the microsegregation, the smoothed profile obtained is referenced profile B. Another sliding average of the gross profile over 50 mm makes it possible to get rid of intermittent macrosegregations, and obtain the continuous macro-segregation profile, the profile obtained being a so-called basic profile, referenced profile C. The profile C is subtracted from the profile B to obtain a profile called corrected, corresponding to the intermittent macrosegregation, the corrected profile being referenced profile D. Such a profile is shown in Figure 5B. As can be seen in this Figure 5B, the corrected profile is primarily representative of intermittent macrosegregation, and is unaffected or unaffected by central continuous macrosegregation and microsegregation. Such a corrected profile makes it possible to characterize the intermittent macrosegregation.

It is then possible to calculate a maximum concentration difference in an analysis zone Z _A located between T / 2.3 and T / 3.3, this maximum difference being able to be expressed according to the following equation:

ACZA = max (C _ZA ) - min (C _ZA ) (4) where max (CZA) and min ( _CZ A) respectively denote the maximum and minimum concentrations of the considered element measured between T / 2.3 and T / 3.3.

The element considered is an element whose content in weight in the alloy is greater than or equal to 0.5%. It may be, preferably, the major element of the alloy, the term major element corresponding to the definition given by The Aluminum Association. The maximum deviation ΔC _Z A can be normalized by the nominal concentration Co of the considered element. The products according to the invention preferably have a value of such a standardized ratio of less than 10% and preferably less than 8% or even less than 6%. However, the solute value of ΔC _Z A can be influenced by the thickness of the product, the nature of the element considered, in particular its partition coefficient and / or its concentration. It is therefore useful for characterizing the products obtained by the process according to the invention to calculate, by way of reference, a maximum difference in a reference zone Z _{R that is} not very sensitive to intermittent macrosegregations, situated between T / 6 and T / 12. this maximum difference can be expressed according to the following equation:

AC _Z R = max (C _ZR ) - min (C _ZR ) (5) where max (C _Z R) and min (C _ZR ) respectively denote the maximum and minimum concentrations of the element under consideration measured between T / 6 and T / 12. A dispersion criterion ε is thus obtained which makes it possible to evaluate intermittent macrosegregation for the element in question:

In order to overcome local variations in composition, it is advantageous, in order to determine AC _Z A and ÙCZR, to calculate an average over at least 5 concentration profiles distant from at least

10 mm.

The smaller ε, the less intermittent macrosegregations are marked. The products obtained by the process according to the invention preferably have a dispersion criterion ε of less than 3.3, preferably less than 3, more preferably less than 2.5, still more preferably less than 2 and preferably less than 1.5.

According to a nomenclature known to those skilled in the art, T / n denotes a distance from an edge of the ingot, along a horizontal axis, T / 2 corresponding to the center of the ingot.

It is also useful to perform a Fourier transform analysis of the raw composition profile and to normalize it by the nominal composition of the element. Such an analysis makes it possible to identify spatial periods characterizing the intermittent macrosegregation. Intermittent macrosegregation has periods of between 8 and 25 mm. When intermittent macrosegregation is important, we observe a peak of the amplitude of the Fourier components for spatial periods between 8 and 25 mm. An adimensional spectral intensity criterion ζ is determined which corresponds to the maximum amplitude of the Fourier components in a range of spatial period between 8 and 25 mm, normalized by the nominal concentration Co of the considered element. 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 dispersion criteria ε and spectral intensity ζ are advantageously applied to the major element of the alloy considered, typically Zn for a 7xxx alloy or Cu for a 2xxx alloy. These criteria can also be applied to the sum of two elements, for example the sum Zn + Cu in some alloys 7xxx or the sum Mg + Si in alloys 6xxx. These criteria may also apply to an element whose content in weight in the alloy is greater than or equal to 0. 5% or the sum of two elements of the alloy whose individual content is greater than 0.5%, In the case where we consider the sum of two elements, the values to normalize the maximum deviation ACZA, and / or the Fourier transform correspond to the sum of the nominal concentrations of the considered elements.

The ingots of rectangular section obtained by the process according to the invention can be used as cast or after wrought, optionally after dissolution and quenching and aging for alloys with structural hardening. Advantageously, ingots of rectangular section obtained by the process according to the invention are rolled and / or forged.

Example 1

 A casting of an AA7035 alloy was performed without electromagnetic stirring. The composition of the cast alloy comprising a nominal Zn concentration of 5.6% by weight, a nominal Mg concentration of 1.3% by weight. The format of the ingot was 1650 mm x 525 mm. This example is representative of the prior art. Grain refining was performed with a 5: 1 AITiB refining concentration of 1 kg / t. The casting speed was 35 mm / min. Figure 3 shows a radiograph of the mid-width ingot in an L / TC plane, on which the negative central macrosegregation and the intermittent macrosegregations are clearly identifiable. FIG. 6A shows different horizontal gross profiles of the Zn content, along an axis TC, and smoothed profiles B are obtained by a sliding average over 2 mm deduced from FIG. 3. The radiography makes it possible to quantify that the elements at the origin of a contrast with respect to aluminum, namely in this case Zn. This remark applies to the following example 2. Negative central macrosegregation is clearly observed, maximum at T / 2, with intermittent macrosegregations observed between T / 2.3 and T / 3.3. FIG. 6B shows the various corrected profiles of the Zn content (profiles D), along an axis TC, obtained after subtraction of each smoothed profile (profile B) by a basic profile (profile C) representative of the macrosegregation continues.

The value of the maximum deviations of the Zn content was 0.75% by weight for & C _Z A and 0.19% by weight for AC _Z ¾ the value of the maximum normalized deviations in the analysis zone and in the reference zone thus being respectively 13.3% and 3.5%. The value of the dispersion criterion ε as defined by equation (6) was 3.9. The Fourier transform of each profile was calculated, and is shown in Figure 7, after normalization by the nominal composition of Zn: 5.6% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. There are different predominant peaks, corresponding to different spatial periods distributed between 8 and 25 mm, and more particularly between 10 mm and 25 mm. The spectral intensity criterion ζ, which corresponds to the maximum amplitude of the Fourier components between 8 and 25 mm, normalized by the nominal Co concentration of Zn, was for the set of profiles at least 0.01. Example 2

In a second example, a casting of an AA7035 alloy with an electromagnetic stirring was performed. The composition of the cast alloy comprised a nominal cen-zinc concentration of 5.6% by weight and a nominal magnesium concentration of 1.3% by weight. The ingot size was 1650 mm x 525 mm. Grain refining was performed with a 5: 1 AITiB refining concentration of 1 kg / t. The casting speed was 35 mm / min. The electromagnetic stirring was obtained by placing, opposite each L / TL face of the ingot, (corresponding to a YZ plane in the reference indicated in FIG. 1A), three inductors oriented along the vertical axis Z, alternating current, of frequency 0.25 Hz, phase-shifted relative to each other by 60 ° and spaced from each other by 0.6 m, thus constituting an electromagnetic pump element. The distance between the inductors and the ingot was 172 mm. The electromagnetic pump elements on each side were facing in the opposite direction. The inductors generated a sliding magnetic field along a horizontal plane, the sliding axis being parallel to the TL direction, the wavelength λ was 3.6 m. The maximum density of the Lorentz force induced in the liquid marsh has been varied between about 180 N / m ³ and 240 N / m ³ with a speed of variation of 2 Nm ³ ss ¹ by modifying the nominal value of the current. The resonance curve, corresponding to these casting conditions, is shown in FIG. 8. The variation of the intensity of the induction current is represented in this figure by a double arrow.

Figure 9 shows a radiograph of the ingot in an L / TC plane, on which the negative central macrosegregation at T / 2 is identifiable. In FIG. 10A, different gross horizontal profiles of the content of Zn (profile A) and smooth (profiles B) are represented along a TC axis. We distinguish the central negative macrosegregation, maximum at T / 2. FIG. 10B shows the different horizontal profiles of the Zn content, along a TC axis, of corrected profile type (D profiles) obtained after subtraction of the profile corresponding to the continuous macrosegregation.

The maximum deviations in the Zn content were 0.24% by weight for & C _Z A and 0.28% by weight for AC _Z ¾ the value of the maximum standard deviations in the analysis zone and in the zone of reference being respectively 4.3% and 5%. The value of the dispersion criterion ε as defined by equation (6) was 0.9: the intermittent macrosegregation in the analysis zone between T / 2.3 and T / 3.3 was eliminated. The Fourier transform of each profile was calculated, and is shown in Fig. 11, after normalization by the nominal composition of Zn: 5.6% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. No more dominant peaks are observed. The spectral intensity criterion ζ, which corresponds to the maximum amplitude of the Fourier components between 8 and 25 mm normalized by the nominal concentration Co of Zn, was for all the profiles less than 0.005.

Example 3 In this example, a casting of AA 7050 alloy was performed without electromagnetic mixing. The composition of the alloy was 6.3% by weight Zn, 2.2% by weight Mg and 2.1% by weight Cu. The format of the ingot was 1650x525mm. The grain is refined using a refining yarn AITIC3: 0.15 with an addition rate of 1 kg / ton. The casting speed was 45mm / min. It constitutes the reference of example 4. FIG. 12 represents a radiography of the ingot according to an L / TC plane, on which the negative central macrosegregation at T / 2 is identifiable. FIG. 13a shows the smoothed horizontal profile of the sum of two elements Zn and Cu (profiles B) along an axis TC, deduced from the radiography of FIG. 12. In fact, the radiography only makes it possible to quantify the elements at the origin of a contrast with respect to aluminum, namely in this case Zn and Cu. This remark applies to the following Examples 4 and 5. FIG. 13b shows the different horizontal profiles of the concentration of Zn + Cu, along a TC axis, of corrected profile type (D-profiles) obtained after subtraction of the profile corresponding to the continuous macrosegregation. The value of the maximum deviations of the sum Zn + Cu was 0.81% by weight for AC _Z A and 0.19% for AC _Z R. The value of the dispersion criterion ε as defined by equation (6) was 4.4. FIG. 14 represents the Fourier transform of each profile, after normalization by the sum of the nominal compositions in Zn and Cu: 8.3% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. The spectral intensity criterion ζ, which corresponds to the maximum amplitude of the Fourier components between 8 and 25 mm normalized by the sum of the nominal compositions in Zn and Cu, was for one of the profiles greater than 0.01 or for all the profiles greater than 0.007.

Example 4 In this example, an AA 7050 alloy casting was performed. The composition of the alloy was 6.3% by weight Zn, 2.2% by weight Mg and 2.1% by weight Cu. The ingot section was 1650x525mm. The grain is refined using a refining yarn AITIC3: 0.15 with an addition rate of 1 kg / ton. The casting speed was 45mm / min. The electromagnetic stirring was obtained by placing, opposite each L / TL face of the ingot, (corresponding to a YZ plane in the reference indicated in FIG. 1A) of three coils oriented along the z axis and traversed by an alternating current which was out of phase, in the central coil, by 90 ° with respect to the current in the extreme coils. The wavelength of the sliding field was 2.4 m. The electromagnetic pump elements thus obtained were arranged in mirror with respect to each L / TL face of the ingot, the direction of sliding being parallel to the cross-long direction, the generated slips diverging from the mid-width of the ingot. Unsteady forcing was obtained by imposing a cyclic variation in the frequency of the alternating electric current flowing through the coils, as illustrated by the double arrow in the frequency vs intensity diagram in FIG. the Lorentz force thus generated by the variation of the frequency between 0.450 and 0.600 Hz has been varied between approximately 110 N / m ³ and 150 N / m ³ over a period of 3 min, which corresponds to a speed of variation of approximately 0.22 N / m ³ / s.

Figure 16 shows a radiograph of the ingot in L / TC plane, on which the negative central macrosegregation at T / 2 is identifiable. The intermittent macrosegregations are greatly attenuated with respect to the reference (FIG. 12), as shown in FIGS. 17a and 17b.

FIG. 17a shows the smoothed horizontal profile of the sum of the elements Zn + Cu (profiles B) along an axis TC, deduced from the radiography of FIG. 16. In FIG. 17b, the various profiles are represented. horizontal axes of the sum of the two elements Zn and Cu, along an axis TC, of corrected profile type (Profiles D) obtained after subtraction of the profile corresponding to the continuous macrosegregation. The value of the maximum deviations of the Zn + Cu content was 0.30% by weight for AC _Z A and 0.14% for AC _Z R. The value of the dispersion criterion ε as defined by equation (6) was 2.2. The intermittent macrosegregation in the analysis zone has therefore been reduced and is represented in FIG. 18, after normalization by the sum of the nominal Zn and Cu compositions: 8.3% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. The spectral intensity criterion ζ, which corresponds to the amplitude The maximum of Fourier components between 8 and 25 mm normalized by the sum of the nominal compositions in Zn and Cu, was for all the profiles less than 0.005.

Example 5

In this example, an AA7050 alloy casting was performed. The composition of the alloy was 6.3% by weight of Zn, 2.2% by weight of Mg and 2.1% by weight of Cu, the contents of the other elements were all less than 0.5% by weight. The ingot section was 1650x525mm. The grain is refined using a refining yarn AITIC3: 0.15 with an addition rate of 1 kg / ton. The casting speed was 45mm / min. The electromagnetic stirring was obtained by placing, opposite each L / TL face of the ingot, (corresponding to a YZ plane in the reference indicated in FIG. 1A) of three coils oriented along the z axis and traversed by an alternating current which was out of phase, in the central coil, by 90 ° with respect to the current in the extreme coils. The wavelength of the sliding field was 2.4 m. The electromagnetic pump elements thus obtained were arranged in mirror with respect to each L / TL face of the ingot, the direction of sliding being parallel to the cross-long direction, the generated slips diverging from the mid-width of the ingot.

The unsteady forcing was obtained by imposing a variation from zero of the intensity of the alternating electric current flowing through the coils, as illustrated by the arrows in the frequency vs. intensity diagram in Figure 19. L The intensity of the Lorentz maximum volumetric force thus generated by the variation in intensity varied typically from 0 N / m ³ to 140 N / m ³ in 4 min, which corresponds to a variation speed of 0.58 N / m 3 / s. Subsequently, the intensity of the Lorentz maximum volume force was varied between 140 N / m ³ and 360 N / m ³ in 5 min, which corresponds to a variation speed of 0.73 N / m 3 / s.

The results obtained are illustrated by the two vertical X-rayed slices shown in FIG. 20a (variation of the intensity between 0 N / m ³ at 140 N / m ³ in 4 min) and FIG. 21 a (variation of the force between 140 N / m ³ to 360 N / m ³ in 5min) which are in continuity of one another.

In FIG. 20b, the smoothed horizontal profile of the sum of the major elements Zn + Cu (profiles B) is represented along an axis TC, deduced from the radiography of FIG. 20a. FIG. 20c shows the different horizontal profiles of the sum of the elements Zn + Cu, along an axis TC, of corrected profile type (D-profiles) obtained after subtraction of the profile corresponding to the continuous macrosegregation. The value of the maximum deviations of the Zn + Cu content was 0.70% by weight for AC _Z A and 0.22% for AC _Z R. The value of the dispersion criterion ε as defined by Equation (6) was 3.2. FIG. 20d represents the Fourier transform of each profile, after normalization by the sum of the nominal compositions in Zn and Cu: 8.3% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. The spectral intensity criterion ζ, which corresponds to the maximum amplitude of the Fourier components between 8 and 25 mm normalized by the sum of the nominal Zn and Cu compositions, was for all the profiles less than 0.01. Note, however, that the spectral intensity criterion ζ shows values greater than 0.005.

In FIG. 21b, the smoothed horizontal profile of the sum of the major elements Zn + Cu (profiles B) is represented along an axis TC, deduced from the radiography of FIG. 21a. In FIG. 21c, the different horizontal profiles of the sum of the major elements Zn + Cu, along a TC axis, of the corrected profile type (D-profiles) obtained after subtraction of the profile corresponding to the continuous macrosegregation are represented. The value of the maximum deviations of the Zn + Cu content was 0.37% by weight for AC _Z A and 0.15% for AC _Z R. The value of the dispersion criterion ε as defined by equation (6) was 2.4. FIG. 21d represents the Fourier transform of each profile, after normalization by the sum of the nominal compositions in Zn and Cu: 8.3% by weight. The abscissa axis represents the spatial period, between 0 and 30 mm. The spectral intensity criterion ζ, which corresponds to the maximum amplitude of the Fourier components between 8 and 25 mm normalized by the sum of the nominal Zn and Cu compositions, was for all the profiles less than 0.005. It is thus observed that the suppression of intermittent macrosegregations is improved if the force is greater than 140 N / m ³ . Indeed, when the force is too weak, it is found that the value of the dispersion criterion ε 00 0000000000 of spectral intensity ζ are greater than the preferred values of the invention. The inventors thus assume that an unsteady forcing which consists in periodically interrupting the sliding field would not make it possible to advantageously suppress the intermittent macrosegregations.

Claims

1. Process for casting an aluminum alloy ingot in a substantially rectangular ingot mold comprising the following steps: - preparation of the aluminum alloy; casting of the aluminum alloy in the ingot mold, along a vertical casting axis, the alloy being cooled, during casting, by a trickle of a cooling liquid in contact with the solidified metal; during casting, application of a magnetic field whose amplitude (B ₀ ) is varied periodically according to a frequency ( ^" ), said magnetic field being generated by at least one magnetic field generator arranged at the periphery of the ingot mold , so as to apply a Lorentz force (F) at different points of a liquid part of the alloy being solidified; the applied magnetic field being a sliding magnetic field, propagating along an axis of propagation, such that a maximum amplitude (B™ ^x ) of the magnetic field propagates along said axis of propagation, defining a wavelength of propagation (λ), said sliding magnetic field causing a propagation, along said propagation axis, of a Lorentz force of maximum intensity [F _max ); the method being characterized in that a magnetic parameter called force, governing Lorentz force of maximum intensity (F _max ), is variable in a predetermined time interval (At), said parameter being: said maximum amplitude (B™ ^x ) of the magnetic field; and/or said frequency ( ) of the magnetic field; ■ and/or the propagation wavelength (λ) of the magnetic field; so as to obtain a modulation, in said time interval, of said Lorentz force of maximum intensity (F _max ) propagating along the propagation axis. 2. Method according to claim 1, in which the section of the ingot mold, in a horizontal plane, defines a thickness (e) and a length (β), the thickness being less than or equal to the length, the thickness being greater at 300 mm and preferably at least 400 mm. 3. Method according to any one of the preceding claims, in which the frequency of the magnetic field is less than 5 Hz, or 2 Hz or 1 Hz. 4. Method according to any one of the preceding claims in which, the Lorentz force of maximum intensity (F _max ), propagating along the axis of propagation, varies by at least 30 Nm ^"3 in a time interval ( At) between 20 seconds and 10 minutes. 5. Method according to any one of the preceding claims in which, the magnetic field is such that the absolute value of the variation in the density of the maximum Lorentz force is greater than or equal to 0.05 Nm ^"3. s ¹ during said interval temporal (At). 6. Method according to any one of the preceding claims, in which the axis of propagation of the maximum amplitude of the magnetic field belongs to a plane parallel to the casting direction. 7. Method according to any one of the preceding claims, in which during casting, the variation of the force parameter is periodic, the period being between 20s and 20 minutes, or between 1 minute and 15 minutes, or between 2 minutes and 10 minutes. 8. Method according to any one of the preceding claims, in which the generators are electromagnetic inductors, each electromagnetic inductor being traversed by a current called induction current, the method comprising, during said time interval: a variation of an intensity induction current; and/or a variation of a frequency of the induction current; and/or a variation of a distance between an electromagnetic inductor and the ingot mold. 9. Method according to claim 8, comprising a variation of the intensity or frequency of the induction current flowing through an inductor, the method comprising: a preliminary step of defining at least one critical value of the intensity and frequency of the induction current generating, at the level of a free surface (l _sup ) of the aluminum alloy flowing in the ingot mold, a resonance wave; - a determination of a range of variation of the intensity or frequency of the induction current as a function of said previously defined critical value. 10. Method according to claim 9 comprising, during said preliminary step, a definition of a plurality of critical values of the intensity and frequency of the induction current, so as to define a resonance curve (), representing the intensity and frequency values generating a resonance of said free surface, the method comprising a determination of a range of variation of the intensity or frequency of the induction current in a domain delimited by said resonance curve. 11. Method according to any one of claims 1 to 7, in which at least one generator is a permanent magnet, the method comprising: - a variation in a distance between the permanent magnet and the ingot mold; and/or a rotation of the permanent magnet, and a variation in the speed of rotation of the magnet; and/or a rotation of two permanent magnets. 12. Method according to any one of the preceding claims, in which the aluminum alloy is chosen from alloys of types 2XXX, 5XXX, 6XXX or 7XXX and in which the thickness is at least 400 mm or 450 mm. 13. Method according to any one of the preceding claims, in which the dimensionless Hartmann number, at at least one point of the liquid part of the alloy, varies at least by a factor of 3, or even by a factor of 5, in said time interval (At). 14. Aluminum alloy ingot, obtained by the process which is the subject of any one of claims 1 to 13. 15. Aluminum alloy ingot, according to claim 14 having, for one element of the alloy, whose content by weight is greater than 0.5%, or for the sum of two elements of the alloy whose individual content in weight is greater than 0.5%, a dispersion criterion less than 3.3, preferably less than 3, more advantageously less than 2.5, even more advantageously less than 2 and preferably less than 1.5, the dispersion criterion being defined according to the following expressions: e = C _ZA / AC _Z « (6) ACZA = max (C _ZA ) - min (C _ZA ) (4), ACZR = max (C _ZR ) - min (C _ZR ) (5), Or : max (C _ZA ) and min (C _ZA ) respectively designate the maximum and minimum concentrations of the element considered or the sum of the two elements considered measured in an analysis zone (ZA), presenting intermittent macrosegregations, for example between T/2.3 and T/3.3; max (C _ZR ) and min (C _ZR ) designate respectively the maximum and minimum concentrations of the element considered or the sum of the two elements considered measured in a reference zone (Z), considered to be little affected by intermittent macrosegregation, for example between T/6 and T/12; said concentrations being measured on at least one profile established at mid-width in a vertical plane L/TC and in the direction TC, said profile being representative of said intermittent macrosegregations of the element considered in the direction TC. 16. Aluminum alloy ingot, according to claim 14 or claim 15, a spectral intensity criterion (ζ) of which is less than 0.01, preferably less than 0.007 and preferably less than 0.005, said intensity criterion spectral being calculated in: Determining a maximum amplitude of a Fourier transform of a profile representative of an intermittent macrosegregation of an element whose content by weight is greater than 0.5% or the sum of two elements of the alloy whose individual content is greater at 0.5%, the profile being established according to said direction TC, said maximum amplitude being determined in a range of spatial periods between 8 and 25 mm, normalizing said maximum amplitude by a nominal concentration Co of said element or by the sum of the nominal concentrations of the two elements considered.