Method for monitoring molding process

Abstract

To detect a presence or absence of a defect in a molded object by lamination molding with high accuracy.SOLUTION: A method according to at least one embodiment of the present disclosure for monitoring a molding process includes the steps of: molding a part of a molded object by irradiating a layer of raw material powder with an energy beam while scanning the energy beam to melt and solidify the raw material powder of the layer; acquiring time-series data of emission intensity of light emitted from the molded object when the energy beam is irradiated while scanning in the molding step; creating a two-dimensional map of the emission intensity based on the time-series data acquired in the acquiring step and information on a scanning trajectory of the energy beam; and detecting a presence or absence of a defect in the molded object based on the two-dimensional map created in the creating step.SELECTED DRAWING: Figure 2

Description

[Technical Field]

[0001] The present disclosure relates to a method for monitoring a build process. [Background technology]

[0002] Among the additive manufacturing methods for additively manufacturing three-dimensional objects, for example, the powder bed method uses an energy beam such as a light beam or an electron beam to irradiate metal powder, which is raw material powder laid out in layers, to repeatedly melt and solidify the powder and form a three-dimensional object (model).

[0003] In the area irradiated by the energy beam, the metal powder melts rapidly, which can cause defects in the molded object, such as the formation of cavities inside the molded object. Therefore, in order to ensure the quality of the molded object, it is necessary to perform non-destructive testing of the molded object after its creation is complete. However, non-destructive testing performed after the creation of the object takes time and is costly.

[0004] Therefore, it has also been proposed to detect defects in a layer of additively manufactured powder based on an image obtained by capturing an image of the surface of the additively manufactured powder before irradiation with an energy beam using an imaging means (see, for example, Patent Document 1). [Prior art documents] [Patent documents]

[0005] [Patent Document 1] Special Publication No. 2023-535930 Summary of the Invention [Problem to be solved by the invention]

[0006] However, in the technology described in the above-mentioned patent document, defects in the layer of additively manufactured powder are detected based on images captured by an imaging means, so it is necessary to predict defects in the layer of additively manufactured powder only from information on the surface of the laid additively manufactured powder. Therefore, the information obtained as the state within the layer of additively manufactured powder is only a prediction based on information on the surface of the laid additively manufactured powder. In addition, in the technology described in the above-mentioned patent document, defects in a layer of additively manufactured powder are detected based on images obtained by capturing images using an imaging means, so the resolution for defect detection is limited by the resolution of the imaging means.

[0007] In view of the above circumstances, at least one embodiment of the present disclosure aims to accurately detect the presence or absence of defects in an object manufactured by additive manufacturing. [Means for solving the problem]

[0008] A method for monitoring a build process according to at least one embodiment of the present disclosure includes: irradiating a layer of raw material powder with an energy beam while scanning the layer to melt and solidify the raw material powder in the layer, thereby forming a part of a shaped object; acquiring time-series data of light emission intensity of light emitted from the object when the energy beam is scanned and irradiated in the modeling step; creating a two-dimensional map of the emission intensity based on the time series data acquired in the acquiring step and information on the scanning trajectory of the energy beam; detecting the presence or absence of a defect in the object based on the two-dimensional map created in the creating step; Equipped with. [Effects of the Invention]

[0009] According to at least one embodiment of the present disclosure, the presence or absence of defects in an object manufactured by additive manufacturing can be detected with high accuracy. [Brief explanation of the drawings]

[0010] [Figure 1] 1 is a schematic diagram showing the overall configuration of a three-dimensional additive manufacturing apparatus, which is an additive manufacturing apparatus to which a method for monitoring a manufacturing process according to at least one embodiment of the present disclosure can be applied; [Figure 2] 10 is a flowchart showing a processing procedure in a method for monitoring a modeling process using a three-dimensional additive manufacturing device. [Figure 3] 10 is a diagram showing an example of time-series data of emission intensity stored in a storage unit of an emission monitor control unit. FIG. [Figure 4] FIG. 10 is a diagram showing an example of a two-dimensional map of emission intensity. [Figure 5] FIG. 10 shows a two-dimensional map after applying a spatial filter. [Figure 6] 10 is a graph showing the number of peculiar regions in a two-dimensional map for each modeling layer after applying a spatial filter to a certain modeled object. [Figure 7] 7 is a graph showing the defect rate for each modeling layer in the model shown in FIG. 6. [Figure 8] 10 is a graph showing an example of the emission intensity of light incident on a light detection unit. DETAILED DESCRIPTION OF THE INVENTION

[0011] Hereinafter, several embodiments of the present disclosure will be described with reference to the accompanying drawings. However, the dimensions, materials, shapes, relative arrangements, etc. of components described as embodiments or shown in the drawings are merely illustrative examples and are not intended to limit the scope of the present disclosure. For example, expressions expressing relative or absolute arrangement such as "in a certain direction," "along a certain direction," "parallel," "orthogonal," "center," "concentric," or "coaxial" not only express such an arrangement exactly, but also express a state in which there is a relative displacement with a tolerance or an angle or distance to the extent that the same function is obtained. For example, expressions such as "identical," "equal," and "homogeneous" that indicate that something is in an equal state not only indicate a state of strict equality, but also indicate a state in which there is a tolerance or a difference to the extent that the same function is obtained. For example, expressions representing shapes such as a square shape or a cylindrical shape not only represent shapes such as a square shape or a cylindrical shape in the strict geometric sense, but also represent shapes including uneven portions, chamfered portions, etc., to the extent that the same effect can be obtained. On the other hand, the expressions "comprise," "include," "have," "includes," or "have" of one element are not exclusive expressions that exclude the presence of other elements.

[0012] (About the 3D additive manufacturing device 1) FIG. 1 is a schematic diagram showing the overall configuration of a three-dimensional additive manufacturing apparatus 1, which is an additive manufacturing apparatus to which a method for monitoring a manufacturing process according to at least one embodiment of the present disclosure can be applied. The three-dimensional additive manufacturing device (additive manufacturing device) 1 is a device for producing a three-dimensional object 15 by performing additive manufacturing by irradiating a light beam 71 as an energy beam onto metal powder, which is raw material powder laid in layers, and is capable of performing additive manufacturing using the powder bed method. The three-dimensional additive manufacturing apparatus 1 shown in FIG. 1 can form parts such as rotor blades and stator blades of turbines such as gas turbines and steam turbines, or inner cylinders, transition pieces and nozzles of combustors.

[0013] The three-dimensional additive manufacturing apparatus 1 shown in FIG. 1 includes a storage unit 31 for raw material powder 30. The three-dimensional additive manufacturing apparatus 1 shown in FIG. 1 also includes a powder bed forming unit 5 having a base plate 2 on which a powder bed 8 is formed by sequentially stacking layers 8a of raw material powder 30 supplied from the storage unit 31. The three-dimensional additive manufacturing apparatus 1 shown in FIG. 1 also includes a light beam irradiation device 20 including a light beam irradiation unit 21 capable of irradiating the powder bed 8 with a light beam 71, and a light beam control unit 22 that controls the light beam irradiation unit 21. The three-dimensional additive manufacturing apparatus 1 shown in FIG. 1 also includes a powder laying unit 10, which will be described later, a drive cylinder 2a of the base plate 2, and a control unit 80 that controls the entire three-dimensional additive manufacturing apparatus 1. The three-dimensional additive manufacturing apparatus 1 shown in FIG.

[0014] The base plate 2 serves as a foundation on which the object 15 is formed. The base plate 2 is disposed inside a generally cylindrical cylinder 4 having a vertical central axis so that it can be raised and lowered by a drive cylinder 2a. During the forming operation, a new layer 8a is formed on the powder bed 8 formed on the base plate 2 by laying raw material powder 30 on the upper layer side each time the base plate 2 is lowered in each cycle.

[0015] The 3D additive manufacturing apparatus 1 shown in FIG. 1 includes a powder laying unit 10 for laying raw material powder 30 on a base plate 2 to form a layer 8a from the raw material powder 30. The powder laying unit 10 supplies raw material powder 30 from a storage unit 31 to the upper surface of the base plate 2 and flattens the surface, thereby forming a layer 8a having a substantially uniform thickness across the entire upper surface of the base plate 2. The powder bed 8, in which the layers 8a formed in each cycle are sequentially stacked, is selectively melted and solidified by irradiation with a light beam 71 from a light beam irradiation unit 21. In the next cycle, raw material powder 30 is again laid on the upper layer by the powder laying unit 10, forming a new layer 8a, and the layers are stacked in layers.

[0016] The raw powder 30 supplied from the powder depositing section 10 is a powdered substance that will be used as the raw material for the molded object 15, and a wide range of metal materials, such as iron, copper, aluminum, or titanium, and non-metallic materials, such as ceramics, can be used.

[0017] In the three-dimensional additive manufacturing apparatus 1 shown in Fig. 1, the powder bed formation unit 5 and the powder laying unit 10 are housed in a chamber 9. The chamber 9 is provided with a window 9a that is capable of transmitting a light beam 71. Protective glass or the like is arranged in the window 9a so that the light beam 71 or a light beam can pass through while maintaining airtightness between the inside and outside of the chamber 9.

[0018] The control unit 80 is a control unit of the 3D additive manufacturing apparatus 1 shown in Fig. 1 and is configured by, for example, an electronic processing device such as a computer. Information related to the scanning position of the light beam 71 for each layer 8a is input to the control unit 80 as information necessary for manufacturing the model 15. The information related to the irradiation position of the light beam 71 for each layer 8a may be input from, for example, an external device and stored in, for example, a memory unit (not shown) of the control unit 80.

[0019] The control unit 80 is configured to control the driving of the drive cylinder 2a of the powder bed forming unit 5. The control unit 80 is also configured to be able to control the detection device 40 in cooperation with the control unit 80.

[0020] (Light beam irradiation device 20) The light beam irradiation device 20 is a device for irradiating the powder bed 8 with a light beam 71, and has a light beam irradiation unit 21 capable of irradiating the light beam 71 as described above, a light beam control unit 22 for controlling the light beam irradiation unit 21, and a galvanometer mirror 23 for scanning the light beam 71. The light beam control unit 22 is configured to include an electronic processing device such as a computer and a memory unit for storing time series data of the scanning trajectory of the light beam 71, and is configured to control the irradiation position of the light beam 71 irradiated by the light beam irradiation unit 21 by controlling the attitude of the galvanometer mirror 23, which can swing in any direction, based on information regarding the irradiation position of the light beam 71 for each layer 8a received from the control unit 80.

[0021] (Detection device 40) The detection device 40 is a device for detecting the presence or absence of defects in the molded object 15 based on the light emitted from the molded object 15 when irradiated with the light beam 71, i.e., the light emission intensity of the light emitted from the molten pool and plume formed when irradiated with the light beam 71, and is, for example, a plasma light emission monitor capable of detecting plasma light emission during the process in real time. The detection device 40 has a light detection unit 41 for detecting the light emitted from the molded object 15, and an emission monitor control unit 42 that can store the light emission intensity of the light detected by the light detection unit 41 as time-series data. The light emission monitor control unit 42 includes an electronic processing unit such as a computer, and a storage unit for storing time-series data of the light emission intensity detected by the light detection unit 41. In the three-dimensional additive manufacturing apparatus 1 shown in FIG. 1, the light emitted from the model 15 when irradiated with the light beam 71 is reflected by the beam splitter 6 and enters the light detection unit 41.

[0022] (Challenges in conventional additive manufacturing processes) In the region irradiated with the energy beam, the raw material powder 30 melts rapidly, which may cause defects in the shaped object 15, such as the formation of cavities inside the shaped object 15. Therefore, in order to ensure the quality of the object 15, it is necessary to perform a non-destructive inspection of the object 15 after the formation of the object 15 is completed. However, not all defects can be detected by non-destructive testing that is performed after the fabrication of the model 15 is completed. In addition, non-destructive testing takes time and is costly.

[0023] As a result of extensive research, the inventors have found that the intensity of light emitted from the object 15 when irradiated with the light beam 71 changes depending on the molten state of the molten pool, and that the molten state of the molten pool changes depending on the state of the layer of raw material powder 30 laid on the powder bed 8 and the state of the object 15a below the layer of raw material powder 30. In other words, the inventors have found that it is possible to detect the presence or absence of defects in the object 15 based on the intensity of light emitted from the object 15 when irradiated with the light beam 71.

[0024] For example, if a cavity exists in the forming layer 15a below the layer of raw material powder 30, when the light beam 71 passes over the cavity, the molten state changes as the molten pool formed by the irradiation of the light beam 71 falls into the cavity, causing a decrease in the intensity of the light emission signal. Therefore, the presence of a cavity can be detected by detecting this decrease in the intensity of the light emission signal.

[0025] Furthermore, for example, if the thickness of the layer of raw material powder 30 is thin or if foreign matter such as sputtering is deposited locally, the intensity of the light emission signal increases locally. Therefore, by detecting this increase in the intensity of the light emission signal, defects in the layer of raw material powder 30 can be identified.

[0026] Note that, when only looking at the one-dimensional time-series data of the emission intensity shown in Fig. 8 (described later), it is relatively difficult to identify the locations where the emission intensity drops and determine whether or not cavities exist in the substrate. On the other hand, in additive manufacturing, the object 15 is formed by the light beam 71 scanning back and forth at narrow intervals, so by displaying the data in a two-dimensional map such as those shown in Fig. 4 and Fig. 5 (described later) based on the one-dimensional time-series data, the light beam 71 passes multiple times over cavities having sizes in the X and Y directions (i.e., the scanning direction of the light beam 71 and the direction perpendicular to the scanning direction), resulting in a significant drop in the emission signal, making it easier to detect cavity defects. By using a two-dimensional map, it becomes easy to detect defects even when the intensity of the light emission signal increases, such as when foreign matter such as sputters is locally deposited.

[0027] Therefore, in the method for monitoring a modeling process according to some embodiments, the presence or absence of a defect in the model 15 during modeling of the model 15 is detected as follows.

[0028] (How to monitor the build process) 2 is a flowchart showing the procedure of a method for monitoring a modeling process using the above-described three-dimensional additive manufacturing apparatus 1. The following will be described with reference to the flowchart in FIG. A method for monitoring a manufacturing process in some embodiments includes step S10 of irradiating an energy beam to manufacture a product, step S20 of acquiring time series data of the luminescence intensity, step S30 of creating a two-dimensional map, and step S40 of detecting the presence or absence of defects.

[0029] (Step S10: Irradiating an energy beam to form a shape) Step S10 of irradiating with an energy beam to form a shape is a step of irradiating layer 8a of raw material powder 30 with a light beam 71 as an energy beam while scanning the layer 8a of raw material powder 30, thereby melting and solidifying the raw material powder 30 in layer 8a, thereby forming a part of the object 15. In the following description, step S10 of irradiating with an energy beam to form a shape is also simply referred to as step S10. Prior to carrying out step S10, the control unit 80, more specifically, the arithmetic unit of the control unit 80, controls the powder laying unit 10 to lay down the raw material powder 30. As a result, a layer 8a is formed from the raw material powder 30 supplied from the storage unit 31. Next, in step S10, the control unit 80 controls the light beam irradiation device 20 to irradiate while scanning the light beam 71. As a result, the light beam irradiation device 20 irradiates the light beam 71 based on information regarding the irradiation position of the light beam 71 for each layer 8a received from the control unit 80. In step S10, one modeling layer 15a is modeled.

[0030] In step S10, the light beam control unit 22 of the light beam irradiation device 20 stores time-series data of the scanning trajectory of the light beam 71 in the storage unit of the light beam control unit 22 when the light beam 71 is irradiated.

[0031] (Step S20 of acquiring time series data of luminescence intensity) Step S20 of acquiring time-series data of the light emission intensity is a step of acquiring time-series data of the light emission intensity of light emitted from the model 15 when the light beam 71 is scanned and irradiated in step S10. In the following description, step S20 of acquiring time-series data of the light emission intensity is also simply referred to as step S20. Step S20 is performed at the same timing as step S10. In step S20, the light emission monitor control unit of the detection device 40 acquires the light emission intensity of the light incident on the light detection unit 41 as time-series data, and stores it in the storage unit of the light emission monitor control unit .

[0032] (Step S30: Creating a 2D map) Step S30 of creating a two-dimensional map is a step of creating a two-dimensional map of emission intensity based on the time series data of the scanning trajectory of light beam 71, which is information related to the scanning trajectory of light beam 71 acquired in step S10, and the time series data of emission intensity acquired in step S20. In the following description, step S30 of creating a two-dimensional map will also be simply referred to as step S30. In step S30, for example, the control unit 80 reads the time series data of the scanning trajectory of the light beam 71 stored in the memory unit of the light beam control unit 22, and creates a two-dimensional map of the light emission intensity based on the time series data of the light emission intensity stored in the memory unit of the light emission monitor control unit 42.

[0033] FIG. 3 is a diagram showing an example of time-series data of light emission intensity stored in the memory unit of the light emission monitor control unit 42. The vertical axis of the graph shown in FIG. 3 represents light emission intensity, and the horizontal axis represents irradiation time. For example, when forming one modeling layer 15a, the light beam 71 is scanned on the surface of the layer 8a multiple times while shifting the scanning position for each scan. The light emission intensity obtained by one scan of the light beam 71 is represented by a single graph line 51 in the graph of FIG. 3. In FIG. 3, multiple graph lines 51 of light emission intensity obtained by scanning the surface of the layer 8a multiple times with the light beam 71 are superimposed. Note that FIG. 8, which will be described later, shows an example of a graph line of light emission intensity obtained by one scan of the light beam 71.

[0034] Fig. 4 is a diagram showing an example of a two-dimensional map of emission intensity. For convenience of illustration, the two-dimensional map 53 shown in Fig. 4 is a grayscale representation of a color contour diagram of emission intensity. In the color contour diagram before grayscaling, areas with weak emission intensity are shown in blue, and areas with strong emission intensity are shown in red, with the color gradually changing from blue to green, yellow, and red as the emission intensity increases. When such a color-displayed contour diagram is converted to grayscale, the areas with weak luminescence intensity, shown in blue, and the areas with strong luminescence intensity, shown in red, become darker, while the areas with intermediate luminescence intensity, shown in green and yellow, which are neither weak nor strong, become lighter. Therefore, the greater the deviation of the emission intensity from the average value of the emission intensity in the two-dimensional map 53, the darker the density will be in FIG. The two-dimensional maps 53 and 63 shown in FIG. 4 and FIG. 5 described later are maps represented on an XY plane in the XY direction, i.e., the scanning direction of the light beam 71 (e.g., the X direction) and a direction perpendicular to the scanning direction (e.g., the Y direction), and represent the distribution of luminescence intensity on the XY plane in the scanning direction of the light beam 71 (e.g., the X direction) and a direction perpendicular to the scanning direction (e.g., the Y direction) on the modeling surface, i.e., the surface of the layer 8a irradiated with the light beam 71.

[0035] (Step S40: Detecting the presence or absence of defects) Step S40 of detecting the presence or absence of defects is a step of detecting the presence or absence of defects in the object 15 based on the two-dimensional map created in step S30. In the following description, step S40 of detecting the presence or absence of defects will also be simply referred to as step S40. In step S40, the worker or the control unit 80 detects whether or not there is a defect in the shaped object 15 based on the two-dimensional map 53 created by the control unit 80 in step S30.

[0036] When detecting the presence or absence of defects in the object 15, the worker observes a two-dimensional map 53, which is, for example, a color contour diagram as described above, displayed on a display device (not shown) or the like. The worker may then detect the presence or absence of defects by estimating that there are defects in areas where a region with weak emission intensity is recognized as a certain size, or in areas where a region with strong emission intensity is recognized as a certain size. Specifically, the worker should determine whether there is a peculiar area that is a continuous area in which the deviation of the light emission intensity from the average value in the two-dimensional map 53 exceeds a first threshold value Th1 that is larger than the average value, or has a light emission intensity that is smaller than a second threshold value Th2 that is smaller than the average value, and that is recognized as an area of ​​a size equal to or larger than the specified range R. Here, the above-mentioned specified range R is, for example, 100 μm in terms of circle equivalent diameter. If the material and particle size distribution of the raw material powder 30 and the irradiation conditions such as the intensity and scanning speed of the light beam 71 are the same, the average values ​​of the emission intensities in the two-dimensional map 53 will also be approximately the same. Therefore, it is possible to set the first threshold value and the second threshold value in advance based on this average value.

[0037] 4, peculiar regions 55 are observed in four locations from the upper left to the lower center of the left and right sides of the two-dimensional map 53, and in one location near the right end of the two-dimensional map 53. In the actual object 15, a defect with a diameter of 100 μm or more exists in a circled region 56. In this way, there is a high possibility that a defect exists at a position where a peculiar region 55 is recognized in the two-dimensional map 53. Therefore, if a spatial filter such as that described below is not applied to the two-dimensional map 53, the worker may estimate that a defect exists in the object 15 at a position corresponding to the above-mentioned peculiar region 55.

[0038] In the object 15, there are no defects in the positions corresponding to the upper left peculiar region 55 and the lower center peculiar region 55 in the two-dimensional map 53 in Fig. 4. This is thought to be due to the following reason. For example, even if a defect such as a cavity occurs during the formation of a certain modeling layer 15a, the cavity may disappear if the layer 8a of raw material powder 30 laid on the modeling layer 15a is irradiated with the light beam 71 to melt and solidify it. In other words, a defect in a modeling layer 15a below the subsequent layer may be repaired during the formation of the subsequent layer. Therefore, even if a peculiar region 55 is recognized in the 2D map 53, it does not necessarily mean that a defect remains in the modeled object 15 at a position corresponding to the peculiar region 55.

[0039] When the control unit 80 detects the presence or absence of defects in the object 15, the control unit 80 may determine that a region is a peculiar region 55 if the region is recognized as a continuous region having an emission intensity whose deviation from the average value of the emission intensity in the two-dimensional map 53 exceeds a first threshold value Th1 that is larger than the average value, or has an emission intensity that is smaller than a second threshold value Th2 that is smaller than the average value, and the region is larger than the specified range R. The control unit 80 may then determine that a defect exists in the object 15 at a position corresponding to the unique region 55 .

[0040] Step S40 may be performed each time a two-dimensional map 53, 63 is obtained in step S30, i.e., each time one modeling layer 15a is modeled, or may be performed after two-dimensional maps 53, 63 are obtained for multiple modeling layers 15a.

[0041] As described above, the method for monitoring a manufacturing process in some embodiments includes step S10 of irradiating an energy beam to manufacture a product, step S20 of acquiring time series data of the luminescence intensity, step S30 of creating a two-dimensional map, and step S40 of detecting the presence or absence of defects. This allows the presence or absence of defects in the object 15 to be detected based on two-dimensional maps 53, 63 that are based on the light emission intensity of the light emitted from the object when the light beam 71 serving as an energy beam is scanned and irradiated.Therefore, since two-dimensional maps 53, 63 with higher resolution than images obtained by, for example, capturing images with an imaging device can be obtained relatively easily, it becomes easier to accurately detect the presence or absence of defects in the object 15 created by additive manufacturing. In particular, since imaging devices use lenses, image distortion occurs due to lens aberration, such as distortion of the peripheral parts of the image capture surface, making it impossible to uniformly evaluate the modeling surface and tending to result in poor accuracy in the peripheral parts.However, according to some of the above-mentioned embodiments, even the peripheral parts of the modeling surface can be evaluated uniformly without distortion. Furthermore, according to the method for monitoring a building process in some embodiments, the light emission intensity of the light emitted from the object 15 when the light beam 71 is irradiated while being scanned is used, and therefore it is possible to obtain information that reflects the melted state of the molten pool at the time of irradiation with the light beam 71. As a result, if there is a defect located inside the object 15 relative to the currently being formed layer 15a that may affect the melted state of the molten pool, for example, a change occurs in the light emission intensity of the light emitted from the object 15 when the light beam 71 is irradiated while being scanned. Therefore, according to the method for monitoring a building process in some embodiments, it is possible to detect the presence or absence of a defect located inside the object 15 relative to the currently being formed layer 15a.

[0042] (Applying spatial filters to 2D maps) Even in a region where the presence of a peculiar region 55 is not recognized in FIG. 4, such as the region 56 in the upper right corner of FIG. 4, a defect actually exists in the shaped object 15. Even in such a case, by applying a spatial filter to the two-dimensional map 53, it is possible to highlight the areas where the emission intensity deviates from the average value. Fig. 5 is a diagram showing a two-dimensional map after a spatial filter has been applied to the two-dimensional map. The two-dimensional map shown in Fig. 5 is two-dimensional map 63 after a spatial filter has been applied to two-dimensional map 53 shown in Fig. 4. The passband of the spatial filter to be applied is determined in advance by experiment or the like so that regions where the emission intensity changes can be efficiently extracted.

[0043] When applying a spatial filter, the spatial filter is applied to a two-dimensional map 53 as shown in Fig. 4 in step S30. That is, in step S30, for example, the control unit 80 reads time-series data of the scanning trajectory of the light beam 71 stored in the storage unit of the light beam control unit 22, and creates a two-dimensional map of the emission intensity based on time-series data of the emission intensity stored in the storage unit of the emission monitor control unit 42. Then, the control unit 80 applies a spatial filter with a passband set in advance to the created two-dimensional map.

[0044] As shown in FIG. 5, in the two-dimensional map 63 after the application of the spatial filter, peculiar regions 65, i.e., continuous regions having luminescence intensities where the deviation from the average value of the luminescence intensity in the two-dimensional map 63 exceeds a first threshold value Th1 that is larger than the average value, or falls below a second threshold value Th2 that is smaller than the average value, and which are recognized as regions of a certain size, stand out compared to the two-dimensional map 53 shown in FIG. 4 before the application of the spatial filter. In the example shown in FIG. 5, a unique region 65 is observed within a circled region 64.

[0045] After the spatial filter is applied, in step S40, the worker or the control unit 80 detects whether or not there is a defect in the shaped object 15 based on the two-dimensional map 63 after the spatial filter is applied.

[0046] When a worker detects whether or not there is a defect in the object 15, the worker may estimate that the defect exists at a position corresponding to the peculiar region 65 in the two-dimensional map 63 after the spatial filter is applied. Regarding the presence or absence of a peculiar region 65, the worker should determine whether or not there is a peculiar region 65 that is recognized as a region having a size equal to or larger than the specified range R, which is a continuous region in which the deviation of the luminescence intensity from the average value in the two-dimensional map 63 exceeds a first threshold value Th1 that is larger than the average value, or has an luminescence intensity that is smaller than a second threshold value Th2 that is smaller than the average value.

[0047] In the object 15, defects exist only at positions corresponding to the peculiar regions 65 in the region 66 marked with a reference numeral 66 within the circled region 64 in the two-dimensional map 63 in Fig. 5, and no defects exist at positions corresponding to the peculiar regions 65 in the other regions 64 not marked with a reference numeral 66. This is thought to be because the defects were repaired during the formation of the subsequent layers, as described above.

[0048] When the control unit 80 detects the presence or absence of defects in the object 15, in step S40, if the control unit 80 recognizes a continuous region having an emission intensity whose deviation from the average value of the emission intensity in the two-dimensional map 63 exceeds a first threshold value Th1 that is larger than the average value, or which is smaller than a second threshold value Th2 that is smaller than the average value, and the region is larger than the specified range R, it may determine the region to be a peculiar region 65. The control unit 80 may then determine that a defect exists in the object 15 at a position corresponding to the unique region 65 .

[0049] In the method for monitoring a modeling process according to some embodiments, in step S30, a spatial filter may be applied to the two-dimensional map 53 to create a spatially filtered two-dimensional map 63. In step S40, the presence or absence of a defect in the model 15 may be detected based on the spatially filtered two-dimensional map 63 created in step S30. Even if a defect exists in the object 15, a clear change in the light emission signal may not be observed in the two-dimensional map 53 obtained during the formation of the formation layer 15a in which the defect exists, before the spatial filter is applied. According to some embodiments of the method for monitoring a modeling process, a spatial filter can be applied to the two-dimensional map 53 to highlight changes in the light emission signal resulting from defects, thereby improving the accuracy of detecting whether or not there is a defect in the model 15.

[0050] In some embodiments of the method for monitoring a modeling process, in step S40, the presence or absence of peculiar regions 55, 65 having a size equal to or greater than a specified range R, in which the deviation of the emission intensity from the average value in the two-dimensional maps 53, 63 exceeds a first threshold value Th1 that is greater than the average value, or the emission intensity falls below a second threshold value Th2 that is smaller than the average value, is detected, and the presence or absence of defects in the model 15 is detected based on the presence or absence of the peculiar regions 55, 65. This allows the presence or absence of defects in the object 15 to be detected based on the presence or absence of the peculiar regions 55, 65, which is a relatively easy method of determination, and therefore makes it easy to detect the presence or absence of defects in the object 15.

[0051] In addition, the first threshold Th1, second threshold Th2, and specified range R for detecting the presence or absence of a peculiar region 55 in the two-dimensional map 53 before applying the spatial filter may be the same as or different from the first threshold Th1, second threshold Th2, and specified range R for detecting the presence or absence of a peculiar region 65 in the two-dimensional map 63 after applying the spatial filter.

[0052] (Relationship between the number of peculiar areas 65 and the defect rate in the model 15) Fig. 6 is a graph showing the number of unique regions 65 for each modeling layer 15a in a two-dimensional map 63 after applying a spatial filter for a certain modeling object 15. The vertical axis of the graph in Fig. 6 represents the number of unique regions 65 in the two-dimensional map 63. The horizontal axis of the graph in Fig. 6 represents the position in the stacking direction of each modeling layer 15a, and is expressed as the number of modeling layers 15a counted from the bottom in the vertical direction during additive manufacturing. Fig. 7 is a graph showing the defect rate for each modeling layer 15a in the model 15 according to the graph in Fig. 6. The vertical axis of the graph in Fig. 7 represents the defect rate of the modeling layer 15a. Note that the defect rate of the modeling layer 15a in the graph in Fig. 7 is the value obtained by dividing the area of ​​defects when the modeling layer 15a is viewed from the stacking direction of the modeling layer 15a by the area of ​​the modeling layer 15a. The horizontal axis of the graph in FIG. 7 represents the position of each modeling layer 15a in the stacking direction, and is expressed as the number of modeling layers 15a counted in order from the bottom in the vertical direction during additive manufacturing.

[0053] In the modeling layer 15a where the number of peculiar regions 65 is relatively small in Fig. 6, the defect rate is extremely low, as shown in Fig. 7. However, in the modeling layer 15a where the number of peculiar regions 65 is significantly large, as shown by the dashed line in Fig. 6, the defect rate is significantly high, as shown by the dashed line in Fig. 7. Thus, a strong correlation is observed between the number of peculiar regions 65 in the modeling layer 15a and the defect rate.

[0054] Therefore, in step S40, the worker or the control unit 80 may determine that a defect exists in the forming layer 15a corresponding to the two-dimensional map 63 in which the number of peculiar regions 65 exceeds a threshold value, or the two-dimensional map 53 in which the number of peculiar regions 55 exceeds a threshold value. This allows the presence or absence of defects to be detected relatively quickly.

[0055] (Use of two-dimensional maps 53 and 63 for multiple modeling layers 15a) As described above, defects in the forming layer 15a below the subsequent layer may be repaired during the formation of the subsequent layer, so even if peculiar areas 55, 65 are recognized in the two-dimensional maps 53, 63, there may be no defects at the positions corresponding to the peculiar areas 55, 65. Therefore, in step S40, the worker or the control unit 80 may extract peculiar areas 55, 65 from the two-dimensional maps 53, 63 for two or more successively formed forming layers 15a, and if peculiar areas 55, 65 are found at the same positions on the two-dimensional maps 53, 63 for at least two or more forming layers 15a, it may be determined that a defect exists at the position corresponding to the peculiar areas 55, 65.

[0056] That is, in the method for monitoring a modeling process according to some embodiments, in step S10, a layer 8a of raw material powder 30 is laid, and a light beam 71 is scanned and irradiated onto the laid layer 8a to form one modeling layer 15a. Step S10 may be repeatedly performed during the modeling of the object 15. In step S20, the time-series data may be acquired each time step S10 is performed. In step S30, two-dimensional maps 53, 63 may be created for each modeling layer 15a. In step S40, the presence or absence of defects in the modeled object 15 may be detected based on the presence or absence of peculiar regions 55, 65 in a first map, which is the two-dimensional maps 53, 63 of the first modeling layer, and a second map, which is the two-dimensional maps 53, 63 of a second modeling layer adjacent to the first modeling layer on at least one side or the other of the first modeling layer in the stacking direction of the modeling layer 15a. If the first modeling layer is the n-th modeling layer 15a of the modeled object 15, the second modeling layer is the (n-1)th or (n+1)th modeling layer 15a of the modeled object 15.

[0057] This allows the presence or absence of defects in the object 15 to be detected based on the presence or absence of peculiar regions 55, 65 in the two-dimensional maps 53, 63 across the multiple object layers 15a, thereby improving the accuracy of detecting the presence or absence of defects in the object 15.

[0058] (Detecting defects in a model 15 using a trained model) As described above, defects in the forming layer 15a below the subsequent layer may be repaired during the formation of the subsequent layer, so even if peculiar areas 55, 65 are recognized in the two-dimensional maps 53, 63, there may be no defects at the positions corresponding to the peculiar areas 55, 65. Therefore, for example, the control unit 80 may be configured to include a trained model 82 stored in a storage unit 81 of the control unit 80.

[0059] The trained model 82 is created by machine learning training data in which the plurality of two-dimensional maps 53, 63 are associated with information on the positions and sizes of defects present in the shaped object 15. As a result, in the trained model 82, when, for example, peculiar areas 55, 65 are recognized in the two-dimensional maps 53, 63, it becomes possible to determine with relatively high accuracy whether or not a defect exists at the position corresponding to the peculiar area 55, 65.

[0060] In step S40, the control unit 80 inputs the two-dimensional maps 53 and 63 to the trained model 82. As a result, the trained model 82 outputs information regarding the presence or absence of a defect in the object 15, and, if a defect is estimated to exist, information regarding the position and size of the defect.

[0061] In this way, in some embodiments of the method for monitoring a modeling process, the presence or absence of defects in the model 15 may be detected using a trained model 82 configured to detect the presence or absence of defects based on the two-dimensional maps 53, 63. This improves the accuracy of detecting whether or not there is a defect in the shaped object 15.

[0062] Furthermore, for example, the trained model 82 may be configured to determine that no defect corresponding to the peculiar region 55, 65 exists even when the peculiar region 55, 65 has a size equal to or larger than the specified range R, in which the deviation of the emission intensity from the average value in the two-dimensional maps 53, 63 exceeds a first threshold value Th1 that is larger than the average value, or is below a second threshold value Th2 that is smaller than the average value, if it is determined that no defect corresponding to the peculiar region 55, 65 exists. Then, in step S40, the trained model 82 configured in this manner may be used to detect the presence or absence of a defect in the object 15. This improves the accuracy of detecting whether or not there is a defect in the shaped object 15.

[0063] (Predicting defects in objects using trained models) For example, by inputting two-dimensional maps 53, 63 into the above-mentioned trained model 82, the trained model 82 may be made to predict that defects will occur in the subsequent layer of the forming layer 15a related to the input two-dimensional maps 53, 63. In this case, for example, the trained model 82 is created by machine learning training data in which the plurality of two-dimensional maps 53, 63 are associated with information on the positions and sizes of defects in the subsequent layers. This makes it possible for the trained model 82 to detect cases in which abnormal surface shapes, etc., that accumulate with each layer may induce an abnormality in the light emission signal (process abnormality) in the subsequent layer.

[0064] Thus, in some embodiments of the method for monitoring a modeling process, in step S40, the presence or absence of defects in the modeled object 15 may be detected using a trained model 82 configured to determine whether or not defects will occur in the modeling layer 15a (subsequent layer) that will be modeled after the modeling layer 15a currently being modeled in step S10. This makes it possible to predict the occurrence of defects in subsequent layers before defects actually occur in the object 15 during modeling, and therefore makes it possible to change modeling conditions, such as the irradiation conditions of the light beam 71, so as to prevent defects from occurring in the modeling layer 15a where defects are likely to occur and at the predicted defect occurrence position in the modeling layer 15a, or to take measures to repair the defects. This makes it possible to efficiently reduce defects remaining in the object 15.

[0065] (Real-time prediction of defects in objects using trained models) For example, by inputting information on the emission intensity of light incident on the light detection unit 41 in real time during the execution of step S10 into the trained model 82 described above, it may be possible to predict that a defect is likely to occur immediately. Then, the control unit 80 may be configured to, for example, notify the user that a defect is likely to occur immediately.

[0066] FIG. 8 is a graph showing an example of the emission intensity of light incident on the light detection unit 41, and shows an example in which a defect occurs at a position irradiated with the light beam 71 at irradiation time T1 in FIG. In the example shown in FIG. 8, it is desirable to predict, before the irradiation time T1 is reached, that there is a high possibility that a defect will occur at the position irradiated with the light beam 71 at the irradiation time T1. To enable such prediction, for example, the trained model 82 may be created by machine learning training data that associates information on the light emission intensity with the irradiation time at which the defect occurred. This enables the trained model 82 to predict the high probability of a defect occurring immediately based on the light emission intensity information input in real time.

[0067] In addition, the above-mentioned detection of the presence or absence of defects in the object 15 using the trained model 82, prediction of the occurrence of defects in the object 15 using the trained model 82, and real-time prediction of the occurrence of defects in the object 15 using the trained model 82 may be performed in any appropriate combination.

[0068] The present disclosure is not limited to the above-described embodiments, but also includes modifications to the above-described embodiments and appropriate combinations of these modifications.

[0069] The contents described in each of the above embodiments can be understood, for example, as follows. (1) At least one embodiment of the present disclosure includes a step S10 of forming a portion of a molded object 15 by scanning and irradiating a layer 8a of the raw material powder 30 with an energy beam (light beam 71) to melt and solidify the raw material powder 30 in the layer 8a, a step S20 of acquiring time series data of the luminescence intensity of light emitted from the molded object 15 when the energy beam (light beam 71) is scanned and irradiated in the molding step S10, a step S30 of creating two-dimensional maps 53, 63 of the luminescence intensity based on the time series data acquired in the acquiring step S20 and information related to the scanning trajectory of the energy beam (light beam 71), and a step S40 of detecting the presence or absence of defects in the molded object 15 based on the two-dimensional maps 53, 63 created in the creating step S30.

[0070] According to the method (1) above, the presence or absence of defects in the object 15 is detected based on two-dimensional maps 53, 63 based on the light emission intensity of the light emitted from the object 15 when the energy beam (light beam 71) is scanned and irradiated. Therefore, it is relatively easy to obtain two-dimensional maps 53, 63 with higher resolution than images obtained by, for example, capturing images with an imaging device, making it easy to accurately detect the presence or absence of defects in the object 15 created by additive manufacturing. Furthermore, according to the method (1) above, the light emission intensity of the light emitted from the object 15 when the energy beam (light beam 71) is irradiated while being scanned is used, so that information reflecting the molten state of the molten pool at the time of irradiation with the energy beam (light beam 71) can be obtained. As a result, if there is a defect located inside the object 15 relative to the currently being formed layer 15a that may affect the molten state of the molten pool, for example, a change occurs in the light emission intensity of the light emitted from the object 15 when the energy beam (light beam 71) is irradiated while being scanned. Therefore, according to the method (1) above, it is possible to detect the presence or absence of a defect located inside the object 15 relative to the currently being formed layer 15a.

[0071] (2) In some embodiments, in the method of (1) above, in the creating step S30, a spatial filter may be applied to the two-dimensional map 53 to create a spatially filtered two-dimensional map 63. In the detecting step S40, the presence or absence of a defect in the object 15 may be detected based on the spatially filtered two-dimensional map 63 created in the creating step S30.

[0072] Even if a defect exists in the object 15, a clear change in the light emission signal may not be observed in the two-dimensional map 53 obtained during the formation of the formation layer 15a in which the defect exists, before the spatial filter is applied. According to the method (2) above, by applying a spatial filter to the two-dimensional map 53, it is possible to highlight changes in the light emission signal resulting from defects, thereby improving the accuracy of detecting whether or not the object 15 has defects.

[0073] (3) In some embodiments, in the method of (1) or (2) above, in the detection step S40, the presence or absence of peculiar regions 55, 65 having a size equal to or greater than a specified range R, in which the deviation of the emission intensity from the average value in the two-dimensional maps 53, 63 exceeds a first threshold value Th1 that is greater than the average value, or the emission intensity falls below a second threshold value Th2 that is smaller than the average value, is detected, and the presence or absence of defects in the object 15 is detected based on the presence or absence of the peculiar regions 55, 65.

[0074] As a result of careful investigation, the inventors have found that defects exist at a relatively high rate in the positions of the shaped object 15 corresponding to the above-mentioned unique regions 55 and 65 . According to the method (3) above, the presence or absence of defects in the object 15 can be detected based on the presence or absence of the unique regions 55, 65, which is a relatively easy method of determination, and therefore the presence or absence of defects in the object 15 can be easily detected.

[0075] (4) In some embodiments, in the method of (3) above, in the modeling step S10, a layer 8a of raw material powder 30 is laid, and an energy beam (light beam 71) is scanned and irradiated onto the laid layer 8a to form one modeling layer 15a. The modeling step S10 is preferably performed repeatedly when forming the model 15. In the acquiring step S20, the time-series data is preferably acquired each time the modeling step S10 is performed. In the creating step S30, two-dimensional maps 53, 63 are preferably created for each modeling layer 15a. In the detection step S40, the presence or absence of defects in the object can be detected based on the presence or absence of peculiar areas in a first map, which is a two-dimensional map 53, 63 in the first modeling layer, and a second map, which is a two-dimensional map 53, 63 in a second modeling layer adjacent to the first modeling layer on at least one side or the other of the stacking direction of the modeling layer 15a.

[0076] For example, even if a defect such as a cavity occurs during the formation of a certain modeling layer 15a, the cavity may disappear if the layer (layer 8a) of raw material powder 30 laid on that modeling layer 15a is irradiated with an energy beam (light beam 71) to melt and solidify it. In other words, a defect in the modeling layer 15a below the subsequent layer may be repaired during the formation of the subsequent layer. Therefore, even if peculiar regions 55, 65 are recognized in the two-dimensional maps 53, 63, it does not necessarily mean that a defect remains in the modeled object 15 at a position corresponding to the peculiar region 55, 65. According to the method (4) above, the presence or absence of defects in the object 15 is detected based on the presence or absence of peculiar regions 55, 65 in the two-dimensional maps 53, 63 across multiple forming layers 15a, thereby improving the accuracy of detecting the presence or absence of defects in the object 15.

[0077] (5) In some embodiments, in the method of (3) or (4) above, in the modeling step S10, a layer 8a of raw material powder 30 is laid, and an energy beam (light beam 71) is scanned and irradiated onto the laid layer 8a to form one modeling layer 15a. The modeling step S10 is preferably performed repeatedly when forming the model 15. In the acquiring step S20, the time-series data is preferably acquired each time the modeling step S10 is performed. In the creating step S30, two-dimensional maps 53, 63 are preferably created for each modeling layer 15a. In the detecting step S40, it is preferably determined that a defect exists in the modeling layer 15a corresponding to the two-dimensional maps 53, 63 in which the number of peculiar regions 55, 65 exceeds a threshold.

[0078] As described above, even if the peculiar regions 55 and 65 are found in the two-dimensional maps 53 and 63, it does not necessarily mean that defects remain in the object 15 at positions corresponding to the peculiar regions 55 and 65. However, after careful consideration, the inventors discovered that there is a significant difference in the number of unique regions 55, 65 in the two-dimensional maps 53, 63 between a modeling layer 15a with defects and a modeling layer 15a without defects. According to the method (5) above, it is possible to detect the presence or absence of a defect relatively quickly by determining that a defect exists in the forming layer 15a corresponding to the two-dimensional map 53, 63 in which the number of peculiar regions 55, 65 exceeds a threshold value.

[0079] (6) In some embodiments, in any of the methods (1) to (5) above, in the detection step S40, the presence or absence of defects in the object 15 may be detected using a trained model 82 configured to detect the presence or absence of defects based on the two-dimensional maps 53, 63.

[0080] As described above, even if the peculiar regions 55 and 65 are found in the two-dimensional maps 53 and 63, it does not necessarily mean that defects remain in the object 15 at positions corresponding to the peculiar regions 55 and 65. Therefore, for example, by using a trained model 82 that has been machine-learned in advance regarding the relationship between the two-dimensional maps 53 and 63 and the defects remaining in the object 15, the accuracy of detecting the presence or absence of defects in the object 15 can be improved. Therefore, according to the method (6) above, the accuracy of detecting the presence or absence of defects in the shaped object 15 can be improved.

[0081] (7) In some embodiments, in the method of (6) above, in the detection step S40, even if there is a peculiar region 55, 65 having a size equal to or larger than the specified range R, where the deviation of the emission intensity from the average value in the two-dimensional maps 53, 63 exceeds a first threshold value Th1 that is larger than the average value, or where the emission intensity is smaller than a second threshold value Th2 that is smaller than the average value, if it is determined that there is no defect corresponding to the peculiar region 55, 65, the presence or absence of a defect in the molded object 15 may be detected using a trained model 82 configured to determine that there is no defect corresponding to the peculiar region 55, 65.

[0082] According to the method (7) above, the accuracy of detecting the presence or absence of defects in the shaped object 15 can be improved.

[0083] (8) In some embodiments, in the method described in (6) or (7) above, in the modeling step S10, a layer 8a of raw material powder 30 is laid, and an energy beam (light beam 71) is scanned and irradiated onto the laid layer 8a to form one modeling layer 15a. The modeling step S10 may be repeatedly performed when forming the object 15. In the acquiring step S20, time-series data may be acquired each time the modeling step S10 is performed. In the creating step S30, two-dimensional maps 53, 63 may be created for each modeling layer 15a. In the detecting step S40, the presence or absence of a defect in the modeled object 15 may be detected using a trained model 82 configured to determine whether a defect will occur in a modeling layer 15a to be formed after the modeling layer 15a currently being formed in the modeling step S10.

[0084] According to the method (8) above, it is possible to predict the occurrence of defects in subsequent layers before defects occur in the object 15 during modeling, and therefore it is possible to change modeling conditions, such as the irradiation conditions of the energy beam (light beam 71), or to take measures to repair defects so that defects do not occur in the modeling layer 15a where defects are likely to occur or in the predicted defect occurrence position in the modeling layer 15a. This makes it possible to efficiently reduce defects remaining in the object 15. [Explanation of symbols]

[0085] 1. 3D additive manufacturing equipment (additive manufacturing equipment) 2 base plates 2a Drive cylinder 4 cylinders 5 Powder bed forming section 6 Beam Splitter 8 Powder Bed 8a layer 9 Chamber 9a Window 10 Powder laying section 15 Sculptures 15a Modeling layer 20 Light beam irradiation device 21 Light beam irradiation unit 22 Light beam control unit 23 Galvanometer mirror 30 Raw material powder 31 Storage Unit 40 Detection Device 41 Light detection unit 42 Light emission monitor control unit 51 Graph Lines 53 2D Map 55 Singular Area 56 areas 63 2D Map 64 areas 65 Singular Area 71 Light Beam 80 Control Unit 81 Storage section 82 trained models

Claims

1. irradiating a layer of raw material powder with an energy beam while scanning the layer to melt and solidify the raw material powder in the layer, thereby forming a part of a shaped object; acquiring time-series data of light emission intensity of light emitted from the object when the energy beam is scanned and irradiated in the modeling step; creating a two-dimensional map of the emission intensity based on the time series data acquired in the acquiring step and information on the scanning trajectory of the energy beam; detecting the presence or absence of a defect in the object based on the two-dimensional map created in the creating step; Equipped with How to monitor the build process.

2. In the creating step, a spatial filter is applied to the two-dimensional map to create the two-dimensional map after the spatial filter is applied; In the detecting step, the presence or absence of a defect in the object is detected based on the two-dimensional map after the spatial filter created in the creating step is applied. The method of claim 1 .

3. In the detecting step, the presence or absence of a peculiar region having a size equal to or larger than a specified range in which the deviation of the emission intensity from the average value in the two-dimensional map exceeds a first threshold value that is larger than the average value, or the emission intensity falls below a second threshold value that is smaller than the average value, is detected, and the presence or absence of a defect in the shaped object is detected based on the presence or absence of the peculiar region. The method for monitoring a manufacturing process according to claim 1 or 2.

4. In the manufacturing step, a layer of the raw material powder is laid, and the laid layer is irradiated with the energy beam while being scanned to form one manufacturing layer; the step of forming the object is repeatedly performed during the formation of the object, In the acquiring step, the time-series data is acquired every time the modeling step is performed, In the creating step, the two-dimensional map is created for each of the modeling layers; In the detecting step, the presence or absence of a defect in the object is detected based on the presence or absence of the peculiar region in a first map, which is the two-dimensional map in a first modeling layer, and a second map, which is the two-dimensional map in a second modeling layer adjacent to the first modeling layer on at least one side or the other side of the first modeling layer in a stacking direction of the modeling layers. The method of claim 3 .

5. In the manufacturing step, a layer of the raw material powder is laid, and the laid layer is irradiated with the energy beam while being scanned to form one manufacturing layer; the step of forming the object is repeatedly performed during the formation of the object, In the acquiring step, the time-series data is acquired every time the modeling step is performed, In the creating step, the two-dimensional map is created for each of the modeling layers; In the detecting step, it is determined that the defect exists in the modeling layer corresponding to the two-dimensional map in which the number of the peculiar regions exceeds a threshold. The method of claim 3 .

6. the detecting step detects the presence or absence of a defect in the object using a trained model configured to detect the presence or absence of the defect based on the two-dimensional map. The method for monitoring a manufacturing process according to claim 1 or 2.

7. In the detecting step, the presence or absence of a defect in the shaped object is detected using the trained model configured to determine that the defect corresponding to the peculiar region does not exist when it is determined that the defect corresponding to the peculiar region does not exist, even if a peculiar region exists in which the deviation of the emission intensity from the average value in the two-dimensional map exceeds a first threshold value that is larger than the average value, or the emission intensity is below a second threshold value that is smaller than the average value. The method of claim 6 .

8. In the manufacturing step, a layer of the raw material powder is laid, and the laid layer is irradiated with the energy beam while being scanned to form one manufacturing layer; the step of forming the object is repeatedly performed during the formation of the object, In the acquiring step, the time-series data is acquired every time the modeling step is performed, In the creating step, the two-dimensional map is created for each of the modeling layers; In the detecting step, the presence or absence of a defect in the object is detected using the trained model configured to determine whether or not the defect will occur in the modeling layer to be modeled after the modeling layer currently being modeled in the modeling step. The method of claim 6 .

Images