Characterisation of bed of metal powder by colorimetry

Abstract

To provide characterisation of a bed of metal powders by colorimetry.SOLUTION: The present invention relates to a method (20) for determining the oxygen concentration (Cox) of a powder (MP) of a metallic material (Mat) taking the form of a powder bed (PB), the method comprising the following steps: A) generating an image (Im) of at least a portion of the powder bed, wherein the image comprises a set of pixels, the pixels having three quantities (G1, G2, G3 ) of colors coded by a colorimetric code including three quantities (G1, G2, G3); and B) determining oxygen concentration of powders according to value of the three quantities associated with the image pixels by using a calibration function (CFMat) defined in advance that is a function of the materials and associates the oxygen concentration with the three quantities.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present invention relates to the field of characterizing metal powders, and more particularly to the field of determining the oxygen concentration of such powders, the latter having a flat surface (powder bed). This characterization is typically useful when carrying out metal addition manufacturing methods for parts. [Background technology]

[0002] Metal powders are prone to oxidation during use, which can hinder their application. This problem arises, for example, when implementing powder bed fusion bonding metal addition (MAM-PBF) methods with selective solidification of the powder bed. This method is typically used in aeronautics, aerospace, biomedical, automotive, and nuclear applications.

[0003] Additive manufacturing involves producing parts by sequentially adding layers of (metallic) material one by one based on a 3D digital file. A powder bed (PB) should be understood as a powder mass (MP) of controlled thickness with a flat surface. The metal material (Mat) constituting the powder is selected from, for example, stainless steel, titanium-based alloys, aluminum-based alloys, and nickel-based alloys.

[0004] Manufacturing requires spreading thin layers of powder (typically between 10 and 100 μm thick) onto one another, with a selective solidification step between each layer deposition. Selective solidification is performed, for example, using one or more laser beams, electron beams, laser sintering, or by buffing. Solidification should be understood as hardening the material by bonding the powder particles together.

[0005] An example of a system 15 using MAM-PBF (indicated as L-PBF) manufacturing of part Pa using a laser beam LB is shown in Figure 1. This system includes a tank PC for metal powder MP, a manufacturing platform MPL, and a powder collection tank PCT. A powder coating device PSD pushes powder from the tank to the platform MPL in a coating direction SD such that a controlled thickness of powder (typically about 10–100 μm), known as the powder bed PB, is deposited. This direction SD is along axis X, the plane of the powder bed is plane XY, and Z is perpendicular. During coating, excess powder is discharged into the collection tank PCT. The coating device PSD is, for example, a scraper, roller, or brush. A vertical transfer device MD is located inside the tank Tk used for manufacturing to move the manufacturing substrate Sub, thereby descending along Z with the substrate and the part Pa to be formed as the layers are deposited and solidified. At the start of manufacturing, the substrate Sub receives the powder bed, and then during manufacturing, the part is attached to this substrate. Here, the solidification device CD includes a set of laser L and mirrors LSD controlled by a processing unit UT0 that deflects the laser beam LB based on data from a 3D file.

[0006] Therefore, the MAM-PBF method includes a cycle for each layer that includes a step of applying powder (and returning the apparatus PSD to its initial position), including the discharge of excess powder, and a solidification step. Next, the substrate is lowered, and the formation of the next layer and so on by a new cycle is carried out until the final layer.

[0007] After manufacturing is complete, the substrate to which the manufactured parts are attached is removed from the machine. The powder filling tank Tk and the powder from tank PCT are recovered, optionally sieved, characterized, and, if appropriate, mixed with new powder for reuse in subsequent manufacturing. In practice, only a very small portion of the powder solidifies into parts during manufacturing, and the unmelted powder can be recovered and reused in subsequent manufacturing. However, due to the interaction between the laser and the material, the high temperatures generated, and the imperfectly controlled atmosphere in the manufacturing chamber, some particles degrade. Thus, powder reuse results in degradation of properties in some particles that are irregularly distributed in the powder. The deterioration of powder quality induces a deterioration in the properties of the manufactured parts. An increase in the oxygen concentration of recycled powder has been documented in the literature on many materials. Some powder particles are surrounded by a somewhat uniform layer of oxides, and their chemical properties vary depending on the material Mat. For example, this phenomenon in the case of stainless steel powder is described in the publication T. Delacroix et al., “Influence of powder recycling on 316L stainless steel feedstocks and printed parts in laser powder bed fusion” (Addit. Mauf., vol.25, pp.84-103, 2019). Therefore, oxidation is one of the good indicators of deterioration in powder quality, especially when typically used in MAM-PBF manufacturing.

[0008] Currently, powder samples are collected between manufacturing cycles, and small quantities are characterized ex situ to assess their quality. These small characterized samples do not necessarily represent the entire powder.

[0009] In some cases, to avoid taking risks, unmelted powder recovered after manufacturing is not inspected. Instead, only new powder is used, and it is directly discarded to guarantee the quality of the manufactured parts, which increases the cost of the parts. [Prior art documents]

Non-Patent Literature

[0010]

Non-Patent Literature 1

Summary of the Invention

Problems to be Solved by the Invention

[0011] One object of the present invention is to improve the above disadvantages by proposing a method for determining the oxygen concentration of a powder of a metallic material in the form of a powder bed, and a related apparatus that is quick, relatively low-cost, and can be carried out directly online during the implementation of a powder bed metal additive manufacturing method.

Means for Solving the Problems

[0012] The present invention is, for the subject matter, a method for determining the oxygen concentration of a powder of a metallic material in the form of a powder bed, comprising: A) generating an image of at least a part of the powder bed, the image including a set of pixels, the pixels having a color encoded by a colorimetric code including three quantities; B) using a predefined calibration function that is a function of the material and that associates the oxygen concentration with the three quantities to determine the oxygen concentration of the powder from the values of the three quantities associated with the pixels of the image; and having steps as described above.

[0013] According to one embodiment, the colorimetric code is in the RGB system and the three quantities are known as R, G, and B.

[0014] According to one embodiment, the calibration function is a linear or quadratic polynomial, has three variables corresponding to three quantities, and their coefficients are a function of the material.

[0015] According to one variant, step A is performed at once using a video camera.

[0016] According to another variant, step A is performed by scanning using a flatbed scanner.

[0017] According to the first method, step B is: B1) determining, for a plurality of pixels of the image, the oxygen concentration known as pixel density, using the calibration function; and B2) determining the oxygen concentration from the average value of the pixel density. It includes the sub-steps as described above.

[0018] According to the second method, step B is: [[ID=2३]] B’1) determining the average value of each of the quantities from the values of the quantities associated with a plurality of pixels of the image; and B’2) determining the oxygen concentration from the average values of the three quantities through the calibration function. It includes the sub-steps as described above.

[0019] In another aspect, the present invention relates to a metal additive manufacturing method by selective solidification of a powder bed, including the step according to the present invention of determining the oxygen concentration of the powder bed, which is performed at least once in the metal additive manufacturing method.

[0020] According to one embodiment of the metal additive manufacturing method according to the present invention, the method for determining the oxygen concentration of the powder bed is performed at the start of the metal additive manufacturing method, during the injection of an inert gas into the manufacturing chamber, and / or during the cooling of the manufactured part at the end of the metal additive manufacturing method.

[0021] According to one embodiment, the metal addition manufacturing method according to the present invention includes the step of adapting the parameters of the method as a function of a determined oxygen concentration value.

[0022] According to one embodiment, the metal addition manufacturing method according to the present invention further includes a step of mixing used powder with new powder when the oxygen concentration exceeds a predetermined threshold, and this mixing step is performed between two manufacturing operations of the part.

[0023] In another aspect, the present invention relates to an apparatus for determining the oxygen concentration of a powder of a metallic material in the form of a powder bed: An image acquisition device configured to generate an image of at least a portion of the powder bed, wherein the image includes a set of pixels, and the pixels have a color encoded by a colorimetric code comprising three quantities, A first processing unit configured to determine the oxygen concentration of the powder from the values ​​of the three quantities related to the pixels of the image, using a predefined calibration function which is a function of the material and links the oxygen concentration to the three quantities, Regarding devices including...

[0024] In another embodiment, the present invention relates to a computer program that includes instructions causing an apparatus according to the present invention to carry out steps of a method for determining oxygen concentration according to the present invention.

[0025] In yet another embodiment, the present invention relates to a system for metal addition by selective solidification of a powder bed, and includes: Apparatus according to the present invention for determining oxygen concentration, A tank intended to contain the base material on which the parts are manufactured, An apparatus for applying the powder onto the surface of a powder bed, A device for solidifying powder, A second processing unit configured to control the execution of metal additive manufacturing, Regarding systems that include this.

[0026] According to one embodiment, the image acquisition device includes a flatbed scanner connected to a coating device.

[0027] According to one embodiment, the image acquisition device includes a high-resolution video camera.

[0028] In a further embodiment, the present invention relates to a method for determining a calibration function for carrying out a method for determining oxygen concentration according to the present invention, and includes: The objective is to obtain multiple samples of the aforementioned powder, wherein each sample has a different known oxygen concentration. The process involves generating an image of each sample and determining the associated average color, where the color is coded by a colorimetric code containing three quantities, and the color with the associated oxygen concentration value is known as calibration data. The calibration function is determined from the calibration data by regression, This concerns a decision-making method that includes the steps described below.

[0029] The following description outlines numerous representative embodiments of the apparatus of the present invention, and these examples are not limitations to the scope of the invention. These representative embodiments possess both essential features of the invention and further features relating to the relevant embodiments.

[0030] The present invention will be better understood and its other features, purposes, and advantages will become apparent through the following detailed description with reference to the attached drawings provided by non-limiting examples. [Brief explanation of the drawing]

[0031] [Figure 1] As already mentioned, this describes a system using metal addition manufacturing by selective solidification of a powder bed. [Figure 2] This invention provides a method for determining the oxygen concentration of a powder in the form of a powder bed. [Figure 3] Two applicable methods for carrying out step B of the method according to the present invention are shown. [Figure 4] A first method for carrying out step B according to the present invention is shown. [Figure 5] A second method for carrying out step B according to the present invention is shown. [Figure 6] The image acquisition device shows a first modified form of the apparatus according to the present invention, which includes an optical system and a matrix detector. [Figure 7] This invention illustrates a system for metal addition manufacturing by selective solidification of a powder bed. [Figure 8] The present invention provides a more detailed description of a system for metal addition manufacturing by selective solidification of a powder bed. [Figure 9] This shows the calibration data for 22 samples that were manufactured, representing the intensities of three components R, G, and B related to various concentrations of the samples. [Figure 10] This shows the colors associated with seven samples, represented in the colorimetric space CIExy 1931. [Figure 11] This shows various oxygen concentration measurements obtained using Method 1 and Method 2 according to the present invention, as well as the ex situ method using melting in an inert gas. [Figure 12] The oxygen concentration results obtained for four samples using Methods 1 and 2 of the present invention, the ex situ method by melting in an inert gas, and theoretical values ​​are shown. [Modes for carrying out the invention]

[0032] A method 20 according to the present invention for determining the oxygen concentration Cox of powder MP of a metallic material Mat in the form of a powder bed PB is shown in Figure 2. This method is based on colorimetric analysis and utilizes the color of the particles to determine the oxygen concentration of the powder, which is a good indicator of the deterioration of powder quality in metal addition manufacturing as described above.

[0033] The weight-based oxygen concentration (Cox) of a particle represents the amount of oxygen present in the particle's composition, primarily on its surface. This is typically measured in parts per million by weight (wppm).

[0034] During the oxidation of metal powder at high temperatures, an oxide surface layer is formed. Different heating conditions result in different layer thicknesses and different oxygen concentrations. The resulting color is observable and is caused by the different thicknesses of these oxide films and the interference between light reflected by the film / metal interface and light reflected by the top of the oxide (film / air interface). Therefore, the color of the particles can be correlated with their oxygen concentration.

[0035] The method according to the present invention performs direct analysis on a powder bed PB, that is, on a layer of powder having a flat surface.

[0036] This includes a first step A, which is to generate an image Im of at least a portion of the powder bed PB. This image comprises a set of pixels Pi, where i is the index of the pixel, and during imaging, the pixels of the image have a color Coli encoded by a colorimetric code containing three quantities G1, G2, and G3. For the visual trivariate system, three numbers are sufficient for color discrimination.

[0037] According to one embodiment, the colorimetric code is an RGB (or RGB additive) system, where G1=R, G2=G, and G3=B. Different colors are reproduced by adding light. Most colors can be created / characterized by the additive blending of three light rays: red (R), green (G), and blue (B). To represent / characterize a particular color, the ratio of each of the three additive primary colors RGB in its composition is determined. Adding these three colors together completely yields white. This model is very widely used because it corresponds to how color monitors work. Each primary color varies as a percentage value between 0% and 100%, or a value between 0 and 255 (8-bit coding, 24 bits total per color), and thus a particular color is specified by indicating the contribution of each primary color. Having 256 shades of each primary color allows for the creation of 16.7 million (256 × 256 × 256) colors. This coding is also used when generating images of the field's colors using a detector equipped with a color filter. When using RGB coding, pixel Pi is associated with a triplet (Ri, Gi, Bi).

[0038] However, all kinds of color coding, such as (cyan, magenta, yellow) coding, CIE XYZ coding derived from RGB coding, or CIE U'V'W', CIE L * a * b * Coding can be used in this invention.

[0039] To implement the method, in one embodiment, the initial coding of the imaging device is used, and in another embodiment, a conversion to several different coding forms is performed (typically a change in the axis system). Importantly, there is available information regarding the color of the pixels in the image of the powder bed PB, and this information is coded by value triplets.

[0040] In the second step B, the oxygen concentration of the powder is a function of the material Mat, and a predefined calibration function CF is used to link the oxygen concentration Cox with three quantities. Mat This is determined from the values ​​of three quantities related to the pixels of the image: Cox=CF Mat (G1, G2, G3)

[0041] Pj represents a pixel of an image used in the method according to the present invention. A pixel Pj can correspond to all of the pixels of the image, or to a portion of those pixels.

[0042] Therefore, the oxygen concentration Cox of the powder MP, which is imaged in the form of a powder bed PB, is determined from a set of triplets (G1j, G2j, G3j) associated with pixel Pj.

[0043] After numerous experiments and studies, the inventors developed a calibration function CF that links a specific color to a specific oxygen concentration within a target range. Mat The relationship between color and oxygen concentration (bijective relationship) was revealed. The range of this study is typically the weight-based oxygen concentration measured in wppm units within the range of [200, 2500]. Furthermore, it was shown that color measurement using commercially available imaging devices corresponds to an accurate determination of oxygen concentration. Note that a particular material Mat may correspond to multiple calibration functions.

[0044] Excellent oxygen concentration measurement results have been obtained. This is a remarkable and surprising result. The mechanism of degradation of laser-molten powder on the powder bed is actually very complex due to the interaction between the laser and the material, the physical phenomena involving convection and discharge of liquid material in the molten pool, and the uptake of particles near the laser due to back pressure. Colored particles are observed in the recycled powder, but particles with surface oxide nodules are also observed that do not have a continuous oxide film that generates color (see, for example, the aforementioned publication by Delacroix et al.). Therefore, it was not prior to be evident that oxygen measurement by colorimetric analysis to monitor the quality of additively manufactured powder would be predictable and equivalent to conventional measurements (see below).

[0045] Method 20 according to the present invention utilizes the fact that the powder bed has a planar structure, allowing for the capture of images of the powder bed. This has the advantage of enabling in-situ, non-contact determination of the oxygen concentration by acquiring images of the powder bed and determining the oxygen concentration of the powder layer by colorimetric analysis of the images. This method does not require handling of the powder, is not used with a sample, is used directly on the powder bed, and can therefore be easily integrated into processes that use a powder bed (see below).

[0046] Furthermore, the implementation requires few means, and the calibration function is determined and stored in memory using an independent method. The inventors have also shown that the calibration function can be expressed as a linear or quadratic polynomial having three variables corresponding to three quantities and coefficients that are functions of the material Mat (see below for an example of determining this calibration function).

[0047] In the first modified form, step A (generation of an image of the powder bed) is a one-step process using a video camera, while in the second modified form, which is described in detail below, step A is performed by scanning using a flatbed scanner.

[0048] In a method that does not limit the present invention, the three quantities RGB are used to code a color for the remainder of the description.

[0049] Step B of the method according to the present invention can be carried out using the two methods shown in Figure 3.

[0050] In the first method, also shown in Figure 4, a calibration function is used to obtain the oxygen concentration of each pixel, and then the average oxygen concentration of the floor is calculated by summing these values ​​over all individual pixels Pj. Thus, in step B1, for multiple pixels Pj, the corresponding oxygen concentration Coxj, known as the pixel density, is determined by the calibration function: Coxj=FC Mat (Rj, Gj, Bj)

[0051] Next, in step B2, the oxygen concentration is determined based on the average Coxj of the pixel density:

number

[0052] In the second method, also shown in Figure 5, the average values ​​(Rm, Gm, Bm) of quantities (G1j, G2j, G3j) associated with multiple pixels Pj in the image are first determined in step B'1.

[0053] Next, in step B'2, the oxygen concentration is determined from the average value of the three quantities using a calibration function. Cox=FC Mat (Rm, Gm, Bm)

[0054] The first method should be theoretically more robust because it uses a calibration function for each pixel, and therefore the average of these concentrations Coxj can be assumed to represent the physical value of the amount of oxygen present in the powder bed. The second method is faster because the image processing for determining the average amounts Rm, Gm, and Bm of the pixel set Pj is very simple, and the calibration function is used only once.

[0055] Considering the diversity of colored particles present on the powder bed of degraded raw materials, it was not a priori apparent that similar results would be obtained using the two methods. Furthermore, from a mathematical standpoint, it could be expected that there would be a significant difference if the correlation function used was not linear. However, when the inventors compared the results obtained using the two methods, surprisingly, the second method yielded results as good as the first method (see the examples below).

[0056] Since the method according to the present invention is performed in situ on a powder bed, it can be easily incorporated into a MAM-PBF method, for example, as described in the prior art. In another embodiment, the present invention relates to a metal addition manufacturing method by selective solidification of a powder bed, comprising the step of determining the oxygen concentration of the powder bed using Method 20 according to the present invention. Method 20 is performed at least once in a MAM-PBF method.

[0057] According to one embodiment, method 20 is performed at the start of the metal addition process, while inert gas is being injected into the manufacturing chamber, and / or at the end of the metal addition process, while the manufactured part is being cooled.

[0058] According to one embodiment, step A of method 20 is performed by simultaneously scanning and applying powder (see below).

[0059] Method 20 according to the present invention provides a new opportunity to inspect the quality of powder directly online, in-machine, and for an overall representative sample used directly in manufacturing.

[0060] In the MAM-PBF method, in-situ oxygen concentration measurement enables adjustments that were previously impossible. In one embodiment, the MAM-PBF method includes the step of adapting the parameters of the method as a function of the determined oxygen concentration value. This means, for example, changing the laser output, the laser scanning speed, or the distance between two laser tracks, all of which affect the local energy density applied to the material and require more energy to be applied to the powder surrounded by the oxide layer in order to remove it by melting or evaporation.

[0061] Furthermore, it is now possible to avoid using low-quality powders, i.e., powders with too high oxygen concentrations. In one embodiment, the metal additive manufacturing method includes the step of mixing the used powder with new powder when the oxygen concentration of the used powder exceeds a predetermined threshold value C0. This mixing step is performed during two manufacturing operations of the part.

[0062] According to another aspect, the present invention relates to an apparatus 10 for determining the oxygen concentration Cox of a powder MP of a metallic material Mat in the form of a powder bed PB. The apparatus 10 includes an image capture device ICD configured to generate an image Im of at least a part of the powder bed, the image Im including a set of pixels Pi, and the pixels of the image having a color encoded by a colorimetric code including three quantities (G1, G2, G3). The apparatus 10 also includes a first processing unit UT1 configured to determine the oxygen concentration of the powder from the values of the three quantities associated with the pixels of the image using a predefined calibration function CF Mat that links the oxygen concentration Cox and the three quantities as a function of the material Mat.

[0063] A first variant of the apparatus 10 according to the present invention is shown in FIG. 6, where the apparatus ICD includes an optical system Opt and a matrix detector MD. The image is captured at once. The image capture device preferably includes a high-resolution video camera.

[0064] In a second variant of the apparatus 10 according to the present invention, the apparatus ICD includes a flatbed scanner that captures an image of the powder bed by scanning.

[0065] In another embodiment, the present invention relates to a system 100 for metal addition manufacturing by selective solidification of a powder bed, as shown in Figure 7.

[0066] System 100 includes a tank Tk intended to contain a substrate Sub on which a part Pa is manufactured, an apparatus PSD for applying powder, and an apparatus CD for solidifying the powder. It preferably also includes an apparatus MD for vertically moving the substrate. System 100 also includes apparatus 10 according to the present invention, shown in a second variant in Figure 7, where apparatus ICD includes a flatbed scanner Scan. System 100 also includes a second processing unit UT2 configured to control the execution of metal additive manufacturing. The first processing unit UT1 is preferably integrated into UT2.

[0067] In one embodiment, the image acquisition device ICD of system 100 includes a flatbed scanner Scan connected to a coating device PSD. Image acquisition is then performed together with a device used for coating powder fixed to it, which is advantageous in terms of machine arrangement. Scanners for surface inspection, which are attached to additive manufacturing machines, have already been described, for example, in U.S. Patent No. 1,0981,225.

[0068] A more detailed implementation of system 100, which includes a scanner, is shown in Figure 8.

[0069] As a note, the theory of the scanner is as follows: A lamp positioned on a movable block scans the entire surface of the document / surface. This operation is performed in steps. This divides the document / surface into virtual lines, which is one step forward of the block that determines the scanner's horizontal resolution. The sensor receives the light reflected from the document / surface and defines the color of the points that make up each line. Two techniques are used in flatbed scanners, CCDs, and CIS (contact image sensors).

[0070] In a CCD scanner, a lamp emits white light, which is then reflected line by line by a set of mirrors. At the end of the movement, the beam passes through a lens, focusing the rays, which are then collected by a CCD sensor consisting of a strip of photosensitive elements. To reproduce the colors of the document / surface, red, green, and blue filters alternately cover them. The sensor measures the amount of light received line by line, converts this into an electric charge, which is then converted into digital data.

[0071] In CIS technology, the light source consists of diodes (LEDs) that emit red, green, and blue light, and the sensor is positioned across the entire width of the scanner and moves simultaneously with the LEDs. The three colors of light emitted from the LEDs are focused towards the sensor by a cylindrical lens. The LEDs, lens, and sensor are all part of the same device, one per pixel across its width.

[0072] These two technologies are compatible with the present invention. While commercially available scanners generally output colorimetric information per pixel encoded with 1 to 255 for three colors R, G, and B, CIS technology is preferred because it provides better colorimetric performance.

[0073] Tests have been conducted with multiple resolutions and have been found to be functional. Preferably, a high resolution that can resolve particle size, typically equivalent to 2400 dpi, or even 4800 dpi, is preferred for more faithful measurement of the powder bed color. Because the dispensing device is relatively fast (e.g., 50 mm / s), in one embodiment, image acquisition is performed by a scanner fixed to the dispensing device not while the powder is being properly applied, but when the dispensing device returns to its initial position at a speed that accommodates the high resolution (e.g., 0.16 mm / s).

[0074] High resolution is preferable because it allows for the resolution of particle size, and the deterioration of the powder bed is not uniform, but consists of colored particles dispersed across the surface.

[0075] Calibration function FC for implementing Method 20 according to the present invention Mat The method demonstrated by experimental examples of the determination is described below.

[0076] First, several available powder samples are needed, each having a different known oxygen concentration.

[0077] SS316L stainless steel powder is considered, and various oxidations of this powder are obtained by subjecting the sample to various combinations of time and temperature in an oven. Next, the oxygen concentration is measured ex situ by melting in an inert gas (inert gas melting), a method well known to those skilled in the art. At low concentrations, the color of the powder changes from gray, then between orange and brown, and then, as the concentration increases, the powder becomes pink, and then blue, which can be observed visually.

[0078] Next, images of each sample are generated, and the associated average color is determined by a colorimetric code containing three quantities (G1, G2, G3). Colors with associated oxygen concentration values ​​are recognized as calibration data.

[0079] This measurement was performed on the sample using a Canon CIS scanner with a resolution of 4800 dpi, and the color was measured at every pixel of the saved image. Next, the average values ​​of R, G, and B were calculated from the measurements of various pixels.

[0080] Figure 9 shows calibration data for 22 samples prepared to have oxygen concentrations within the target range of [200, 2500]wppm, representing the intensities of three components R, G, and B (each coded from 1 to 255) related to various concentrations in the samples.

[0081] As an example, the colors associated with seven samples numbered 1 through 7, selected from 22 samples (see Table I below), were calculated and are shown in Figure 10 using the commonly used colorimetric space CIExy 1931.

[0082] [Table 1]

[0083] From the positions of the seven colors associated with seven samples of 316L stainless steel powder with increasing oxygen concentrations, it can be seen that the colors actually differ across the target Cox range, changing from gray to orange, and then from pink to blue.

[0084] Finally, the calibration function is determined by regression based on the calibration data. Linear or quadratic polynomial regression is preferably used, and good results have been obtained.

[0085] The variables R, G, and B are represented by x, y, and z, respectively, in the following equation.

[0086] In the first example applied to the data from Figure 9, (1,x,y,z,xy,xz,yz,x 2 ,y 2 ,z 2 The calibration function CF1 is obtained by regression considering a polynomial of the type ) Mat It was decided: CF1 Mat=4036-49x+0.24x 2 +45y-0.60xy+0.05y 2 -18z+0.3xz+0.36yz-0.40z 2 This was obtained.

[0087] Using this first regression, the calibration coefficient R 2 The result calculated was 0.99824.

[0088] In the second example applied to the data from Figure 9, the calibration function CF2 is obtained by regression considering a polynomial of the type (1,x,y,z,xy,xz,yz). Mat It was decided: CF2 Mat =6117-95x+178y-0.77xy-127z+1.6xz-0.77yz This was obtained.

[0089] Using this second regression, the coefficient R 2 The result calculated was 0.99642.

[0090] In the third example applied to the data from Figure 9, (1,x,y,z,x 2 ,y 2 ,z 2 The calibration function CF3 is obtained by regression considering a polynomial of the type ) Mat It was decided: CF3 Mat =2906+13x-0.30x 2 -2.5y +0.40y 2 -6.22z-0.22z 2 This was obtained.

[0091] Using this third regression, the coefficient R 2 The result calculated was 0.99676.

[0092] R 2 Since the value is very close to 1, it can be seen that good results can be obtained from these three regressions.

[0093] The calibration function CF1 described aboveMat The method 20 according to the present invention was tested using the following. For this purpose, samples of new 316L stainless steel powder (R0) and powder that had been recycled a number of times (1 time (R1), 5 times (R5), 10 times (R10), and 15 times (R15)) in a MAM-PBF system (by laser melting) were available. This recycling is the subject of the study in the aforementioned publication by Delacroix et al., corresponding to the production of several parts on a tray by laser melting onto a powder bed, where all powder that had not solidified in the parts was recovered, the powder was sieved to remove the coarsest particles, and this sieved powder was reintroduced into the machine for a new production cycle without adding new powder.

[0094] The powder bed is scanned as it passes through the machine. Digital zoom of these scans highlights the varying non-uniform structure of the particles, with numerous particles coated with oxides of different colors being shown in the R15 scan.

[0095] It is preferable to use the same image acquisition device for determining the calibration function and for subsequent powder characterization.

[0096] From the images generated by scanning each powder bed R0 to R15, the oxygen concentration was determined using Methods 1 and 2, and compared with measurements performed ex situ by melting in an inert gas. Figure 11 shows the various Cox values ​​obtained.

[0097] The Cox values ​​obtained using the method according to the present invention show a good correlation with the increasing oxygen concentration when the powder is reused.

[0098] Remarkable results have been obtained showing that the results obtained by the methods according to the present invention (Methods 1 and 2) are in perfect agreement with ex situ measurements. The results obtained using Methods 1 and 2 according to the present invention are slightly higher than the ex situ results by up to 10-15 wppm, but are substantially always within the standard deviation of the chemical analysis values.

[0099] Another positive result is that the second method yields results very close to those obtained using the first method (difference of less than 5 wppm). Therefore, the second method alone can be used for the analysis of powder bed scanning because it provides oxygen concentration results almost instantaneously (100 mm). 2 In this zone, 70ms / 10cm 2 In the case of this zone, coding is performed in 7s). Nevertheless, the first method is also possible and more rigorous.

[0100] As described above, it was not a priori clear that measuring oxygen by colorimetric analysis to monitor powder quality in additive manufacturing was predictable and equivalent to conventional measurements. Furthermore, theoretical methods that begin with the assumption that color is uniquely determined by the film thickness around the particles are not always entirely accurate, and it is noteworthy and surprising in itself that the method according to the present invention can yield accurate oxygen concentration results in highly recycled powders (R10 and R15).

[0101] Furthermore, oxidation in the method is carried out in an inert gas atmosphere (generally argon, nitrogen, or helium), with a very low oxygen partial pressure and a temperature to which the unknown and non-uniform powder is exposed. Therefore, it was not even more obvious that the correlation between oxygen and the color produced by controlled oxidation in an oven in constant-temperature air (for determining the calibration function) could be similar to that observed with colored particles due to oxidation during the laser melting process on the powder bed.

[0102] To further test the robustness of the method according to the present invention, samples of "artificially" degraded powder were also analyzed. Mixtures were prepared from new powder containing different proportions (5% and 10% by weight) of powder oxidized in an oven at different oxygen levels: L1 at 1580 wppm and L2 at 2350 wppm. Thus, four samples were present: Sample [L1-5%]: 95% fresh powder - 5% powder oxidized to level L1 Sample [L1-10%]: 90% fresh powder - 10% powder oxidized to level L1 Sample [L2-5%]: 95% fresh powder - 5% powder oxidized to level L2 Sample [L2-10%]: 90% new powder - 10% powder oxidized to level L2

[0103] Figure 12 shows the results of the Cox concentration obtained for four samples using methods 1 and 2 of the present invention, the ex situ method by melting in an inert gas, and theoretical values. These theoretical values ​​represent the expected oxygen concentration of the powder mixture based on the weight fractions and oxygen concentrations of the two components.

[0104] The measurements obtained by melting the four samples in inert gas are in near perfect agreement with the theoretically calculated values.

[0105] Regarding the results obtained using methods 1 and 2 according to the present invention, both methods result in a slight overestimation compared to the values ​​obtained by chemical analysis. These tendencies persist, and the results of these two methods are substantially the same in this case as well. For a certain proportion of oxidized particles in the coating, the overestimation of Cox is more pronounced in the case of the L1 sample. The L1 sample consists of light orange particles, while the L2 sample contains blue particles. The periphery of the colored particles appears darker in the acquired image, which may lead to the overestimation. The L2 oxidized particles are already relatively dark, the edge effect is not very pronounced, and the color variation is small, which can explain why the difference measured in the case of the L2 mixture is smaller. Furthermore, at certain levels, this difference is larger in the case of a sample with 10% colored particles. As the number of colored particles increases, a more abundant and therefore more divergent edge effect is induced. Nevertheless, it is noteworthy and quite surprising that the methods according to the present invention enable a more realistic evaluation of Cox in the case of powders containing mixtures of particles of different oxidations. [Explanation of symbols]

[0106] 10 equipment 100 Systems 20 ways CD solidification equipment CFMat calibration function Cox oxygen concentration ICD Image Acquisition Device Im Image Mat metal material MP metal powder P pixel set PB powder bed PSD Powder Dispersion Apparatus Scan flatbed scanner Tk Tank UT1 First Processing Unit UT2 Second Processing Unit

Claims

1. A method (20) for determining the oxygen concentration (Cox) of powder (MP) of a metal material (Mat) in the form of a powder bed (PB), wherein: A. To generate an image (Im) of at least a portion of the powder bed, wherein the image includes a set of pixels (P), and the pixels have a color encoded by a colorimetric code including three quantities (G1, G2, G3), B A predefined calibration function (CF) which is a function of the material and links the oxygen concentration with the three quantities. Ｍａｔ Using the above, the oxygen concentration of the powder is determined from the values ​​of the three quantities related to the pixels of the image, A method that includes the steps described.

2. The method according to claim 1, wherein the colorimetric code is in the RGB system, and the three quantities are known as R, G, and B.

3. The method according to claim 1 or 2, wherein the calibration function is a linear or quadratic polynomial having three variables corresponding to the three quantities and a coefficient that is a function of the material.

4. The method according to claim 1 or 2, wherein step A is performed in one step using a video camera.

5. The method according to claim 1 or 2, wherein step A is performed by scanning using a flatbed scanner.

6. Step B is: B1. The calibration function is used to determine the oxygen concentration (COxj), known as the pixel density, for multiple pixels of the image. B2 Determining the oxygen concentration from the average value of the pixel density, The method according to claim 1 or 2, comprising the substeps described in the following:

7. Step B is: B'1. Determine the average value (G1m, G2m, G3m) of each of the quantities (G1j, G2j, G3j) associated with multiple pixels of the aforementioned image, B'2 The oxygen concentration is determined from the average value of the three quantities via the calibration function, The method according to claim 1 or 2, comprising the substeps described in the following:

8. A method for metal addition by selective solidification of a powder bed, according to claim 1 or 2, comprising the step of determining the oxygen concentration of the powder bed, which is performed at least once in the metal addition method.

9. The metal addition manufacturing method according to claim 8, wherein the method for determining the oxygen concentration of the powder bed according to claim 1 or 2 is performed at the start of the metal addition manufacturing method, while inert gas is being injected into the manufacturing chamber, and / or at the end of the metal addition manufacturing method, while the manufactured part is being cooled.

10. The metal addition manufacturing method according to claim 8, comprising the step of adapting the parameters of the method as a function of the determined oxygen concentration value.

11. The metal addition manufacturing method according to claim 8, further comprising the step of mixing used powder with new powder when the oxygen concentration exceeds a predetermined threshold, wherein the mixing step is performed between manufacturing two parts.

12. Apparatus (10) for determining the oxygen concentration (Cox) of powder (MP) of a metal material (Mat) in the form of a powder bed (PB): An image acquisition device (ICD) configured to generate an image (Im) of at least a portion of the powder bed, wherein the image includes a set of pixels (P), and the pixels have a color encoded by a colorimetric code including three quantities (G1, G2, G3), A predefined calibration function (CF) that is a function of the material and links the oxygen concentration with the three quantities. Ｍａｔ A first processing unit (UT1) is configured to determine the oxygen concentration of the powder from the values ​​of the three quantities related to the pixels of the image, using the above method. A device that includes this.

13. A computer program comprising an instruction to cause the apparatus according to claim 12 to perform the steps of the method according to claim 1 or 2.

14. A system (100) for metal addition manufacturing by selective solidification of a powder bed, comprising: The apparatus (10) according to claim 12 for determining the oxygen concentration, A tank (Tk) intended to contain the base material on which the parts are manufactured, A device (PSD) for applying the powder onto the surface of the powder bed, A device (CD) for solidifying the aforementioned powder, A second processing unit (UT2) configured to control the execution of the metal additive manufacturing process, A system that includes this.

15. The metal additive manufacturing system according to claim 14, wherein the image acquisition device includes a flatbed scanner (Scan) connected to the coating device (PSD).

16. The metal additive manufacturing system according to claim 14, wherein the image acquisition device includes a high-resolution video camera.

17. A method for determining a calibration function for carrying out the method described in claim 1 or 2, the method being: The objective is to obtain multiple samples of the aforementioned powder, wherein each sample has a different known oxygen concentration. The process involves generating images of each sample and determining the associated average color, where the color is coded using a colorimetric code containing three quantities (G1, G2, G3), and the color with the associated oxygen concentration value is known as calibration data. The calibration function is determined from the calibration data by regression, A decision method that includes the steps described below.

Images