JP6939082B2 - Powder bed evaluation method - Google Patents

Description

The present invention relates to a powder bed evaluation method, and more specifically, to irradiate the powder bed with energy rays such as a laser beam to evaluate the uniformity of distribution of the powder material in the powder bed when producing a three-dimensional model. It is about the method of.

In recent years, the development of additive manufacturing technology (AM) has been remarkable as a new technology for manufacturing three-dimensional shaped objects. As a kind of additional manufacturing technology, there is a layered manufacturing method that utilizes solidification of a powder material by irradiation with energy rays. Examples of the layered manufacturing method using a powder material include a selective laser melting method (SLM), a selective laser sintering method (SLS), and an electron beam melting method (EBM). Method can be mentioned. In these methods, a powder material composed of a resin, ceramics, metal, or a composite material thereof is supplied onto a base material to form a powder bed, and the powder is powdered based on three-dimensional design data. A laser beam, an electron beam, or other energy beam is applied to a predetermined position on the floor. Then, the powder material of the irradiated portion is solidified by melting, re-solidifying, or sintering, and a modeled body is formed. A three-dimensional model can be obtained by repeating the supply of the powder material to the powder bed and the modeling by irradiation with energy rays to sequentially form the model in layers.

In the obtained three-dimensional model, a structure in which the distribution of constituent materials such as voids is not uniform occurs at a position corresponding to the surface of each layer when the material is supplied to the powder bed and the energy rays are irradiated in layers. In some cases. It is desirable to suppress the formation of such a non-uniform structure as much as possible. For example, in Patent Document 1, a static magnetic field is applied perpendicularly to the surface of the metal powder layer to suppress marangoni convection caused by surface tension on the surface of the molten pool, thereby causing deposit surface ripple defects and the like. I am trying to eliminate.

JP-A-2017-25401

In the production of a three-dimensional model by irradiating a powder bed with energy rays, the cause of the uneven distribution of constituent materials inside the three-dimensional model is the energy as described in Patent Document 1. There can be various factors as well as the phenomenon that occurs when irradiating a line. Among them, the state of the powder bed before irradiation with energy rays can also have a great influence on the state of the obtained three-dimensional model. As a prerequisite for taking measures to prevent the state of the powder bed from affecting the state of the three-dimensional model, the state of the powder bed should be evaluated, and the state of the powder bed and the obtained three-dimensional model should be evaluated. It is important to investigate the correlation with the state of. For example, if the distribution of the powder material on the surface of the powder bed is non-uniform, the non-uniformity may be reflected as the non-uniformity of the distribution of the constituent materials in the three-dimensional model.

The problem to be solved by the present invention is to provide a powder bed evaluation method capable of evaluating the uniformity of distribution of a powder material in a powder bed when producing a three-dimensional model by irradiating the powder bed with energy rays. To do.

In order to solve the above problems, the powder bed evaluation method according to the present invention includes a powder supply step of laying powder materials in layers to form a constituent layer of the powder bed, and irradiating the surface of the constituent layer while scanning energy rays. Then, the molding step of solidifying the powder material of the irradiated portion is alternately repeated, and when the three-dimensional shaped object is manufactured, the surface of the constituent layer is between the powder supply step and the molding step. An imaging step of acquiring an image by photographing the image, and an evaluation step of acquiring evaluation data for detecting a portion where the distribution of the powder material in the constituent layer is uneven based on the image acquired in the photographing step. It is a thing.

Here, based on the images obtained in each of the photographing steps corresponding to the plurality of powder supply steps, the evaluation data is shown in the evaluation step, and the distribution of the powder material in the powder bed is shown in three dimensions. It is good to get as.

Further, the irradiation data indicating the position where the energy rays are irradiated in the modeling step is superimposed on the evaluation data obtained in the evaluation step, and the distribution of the powder material in the powder bed is displayed on the three-dimensional modeled object. It is good to correspond to the position in.

Then, in the evaluation step, it is preferable to associate the distribution of voids in the three-dimensional model with the distribution of the powder material in the powder bed.

Then, in the evaluation step, information on the distribution of the powder material is extracted by the filtering process and the location where the distribution of the powder material is non-uniform is identified by binarization with respect to the image obtained in the photographing step. And should be done in this order.

In the method for evaluating a powder bed according to the above invention, when a powder material is spread on the surface of a powder bed containing a three-dimensional model in the process of formation and a new constituent layer of the powder bed is formed, the surface of the constituent layer is formed. Is photographed, and based on the obtained image, a portion where the distribution of the powder material is uneven is detected. As a result, if the distribution state of the powder material is non-uniform in the constituent layer due to the shape of the three-dimensional modeled object, the supply conditions of the powder material, etc., the non-uniform distribution is quantitatively detected. be able to. Then, based on the obtained information, it is possible to estimate the degree of uniformity of the distribution of the constituent materials for the three-dimensional model to be manufactured, and to perform each process of laminated modeling such as powder bed formation. It is possible to verify and improve such conditions.

Here, in the case of acquiring the evaluation data as data showing the distribution of the powder material in the powder bed in three dimensions in the evaluation process based on the images obtained in each of the photographing processes corresponding to the plurality of powder supply processes. Is easy to verify and interpret the distribution state of the powder material in the powder bed in correspondence with the three-dimensional formation of the modeled object by sequentially modeling in layers.

Further, when the irradiation data indicating the position where the energy ray is irradiated in the modeling process is superimposed on the evaluation data obtained in the evaluation process and the distribution of the powder material in the powder bed is associated with the position in the three-dimensional modeled object. Can correspond to a specific position in a three-dimensional modeled object when there is a portion of the powder bed in which the distribution of the powder material is uneven. As a result, the position where the distribution of the constituent materials is non-uniform in the three-dimensional model is estimated with high accuracy, and the cause of the non-uniform distribution of the powder material in the powder bed is included in the powder bed. It is possible to accurately verify the correlation between the non-uniformity of the powder bed and the shape of the three-dimensional model, such as whether or not it is in the shape of the already formed three-dimensional model existing in the lower layer.

Then, in the evaluation process, when the distribution of voids in the three-dimensional model is associated with the distribution of the powder material in the powder bed, the three-dimensional model is caused by the non-uniformity of the distribution of the powder material in the powder bed. It is possible to estimate the existence of voids in the powder bed and to verify the correlation between the non-uniformity of the distribution of the powder material in the powder bed and the voids in the three-dimensional model.

Then, in the evaluation step, for the image obtained in the photographing step, extraction of information on the distribution of the powder material by the filtering process and identification of the portion where the distribution of the powder material is non-uniform by binarization are performed in this order. When this is done, the non-uniform distribution of the powder material in the powder bed can be detected with high accuracy and easily.

It is a figure explaining an example of the laminated modeling apparatus which applies the powder bed evaluation method which concerns on one Embodiment of this invention. It is an example of a photographed image of the surface of the powder bed, (a) is a good surface, (b) is a surface where the exposure of the three-dimensional model is seen, and (c) is the exposure of the three-dimensional model and the powder bed. Shows the surface where the depression is seen. It is a figure explaining the creation of a three-dimensional distribution image and a cross-sectional image, (a) shows a state in which a plurality of photographed images are laminated, (b) shows a three-dimensional image, and (c) shows a cross-sectional image near a surface. .. It is a figure explaining the quantitative analysis in a cross-sectional image, (a) is an unprocessed image, (b) is an image after filtering, (c) is an image after binarization, (d) is a third order of unevenness distribution. It is the original display. It is a figure which shows the design shape of the sample shaped object produced in an Example, (a) shows a block-shaped sample shaped object, and (b) shows a beam-shaped sample shaped object. It is a figure which shows the comparison of the laminated image which shows the unevenness of the distribution of a powder material, and the cross-sectional photograph of the obtained 3D modeled object in the case of using each powder material of an Example. It is a figure which shows the relationship between the unevenness ratio in a powder bed, and the porosity in a three-dimensional modeled object. It is a figure which shows the modification of the design shape of a sample model, (a) shows a block-shaped sample model, and (b) shows a beam-shaped sample model.

Hereinafter, the powder bed evaluation method according to the embodiment of the present invention will be described in detail. The powder bed evaluation device according to an embodiment of the present invention irradiates a powder bed on which a powder material is spread with energy rays such as a laser beam and an electron beam in a predetermined two-dimensional pattern to form a modeling layer (layered modeling body). ) Is sequentially repeated to create a three-dimensional model, and the distribution state of the powder material on the powder bed is evaluated.

[Laminate modeling equipment]
First, an example of the laminated modeling apparatus 1 to which the powder bed evaluation method according to the embodiment of the present invention is applied will be briefly described.

FIG. 1 shows an outline of the laminated modeling apparatus 1. The laminated modeling device 1 performs laminated modeling on a powder material by a selective laser melting method (SLM). The laminated modeling device 1 includes a powder bed container 10, a powder supply device 20, a laser irradiation device 30, a camera 40, and a control device 50. In the laminated modeling device 1, the powder bed B is formed in the powder bed container 10 by using the powder supply device 20, and the surface of the powder bed B is irradiated with a laser beam by using the laser irradiation device 30 to perform three-dimensional modeling. Manufacture thing P.

The powder bed container 10 has a base material 11 and a side wall 12, and can accommodate the powder material and hold the powder bed B inside the space surrounded by the base material 11 and the side wall 12. The base material 11 serving as the bottom surface of the powder bed container 10 is driven by a driving device (not shown) to move in the vertical direction M1.

The powder supply device 20 includes a tank 21, a powder supply path 22, and a blade 23. The powder supply path 22 and the blade 23 are integrally held by a common holding portion 24, and an in-plane movement M2 of the holding portion 24 in a horizontal plane is driven by a driving device (not shown). The tank 21 stores the powder material and can supply the powder material to the powder supply path 22 at a predetermined supply rate. The powder supply path 22 has an opening at the lower end, and the powder material supplied from the tank 21 can be discharged from the opening to the powder bed container 10 by gravity. The blade 23 is a plate-shaped member made of an elastic material such as rubber and having an edge at the tip, and is moved in a horizontal plane so as to slide with the edge in contact with the surface of the powder bed B. , The surface of the powder bed B can be smoothed (squeezing).

When the powder material is supplied to the powder bed container 10 using the powder supply device 20, the powder is driven while driving the in-plane motion M2 of the powder supply device 20 so that the powder supply path 22 is in the front and the blade 23 is in the rear. The powder material is discharged from the opening of the supply path 22. As a result, the layer of powder material discharged from the powder supply path 22 is squeezed by the blade 23 immediately after the discharge.

The laser irradiation device 30 includes a laser unit 31 and a galvano scanner 32. The laser unit 31 generates a laser beam by laser oscillation. The wavelength and energy density of the laser beam may be selected so that the powder material used can be melted. Further, the beam diameter may be set according to the spatial accuracy required for the three-dimensional model P to be manufactured.

The galvano scanner 32 is composed of a combination of optical components such as a mirror, and can irradiate a laser beam emitted from the laser unit 31 at a predetermined position on the surface of the powder bed B formed in the powder bed container 10. In the galvano scanner 32, the irradiation position of the laser beam on the surface of the powder bed B can be two-dimensionally scanned by adjusting the alignment (angle, arrangement, etc.) of each optical component. Design data, for example, three-dimensional CAD data in which a position to irradiate a laser beam is set according to the design shape of the three-dimensional model P to be manufactured is input to the galvano scanner 32 from a control means 50 including a computer or the like. Then, the laser beam is scanned according to the design data.

The camera 40 is attached above the powder bed container 10 and can take an image of the surface of the powder bed B. The camera 40 may be capable of obtaining an image reflecting the distribution of powder on the surface of the powder bed B, and can obtain the entire surface of the powder bed B within one field of view by photographing at a fixed point. Is preferable. The camera 40 outputs the obtained captured image to the control device 50. In addition, in this kind of laminated modeling apparatus 1, for the purpose of monitoring defects and abnormalities in the constituent members of the apparatus and the powder bed during modeling, and confirming whether the modeling is proceeding normally, etc. A camera for photographing the surface of the powder bed B is often provided, and here, such a provided camera may be used as it is. Further, in order to assist the camera 40 in photographing the powder bed B, a lighting device such as a flash device that irradiates the surface of the powder floor B with light may be appropriately provided.

Next, a method of manufacturing the three-dimensional model P using the laminated modeling device 1 described above will be described. Prior to production, a powder material as a raw material for the three-dimensional model P is prepared in the tank 21 of the powder supply device 20. As the powder material, a material made of metal, ceramics, cermet or the like can be used. Preferably, one made of metal is used. The specific component composition of the powder material may be selected according to the material of the three-dimensional model P to be manufactured. The particle size of the powder may be appropriately selected in consideration of the spatial accuracy required for the three-dimensional model P and the ease of forming the powder bed B, but is generally used as a material powder in SLM. The average particle size of the above is about 15 to 60 μm.

The production of the three-dimensional model P is sequentially performed in layers. That is, in the powder bed B, a step of creating a constituent layer B1 in which a powder material is filled in a thin layer and irradiating the surface of the constituent layer B1 with a laser beam to form a forming layer which is a layered solidified material is performed. repeat. By sequentially laminating such modeling layers, a three-dimensional model P having a three-dimensional shape can be completed.

Specifically, first, in the powder supply step, the constituent layer B1 of the powder bed B is formed on the base material 11 of the powder bed container 10. At this time, the powder material is discharged from the powder supply path 22 of the powder supply device 20, and the surface of the portion where the powder material is discharged is skied with the blade 23, and the powder material is spread in layers.

When the formation of the constituent layer B1 is completed, the camera 40 executes a photographing step of photographing the surface of the constituent layer B1 and acquiring an image. The acquired image is sent to the control device 50.

Then, the surface of the formed constituent layer B1 is irradiated with a laser beam from the laser irradiation device 30, and the modeling step is executed. At this time, based on the design data input from the control device 50, irradiation is performed while scanning the laser beam in a predetermined two-dimensional pattern. The design data used here is three-dimensional CAD data or the like obtained by dividing the shape of the three-dimensional model P to be manufactured into a thin layer structure corresponding to each layer of the constituent layer B1. In the in-plane portion of the constituent layer B1 of the powder bed B that has been irradiated with the laser beam, substances constituting the powder material, such as metal, are heated by the laser beam and melted. When the irradiation of the laser beam at the site is completed, the irradiated site is cooled, and the once melted substance solidifies (recoagulates) at the site. As a result, a modeling layer corresponding to the constituent layer B1 for one layer is formed. The formed modeling layer has a shape corresponding to the two-dimensional pattern obtained by scanning the laser beam according to the design data in the plane of the constituent layer B1.

As described above, the formation of one modeling layer constituting the three-dimensional modeling object P is completed. In this state, the formed modeling layer is surrounded by the powder material and encapsulated in the powder bed B.

Next, the powder supply step is executed again. At this time, first, the motion M1 of the base material 11 of the powder bed container 10 is driven to lower the position of the entire powder bed B downward. Then, the powder material is discharged and squeezing is performed by the blade 23 to form a new constituent layer B1. At this time, the new constituent layer B1 is laminated on the surface of the existing powder bed B containing the previously formed modeling layer. The distance for moving the base material 11 downward before forming the new constituent layer B1 may be set so that the height position of the surface of the powder bed B does not change before and after forming the new constituent layer B1. If the thickness of the constituent layer B1 to be formed is the same each time, the base material 11 may be moved downward by a distance equal to the thickness of the constituent layer B1. The thickness of the constituent layer B1 is preferably set to be equal to or larger than the average particle size of the material powder, and is generally about 20 to 200 μm in the SLM method.

When the new constituent layer B1 is completed, the photographing step is executed again, the surface of the constituent layer B1 is photographed, and the image obtained by the photographing is transferred to the control device 50. As described above, due to the movement of the base material 11, the height position of the surface of the constituent layer B1 is the same as that in the previous shooting process, so that the shot images of substantially the same quality under the same shooting conditions. Can be obtained.

Then, the surface of the constituent layer B1 is irradiated with a laser beam again to execute the modeling step. At this time, the laser beam scan is based on the data corresponding to the layer above the layer formed in the previous modeling process among the design data consisting of the shape of the three-dimensional model P divided into thin layer structures. Will be done. At this time, if the shape of the next lower layer and the shape of the newly formed layer are continuous in the design data, the melting and re-solidification of the powder material by the laser irradiation this time has already been formed, and the powder bed A new modeling layer is formed in a state where it occurs directly above the surface constituting the three-dimensional modeling object P contained in B and is joined to the existing three-dimensional modeling object P. As described above, due to the movement of the base material 11, the height position of the surface of the constituent layer B1 is the same as that in the previous molding process. Therefore, if the laser irradiation is performed under the same conditions, the powder material is melted. And re-coagulation proceed substantially in the same way.

In this way, each process is repeated many times in the order of powder supply process → photographing process → modeling process → downward movement of the base material 11 → powder supply process → ... By doing so, a three-dimensional model P having a three-dimensional shape can be manufactured. Here, the photographing step is carried out corresponding to the formation of each modeling layer, and the photographed image of the surface of the constituent layer B1 of the powder bed B formed in each powder supply step is obtained. In the present laminated modeling apparatus 1, the powder bed evaluation method described below is executed for those photographed images, and the distribution state of the powder material in the powder bed B used for manufacturing the three-dimensional model P is evaluated. ..

The additive manufacturing device 1 described here performs additive manufacturing by the selective laser melting method (SLM), but powder materials such as the selective laser sintering method (SLS) and the electron beam melting method (EBM) can be used. Even when other additive manufacturing methods used as raw materials are used, if a camera 40 capable of photographing the surface of the powder bed B is provided in each device, the powder bed evaluation method described below can be similarly applied. .. In the case of SLS, a material made of a resin material is mainly used as the powder material, the average particle size is about 20 to 80 μm, and the thickness of the constituent layer B1 of the powder bed B is about 60 to 180 μm. Further, in the case of EBM, a material made of metal is mainly used as the powder material, the average particle size is about 45 to 110 μm, and the thickness of the constituent layer B1 of the powder bed B is about 50 to 200 μm. Further, in the laminated molding apparatus 1 described here, the powder material is supplied to the powder bed B from above the powder bed B, and the surface of the powder bed B is smoothed by the blade 23. In the form in which the powder material is supplied from below the powder bed B by pushing up the material 11 and in the form in which the surface of the powder bed B is smoothed by a roller, the surface of the powder bed B is photographed by each device. If a capable camera 40 is provided, the powder bed evaluation method described below can be similarly applied.

Further, in the design data that defines the shape of the three-dimensional model P to be manufactured, in addition to the three-dimensional model P that is finally desired to be manufactured, a structure for assisting the manufacture of the three-dimensional model P is appropriately added. May be good. For example, FIGS. 5 (a) and 5 (b) show the design shapes adopted in the later examples, which extend downward from the bottom of the block-shaped and beam-shaped three-dimensional shaped objects P, respectively. A tooth-shaped support p1 is provided. The support p1 contributes to prevent deformation of the three-dimensional modeled object P during modeling. Further, the occurrence of defects in the three-dimensional model P can also be suppressed by the support p1. The support p1 is integrally modeled with the three-dimensional modeled object P based on the design data in the modeling process. After completing all of the plurality of modeling steps, the support p1 is appropriately removed to finally obtain a three-dimensional model P having a desired shape. Further, FIGS. 8A and 8B show, as a modification, a form in which a thin plate-shaped mesh portion p2 connecting between the supports p1 is provided in addition to the support p1. By providing the mesh portion p2, it becomes easy to remove the support p1.

[Powder bed evaluation method]
Next, the powder bed evaluation method according to the embodiment of the present invention will be described. In this powder bed evaluation method, when the three-dimensional model P is manufactured by repeating the powder supply step, the photographing step, and the modeling step in order using the laminated molding apparatus 1 described above, the powder bed B is configured in the powder supply step. Each time the layer B1 is formed, the distribution state of the powder material in the powder bed B is evaluated based on the photographed image obtained by photographing the surface of the constituent layer B1 with the camera 40 in the photographing step. The captured image acquired by the camera 40 is sent to the control device 50, and the control device 50 performs an evaluation step on the captured image.

(Non-uniformity of distribution of powder material on powder floor surface)
In the laminated modeling apparatus 1, the powder material is supplied and the modeling is performed by laser irradiation in a layered manner, and a new constituent layer B1 is formed on the surface of the powder bed B containing the three-dimensional model P that has already been formed halfway. It is laminated. Therefore, the three-dimensional model P contained in the powder bed B may affect the state of the newly formed constituent layer B1.

For example, in a portion where the three-dimensional model P, which has already been formed halfway and is contained in the powder bed B, protrudes upward, the powder material is sufficiently distributed in the next powder supply step as compared with other portions. It may not be done. In an extreme case, the upper surface of the three-dimensional model P may have an exposed portion D1 exposed on the surface of the constituent layer B1 to be formed next. Alternatively, the powder material may be laminated on the portion of the protruding three-dimensional model P, and a thick layer of the powder material may be formed only at that portion. On the contrary, in the part where the upper surface of the three-dimensional modeling part included in the powder bed B is low, even if the powder material is supplied in the same way as other parts in the next powder supply step, only that part is used. A depressed portion (egre) D2 in which the surface of the constituent layer B1 is recessed downward may be formed. Even after squeezing with the blade 23, it is often not possible to completely eliminate the non-uniform distribution of the powder material such as the exposed portion D1 and the depressed portion D2. Further, during squeezing, the blade 23 forms a streak-like non-uniform distribution on the surface of the constituent layer B1, and the blade 23 is caught on the already formed three-dimensional model P, so that the blade 23 is squeezed. In some cases, the non-uniformity of the distribution of the powder material may be increased.

(Construction of evaluation data)
If there is a non-uniform distribution of the powder material such as the exposed portion D1 and the depressed portion D2 in the constituent layer B1 of the powder bed B, such a non-uniform distribution will appear in the photographed image acquired by the camera 40 in the photographing process. It will be reflected. FIG. 2 shows an example of a photographed image. Here, metal powder is used as the powder material. In the image of FIG. 2A, there is substantially no non-uniform distribution of the exposed portion D1 and the depressed portion D2, and the powder material is squeezed by the blade 23, and the powder material has a high uniformity and is a powder bed. It is distributed throughout the plane of B. In the image of FIG. 2B, a large number of structures in which bright and dark parts are striped inside a long and narrow region having an inclination with respect to the side of the image can be seen in the part surrounded by the broken line. Has been done. These structures correspond to the exposed portion D1 in which the metal surface corresponding to the upper surface of the three-dimensional model P, which has already been formed halfway, is exposed. In the image of FIG. 2C, the exposed portion D1 observed as a point-like bright portion in the region indicated by reference numeral A and the depressed portion observed as a darkened portion in the region indicated by reference numeral B. D2 and are confirmed. As described above, by using the photographed image acquired by the camera 40, the uniformity of the distribution of the powder material in the constituent layer B1 of the powder bed B can be evaluated.

In this powder bed evaluation method, evaluation data is constructed by detecting a portion of the powder bed B in which the distribution of the powder material is uneven, based on the photographed image of the constituent layer B1 as shown in FIG. An example of a method for constructing evaluation data will be described below.

(1) Creation of three-dimensional distribution image Three-dimensional distribution showing the distribution of the powder material in the powder bed B in three dimensions based on the two-dimensional photographed image of the surface of each constituent layer B1 as shown in FIG. You can get the image. FIG. 3 shows an example of creating a three-dimensional distribution image.

In order to obtain a three-dimensional distribution image, in each of the photographing steps corresponding to a plurality of powder supply steps, the two-dimensional photographed images obtained for each constituent layer B1 are made to correspond to the stacking order of the constituent layers B1. Then, the images are sequentially laminated to obtain a laminated image. An example of the laminated image is shown in FIG. 3 (a). In the figure, the direction of squeezing is indicated by an arrow. From the viewpoint of analyzing the distribution of the powder material in the powder bed B with high accuracy, it is preferable to construct the laminated image using the photographed images corresponding to all the constituent layers B1, but when the data capacity is limited or the like. May form a laminated image through thinning out of one photographed image for each of a plurality of images and averaging of images used by averaging the images for each of a plurality of images.

FIG. 3B shows a three-dimensional distribution image obtained by three-dimensionally displaying the laminated image obtained above. This corresponds to the distribution state of the powder material in the actual powder bed B formed as an aggregate of the constituent layers B1 sequentially laminated in the laminated modeling apparatus 1. In FIG. 3B, many structures having irregularities inside the elongated region are seen on the surface. On the surface of the actual powder bed B obtained from this three-dimensional distribution image, a structure having irregularities is formed by the surface of the powder material and the three-dimensional model P exposed on the surface. Such a surface condition is reproduced. The three-dimensional distribution obtained here does not indicate the distribution of the powder material in the powder bed B finally obtained by laminating the constituent layers B1 in multiple layers while sandwiching the modeling process by laser irradiation. The distribution of the powder material in the constituent layer B1 at the time when each constituent layer B1 is formed is accumulated. Both distributions may match, but generally they do not.

The distribution of the powder material in the cross section of the powder bed B can be analyzed by creating a cross-sectional image cut at an arbitrary position with respect to the obtained three-dimensional distribution image. In the three-dimensional distribution image of FIG. 3 (b), the cross section is displayed at the end portion, and FIG. 3 (c) shows an enlarged cross-sectional image of a portion near the surface of the end portion. According to the cross-sectional image of FIG. 3C, in the portion where the uneven non-uniform structure is seen on the surface of the three-dimensional distribution of the material powder, the non-uniform structure in which the portion brighter and the darker than the surroundings are mixed even below. Extends in a columnar shape. That is, it can be seen that the structure in which the distribution of the powder material is non-uniform exists in a columnar shape. It is interpreted that this is due to the influence of the three-dimensional model P formed as a columnar body.

(2) Quantitative Evaluation of Non-uniform Distribution of Powder Material Next, a method of quantitatively evaluating the non-uniform distribution of the powder material in the powder bed B by analyzing the obtained three-dimensional distribution image will be described. It is possible to visually recognize that there is non-uniformity in the distribution of the powder material with respect to the two-dimensional or three-dimensional image obtained above, but the non-uniformity should be analyzed quantitatively. Therefore, the distribution state of the powder material in the powder bed B can be analyzed more clearly.

FIG. 4A shows a cross-sectional image created with respect to the three-dimensional distribution image obtained in the same manner as described above. In FIG. 4A, no processing or operation is performed on the image other than stacking and three-dimensionalizing the photographed image taken by the camera 40 and cutting out the cross section, resulting in an unprocessed cross-sectional image. The image of FIG. 4A was obtained for a powder bed when forming a beam-shaped sample using powder A as a raw material in a later example.

The captured image as it is captured by the camera 40 is non-essential, not derived from the distribution state of the powder material on the powder bed B, such as shadows, light reflection on the surface of the powder bed B, and noise during photographing. The elements are superimposed. Such elements are carried over to the three-dimensional distribution image and the cross-sectional image. The presence of such non-essential elements in the image may hinder the quantitative analysis of the distribution of the powder material. Therefore, it is preferable to remove such non-essential elements from the image and perform a process of extracting essential information derived from the distribution of the powder material from the image. Such processing can be performed, for example, by image filtering processing. The filtering process is an operation of removing or extracting a predetermined structure in an image by processing the image data according to a predetermined algorithm. Filtering methods for removing shadows, reflections, noise, etc. in an image are known, and they can be applied here. Further, in each constituent layer B1 of the powder bed B, when the metal surface of the already formed portion of the three-dimensional model P is exposed, the structure in the image derived from such a metal surface is used as the powder material. It can also be distinguished from the derived structure by filtering.

The unprocessed image of FIG. 4 (a) is filtered to extract essential information derived from the distribution of the powder material, which is shown in FIG. 4 (b). Here, the contrast of the entire image is clearer than that of the unprocessed image of FIG. 4A. In the image, the brighter the image is, the higher the distribution density of the powder material is.

Next, in the powder bed B, a process of identifying and extracting a portion where the powder material is unevenly distributed is performed. For that purpose, the binarization process can be applied to the image. In the binarization process, a threshold value of the signal strength that considers the distribution of the powder material to be non-uniform is set, and for each data point in the image, the signal strength is equal to or higher than the threshold value (or exceeds the threshold value). It is discriminated whether it is below the threshold value (or below the threshold value) and replaced with two levels of data (for example, 0 and 1) indicating each state. Thereby, it is possible to determine whether or not the non-uniform distribution of the powder material occurs at each data point. When the signal strength is above the threshold value (or above the threshold value), it is considered that a non-uniform distribution has occurred, or when it is below the threshold value (or below the threshold value), it is considered that a non-uniform distribution has occurred. Whether it is considered is the type of non-uniform distribution that can occur, for example, whether the part where the density of the powder material is lower than other parts is likely to occur, or conversely, the part where the density is high is likely to occur, and the type of powder material. , Since it depends on the shooting conditions of the camera 40 and the like, it may be appropriately selected.

The result of binarizing the image of FIG. 4 (b) is shown in FIG. 4 (c). Here, in the darkly displayed region, the distribution density of the powder material is lower than in the brightly displayed region. Since the brightly displayed area and the darkly displayed area are mixed, it can be considered that the powder material has a non-uniform distribution (unevenness) in the powder bed B. In this way, after the binarization treatment, it is possible to identify a portion in the powder bed B where the distribution of the powder material is non-uniform.

The distribution of unevenness of the powder material obtained with respect to the cross-sectional image as described above can also be displayed three-dimensionally. The cross-sections corresponding to each position of the powder bed B are similarly filtered and binarized, and the obtained cross-sectional images are arranged and displayed three-dimensionally to display the uneven distribution. It can be visualized three-dimensionally. FIG. 4D shows an example of the three-dimensional distribution of unevenness obtained in this way. Here, the area shown in dark gray indicates that the density of the powder material is lower than that of other parts. It can be seen that such regions are distributed throughout the powder bed B.

Furthermore, the abundance of unevenness can be quantitatively estimated in the two-dimensional image (FIG. 4 (c)) or the three-dimensional image (FIG. 4 (d)) obtained through filtering and binarization. The area or volume ratio of the portion determined to have unevenness to the entire powder bed B in the image through binarization may be calculated. For example, in the second-dimensional image of FIG. 4C, the area occupied by the region where the density of the powder material is lower than the other regions displayed in dark is 24.9%.

Here, a mode is described in which a three-dimensional distribution image is created from a captured image obtained by the camera 40, a cross-sectional image is obtained, and unevenness is quantified by image processing on the cross-sectional image. Such image processing may be performed on a two-dimensional captured image obtained by the camera 40. In this case, the image that has undergone the processing may be laminated, three-dimensionalized, and a cross-sectional image may be created. Alternatively, the three-dimensional distribution image formed based on the two-dimensional image obtained by the camera 40 may be subjected to image processing as it is in three dimensions.

(Correspondence between the distribution of powder material in the powder bed and the shape of the three-dimensional model)
As described above, after evaluating the uniformity of the distribution of the powder material in the powder bed B as a raw material for forming the three-dimensional model P, the shape of the three-dimensional model P to be manufactured is further evaluated. Can be associated with.

As described above, in the modeling process for each layer, the position to irradiate the laser beam is performed based on design data such as three-dimensional CAD indicating the shape of the three-dimensional modeled object P to be manufactured. Therefore, by superimposing the design data on the distribution of the powder material obtained above, the distribution of the powder material on the powder bed B can be associated with the position on the manufactured three-dimensional model P. That is, it is possible to associate what kind of portion of the entire three-dimensional model P is the portion of the powder bed B in which the distribution of the powder material is uneven.

In the powder bed B, the uneven portion where the distribution density of the powder material was low became low in the density of the constituent material or the constituent material was missing when the three-dimensional model P was formed by receiving laser irradiation. There is a high probability that it will form a void. Utilizing this, the distribution of the powder material in the powder bed B can be associated with the distribution of voids in the three-dimensional model P. Specifically, the distribution of the powder material on the powder bed B and the design data indicating the irradiation position of the laser beam are superimposed, and a gap is formed in the three-dimensional model B in the portion where the density of the powder material is low in the powder bed B. It is possible to estimate the position where the voids are formed in the three-dimensional model P and the ratio of such voids to the three-dimensional model P by considering that the site is to be formed.

As described above, there is a correlation between the uneven distribution of the powder material in the powder bed B and the density of the constituent materials in the produced three-dimensional model P, which will be shown in later examples. However, since the powder material constituting the powder bed B is once melted and re-solidified to become a constituent material of the three-dimensional model P, the ratio of unevenness (porosity) in the powder bed B remains the same as the three-dimensional model P. In, it does not mean that it is the ratio of voids (porosity) in which the constituent material is not filled. Further, since the residual stress in the three-dimensional model P also affects the formation of the voids in the three-dimensional model P, the degree of void formation also depends on the specific shape of the three-dimensional model P. However, by a preliminary test or the like of forming a three-dimensional model P having the same or similar shape using a powder bed B having various distribution states, the unevenness ratio in the powder bed B and the voids in the three-dimensional model P If the correlation of the rates is quantitatively estimated, the void rate in the three-dimensional model P can be quantitatively estimated from the unevenness rate in the powder bed B. The porosity of the three-dimensional model P used for associating with the unevenness rate can be estimated by analyzing a cross-sectional photograph of the three-dimensional model P, measuring the density, and the like.

(Use of information obtained by evaluating powder beds)
As described above, the uniformity of the distribution of the powder material in the powder bed B is analyzed by analyzing the distribution of the powder material in the powder bed B based on the photographed image of the surface of the constituent layer B1 constituting the powder bed B. Can be evaluated. In particular, in the powder bed B, a portion where the distribution of the powder material becomes non-uniform can be detected, and the non-uniformity can be quantitatively evaluated. If there is unevenness in the distribution of the powder material in the powder bed B which is the raw material of the laminated molding, the unevenness is inherited as the unevenness of the constituent materials in the three-dimensional model P obtained by irradiating the powder bed B with a laser. Probability is high. Therefore, the uniformity of the distribution of the constituent materials in the manufactured three-dimensional model P can be estimated based on the information on the distribution of the powder material in the powder bed B obtained by the evaluation. In order to investigate the density distribution of the constituent materials inside the manufactured 3D model P, it is necessary to cut and observe the 3D model P or perform non-destructive analysis using an analyzer. However, since the density distribution of the constituent materials in the three-dimensional model P can be estimated from the distribution state of the powder material in the powder bed B, it becomes unnecessary to perform such an analysis. In addition, when the constituent layers B1 are sequentially formed and laser irradiation is performed to produce the three-dimensional model P in layers, the distribution of the powder material on the powder bed B is analyzed in real time, and it is found that there is a lot of unevenness. Judges that a high-quality three-dimensional model P cannot be obtained even if the modeling is continued, and stops the modeling. It is also possible to use the evaluation.

In addition, by detecting unevenness in the distribution of the powder material in the powder bed B and quantitatively evaluating it, the conditions for forming the powder bed B can be verified, and the cause of such unevenness can be analyzed. Based on the analysis result, the conditions for forming the powder bed and the like can be improved to reduce unevenness in the powder bed B. The causes of uneven distribution of the powder material in the powder bed B include the characteristics of the powder material (particle size, particle shape, surface treatment, etc.), the supply speed of the powder material, the configuration of the blade 23 (material and shape), and squeezing. The moving speed of the blade 23 at the time, the thickness of each constituent layer B1 and the like can be mentioned. For example, by changing one of these parameters and comparing the distribution unevenness of the powder material in the powder bed B, it is estimated what kind of parameter causes the non-uniformity of the distribution. By adjusting the parameters, it is possible to reduce the uneven distribution of the powder material. Further, as described above, the shape of the already formed lower layer portion of the three-dimensional model P often affects the formation of the exposed portion D1 and the depressed portion D2 in the constituent layer B1 of the powder bed B. If possible, it may be possible to reduce the uneven distribution of the powder material by changing the design shape of the three-dimensional model P.

On the other hand, although the distribution of the powder material on the powder bed B is highly uniform, the porosity may be high in the obtained three-dimensional model P. In such a case, it can be estimated that factors other than the distribution of the powder material in the powder bed B cause voids in the three-dimensional model P. Examples of such factors include the physical characteristics of materials such as metals constituting the powder material and the irradiation conditions of the laser beam. In such a case, it is sufficient to consider changing the type of powder material to be used and the irradiation conditions (wavelength and energy density) of the laser beam.

In the powder bed B, any specific analysis method may be used for detecting unevenness in which the distribution of the powder material is uneven based on the photographed image of each constituent layer B1. By using a method of extracting the essential image information showing the distribution of the powder material by the filtering process and then identifying the unevenness by the binarization process, the presence of the unevenness in the powder bed B can be easily and accurately determined. Can be detected. In addition, it is easy to quantify the unevenness rate in the powder bed B. Since the surface of each of the constituent layers B1 of the powder bed B is photographed, the amount of photographed images to be processed may be enormous, but it is automatically uneven by the filtering process and the binarization process. If is identified, the data processing of the captured image can be collectively and efficiently executed. Further, even if the type of powder material constituting the powder bed B and the shooting conditions with the camera 40 change, the parameters used for the filtering process and the threshold value at the time of binarization can be adjusted by the same method for unevenness. Can be quantitatively detected.

The evaluation of the distribution of the powder material in the powder bed B can be performed on the two-dimensional photographed images obtained for each constituent layer B1 of the powder bed B, but such two-dimensional images are laminated and obtained. By evaluating the three-dimensional information to be obtained, the distribution of the powder material in the powder bed B can be evaluated in correspondence with the three-dimensional modeling object P being formed three-dimensionally by laminating the modeling layers. , It becomes easy to verify and interpret the influence on the obtained three-dimensional model P. Furthermore, by superimposing the design data used for laser irradiation and the three-dimensional distribution image of the powder material on the powder bed B and making it correspond to the shape of the three-dimensional model P and the distribution of the powder material on the powder bed B, the powder as a raw material It is possible to accurately estimate the location where the distribution of the constituent materials of the three-dimensional model P is uneven due to the uneven distribution of the materials. In addition, the correlation between the distribution of the powder material on the powder bed B and the shape of the three-dimensional model P can be accurately verified, and the uneven distribution of the powder material on the powder bed B can be improved and the three-dimensional model P can be used. It is possible to improve the uniformity of the constituent materials. In particular, by quantitatively associating the unevenness ratio in the powder bed B with the void ratio in the three-dimensional model P, the existence position and the ratio of the voids in the three-dimensional model P caused by the unevenness of the powder bed B are quantitatively associated. To estimate, to examine the correlation between the distribution of the powder material in the powder bed B and the voids in the three-dimensional model P, and to improve the conditions of each process related to the laminated modeling based on the examination results. Is possible.

Hereinafter, the present invention will be described in more detail with reference to Examples.

[Test method]
(Preparation of sample model)
Using the SLM method, a sample model was prepared using metal powder as a raw material. A metal 3D printer (SLM280HL) manufactured by SLM Solutions was used to prepare the sample model.

As the material powder, the following four types of metal powder were prepared.
-Powder A: Fe-based alloy gas spray powder (average particle size: 45 μm)
-Powder B: High-pressure water spray powder of Fe-based alloy (average particle size: 45 μm)
-Powder C: Powder A mixed with 0.05% nanosilica-Powder D: Ni-based alloy powder (average particle size: 45 μm)

FIG. 5 shows the shape of the sample model based on the three-dimensional CAD data used for laser irradiation when the sample model is produced. Using each material powder, two types of sample shaped objects, a block-shaped sample shown in FIG. 5 (a) and a beam-shaped sample shown in FIG. 5 (b), were prepared. Each sample is provided with a support (p1) at the bottom.

When preparing the sample model, the thickness of each constituent layer to be sequentially laminated on the powder bed was set to 50 μm. Further, the energy density of the laser beam irradiated at the time of modeling is shown in FIG.

(Evaluation of unevenness of powder bed)
The photographed images acquired for each constituent layer of the powder bed during the preparation of the sample model are laminated to form a three-dimensional distribution image, which corresponds to the design data as shown in FIG. 5, and the AA cross section in FIG. And a cross-sectional image of the position corresponding to the BB cross section was created. Then, the cross-sectional image was filtered. In FIG. 6, the processed image is shown in the “stacked image” column. In the region corresponding to the cross section shown by the broken line in the image, the higher the uniformity of the brightness of the data points, the higher the uniformity of the distribution of the powder material in the powder bed. For example, in the beam-shaped sample of powder A, bright parts and dark parts are mixed in a streak pattern in the region, indicating that the distribution of the powder material is uneven and there are many irregularities. ing. Further, after performing the binarization process, the area of the region that appears dark in the image was estimated, and the unevenness ratio was calculated as a ratio to the area of the entire cross section.
The value of the obtained unevenness rate is shown in the column of "stacked image" in FIG.

(Evaluation of voids in the sample model)
Each sample model actually produced was cut at a position corresponding to the AA cross section and the BB cross section in FIG. 5, and a cross section was obtained. In FIG. 6, in the column of "cross section", a photograph of those cross sections is shown. In the photograph, the region that is darkly observed like a hole in the cross section is a void. The total area of such voids was estimated from the photograph, and the porosity was calculated as a ratio to the area of the entire cross section. The obtained porosity values are shown in the "Cross section" column in FIG.

[Results and discussion]
(Correlation between the unevenness rate of the powder bed and the porosity of the sample model)
FIG. 6 shows a cross-sectional image (laminated image) showing the distribution of the powder material in the powder bed and a cross-sectional photograph of the actually produced sample model in the case where a sample model of each shape is produced using each metal powder. The comparison of (cross section) is shown. Both correspond to the AA cross section and the BB cross section in FIG.

Looking at FIG. 6, except for the case where the energy density of the powder D is 25.0 J / mm ³ , the degree of mixture of light and dark is large in the laminated image of the powder bed, and the larger the uneven distribution of the powder material in the powder bed, the greater the difference. In the cross-sectional photograph of the sample model, it can be seen that many hole-shaped voids are distributed. In order to confirm this tendency, FIG. 7 shows a plot of the values of the unevenness rate and the porosity calculated from each image.

According to FIG. 7, it can be confirmed that there is a positive correlation between the unevenness ratio in the powder bed and the porosity in the cross section of the sample model. In the figure, the result of linear approximation of the data points is also shown by a broken line, but a good approximation is established to some extent, and there is a gap ratio between the unevenness ratio in the powder bed and the porosity in the cross section of the sample model. , It can be said that there is a linear relationship. As described above, the fact that a linear relationship is observed between the unevenness of the powder material in the powder bed and the voids formed in the sample model obtained from the powder bed as a raw material is a major feature of the voids in the sample model. It is shown that the cause is the non-uniform distribution of the powder material in the powder bed as a raw material.

On the other hand, focusing on the case where a block-shaped sample is prepared using powder D, when the energy density of the laser beam used for modeling is low even though the uniformity of the distribution of the powder material on the powder bed is almost the same. , The porosity is large. This indicates that, at least when powder D is used as a raw material, the formation of voids in the produced sample model depends not only on the uneven distribution of the powder material on the powder bed but also on the conditions of laser irradiation. .. For example, the behavior of melting and recoagulation during laser irradiation due to the material physical properties of the alloys that make up the powder material depends on the energy density of the irradiated laser, and that dependence is also related to the formation of voids. Can be interpreted as being.

(Relationship between sample shape, porosity, and unevenness)
Looking at FIG. 6, except when the powder D is used as a raw material, the unevenness ratio of the powder bed and the porosity in the sample model tend to be larger in the beam-shaped sample than in the block-shaped sample. .. That is, when the beam-shaped sample is produced, the unevenness of the powder material in each constituent layer of the powder bed is larger than when the block-shaped sample is produced, and as a result, the porosity of the sample model obtained by laser irradiation is large. It's getting bigger. It is interpreted that this is because the beam-shaped sample is more susceptible to thermal deformation due to laser irradiation than the block-shaped sample. In laminated molding, when a powder material is spread over a sample model in the process of formation to form a new constituent layer, the sample model in the process of formation is subjected to thermal deformation, so that the sample is formed on the sample model. It is considered that the powder material is difficult to be uniformly distributed in the layer.

When powder D was used, the unevenness ratio of the powder bed and the porosity of the sample model were almost the same between the block-shaped sample and the beam-shaped sample. This means that for powder D, in the distribution of the powder material in the powder bed and the formation of voids in the sample model, the material physical characteristics (physical characteristics of the alloy material itself) and the powder characteristics (the shape and size of the powder) peculiar to the powder D are related. (Characteristics) means that it has a greater effect than the sample shape. The above discussion on the correlation between the unevenness ratio of the powder material and the porosity of the sample model also suggests the large contribution of the material properties.

The embodiments and examples of the present invention have been described above. The present invention is not particularly limited to these embodiments and examples, and various modifications can be made.

1 Laminated molding device 10 Powder bed container 11 Base material 20 Powder supply device 22 Powder supply path 23 Blade 30 Laser irradiation device 31 Laser unit 32 Galvano scanner 40 Camera 50 Control device B Powder bed B1 Constructible layer D1 Exposed part D2 Depressed part P Tertiary Original model

Claims (2)

A powder supply step of laying powder materials in layers to form a constituent layer of a powder bed, and a molding step of irradiating the surface of the constituent layers while scanning energy rays to solidify the powder material at the irradiated portion. This is a method of performing a photographing process, an evaluation process, and an information utilization process to produce a three-dimensional model by laminated modeling and to evaluate the powder bed .
A laminated molding device equipped with a camera that photographs the surface of the powder bed for the purpose of monitoring defects or abnormalities in the components of the device or the powder bed, or confirming that the molding is proceeding normally. Using,
In the photographing step, between the powder supply step and the modeling step, the surface of the constituent layer is photographed by the camera to acquire an image .
In the evaluation step, based on the said image obtained by each of the photographing process corresponding to the plurality of times of the powder supplying step, the three-dimensional distribution image showing the distribution of the powder material in the powder bed in a three-dimensional By creating a cross-sectional image obtained by cutting the three-dimensional distribution image and analyzing the distribution of the powder material in the cross-sectional image, the distribution of the powder material on the powder bed is uneven. Identify a certain unevenness and
The information utilization step is a step of utilizing the identification result of uneven distribution of the powder material in the powder bed obtained in the evaluation step.
Based on the large amount of unevenness in the distribution of the powder material in the powder bed identified in the evaluation step, it is determined whether or not the laminated molding should be stopped, and in the determination, the distribution of the powder material is uneven. When it turns out, the first step to stop the laminated molding and
A second step of analyzing the cause of uneven distribution of the powder material in the powder bed identified in the evaluation step, and a second step.
A third step of analyzing the cause of uneven distribution of the powder material in the powder bed identified in the evaluation step and improving the conditions for forming the powder bed based on the result of the analysis.
After associating the uneven distribution of the powder material in the powder bed obtained in the evaluation step with the voids in the three-dimensional model, the uneven distribution of the powder material in the powder bed is small. Despite the confirmation, when it is found that the void ratio indicating the ratio of voids is high in the three-dimensional model, the change of the irradiation conditions of the powder material to be used and the energy ray is examined . Perform at least one of the steps ,
When the first step is carried out as the information utilization step, the set of the powder supply step, the photographing step, the evaluation step, and the determination in the first step as the information utilization step is repeatedly carried out. When the determination in the first step reveals that the distribution of the powder material is uneven, the laminated molding is stopped.
When at least one of the three steps of the second step, the third step, and the fourth step is carried out as the information utilization step, the powder supply step, the photographing step, the evaluation step, and the modeling step are performed. A powder bed evaluation method, characterized in that at least one of the three steps as the information utilization step is carried out after the three-dimensional model is obtained by repeating the process a plurality of times. In the evaluation step, with respect to the image or the cross-sectional image obtained in the photographing step, information on the distribution of the powder material is extracted by a filtering process, and the distribution of the powder material is non-uniform by binarization. the identification, the at Ue Tsu line in this order, in the cross-sectional image, the powder bed evaluation method according to claim 1, wherein the identifying the uneven distribution of the powder material in the powder bed.

Images