Determination system and determination method

Abstract

To provide a determination system that determines whether or not a desired shaped article can be manufactured at high yield.SOLUTION: A determination system 110 comprises: a detection unit 111 which detects a state of a powder material laid on a powder bed provided in an additive manufacturing apparatus 200 of a binder jetting method, which affects a shape of an additively manufactured shaped article, for each existence region of the powder material; and a determination unit 112 which determines whether or not a state of the powder bed is good by determining whether or not the powder bed is in a state of being ready to additively manufacture a desired shaped article, on the basis of the state of the powder material for each existence region detected by the detection unit 111.SELECTED DRAWING: Figure 1

Description

[Technical field]

[0001] The present disclosure relates to a determination system and a determination method.

[0002] Additive manufacturing (layered manufacturing) is known to include, for example, powder bed fusion and directed energy deposition. In powder bed fusion, additive manufacturing is performed by irradiating a light beam (laser beam, electron beam, etc.) onto a flatly spread powder material (e.g., metal powder). Powder bed fusion methods include selective laser melting (SLM) and electron beam melting (EBM). In directed energy deposition, additive manufacturing is performed by controlling the position of a head that irradiates a light beam and ejects powder material. Directed energy deposition methods include laser metal deposition (LMD) and direct metal deposition (DMP).

[0003] Another method is the binder jetting method. In the powder bed fusion method and the directed energy deposition method, an additive manufacturing product (laminated object) is formed by directly melting and solidifying the powder material using a beam as a heat source. On the other hand, in the binder jetting method, a binder is applied to the powder material (e.g., metal powder) according to the shape of the object, which bonds the powder material together. Next, the binder is removed (debinding heat treatment) and then sintering heat treatment is performed to produce a three-dimensional part.

[0004] Metal Injection Molding (MIM) is a manufacturing method similar to the binder injection method. In MIM, metal powder and binder (binding material, plasticizer, lubricant) are pressurized and mixed to produce pellets (compounds). The pellets are then placed in an injection molding machine where they are plasticized by heating, and the MIM material is injection molded into the cavity of the mold. The molded body is called a green part, and the molded body that has been solvent degreased and heated and degreased is called a brown part. Sintering the brown part produces a sintered product called a silver part.

[0005] As a technology related to additive manufacturing, claim 1 of Patent Document 1 describes an additive manufacturing method for manufacturing a structure by stacking multiple layers one by one from the top surface side, comprising: a first layer formation step of forming at least one first layer; a second layer formation step of forming at least one second layer on the first layer; a surface layer inspection step, after the second layer formation step, of inspecting defects in a surface layer inspection area in the first layer that is shallower than a predetermined surface layer inspectable depth position; a determination step of determining whether or not reinspection is necessary based on the inspection results obtained by the surface layer inspection step; and a reinspection step of reinspecting defects in the first and second layers when it is determined by the determination step that reinspection is necessary. [Prior art documents] [Patent documents]

[0006] [Patent Document 1] JP 2020-7604 A Summary of the Invention [Problem to be solved by the invention]

[0007] In the binder jet additive manufacturing device described in Patent Document 1, the top surface of the solidified layer is heated from above for a predetermined period of time to determine whether there are any defects inside the solidified layer (paragraphs 0025, 0039, 0042, 0085). A solidified layer in which unintended defects are detected cannot be considered a desired object. For this reason, additive manufacturing is performed again to manufacture the desired object. As a result, the yield rate during additive manufacturing decreases. The problem to be solved by the present disclosure is to provide a determination system and a quality determination method for determining whether a desired object can be manufactured with a high yield. [Means for solving the problem]

[0008] The judgment system disclosed herein includes a detection unit that detects a state of a powder material spread on a powder bed provided in a binder jetting additive manufacturing device, the state affecting the shape of a model to be additively manufactured, for each region of the powder material, and a judgment unit that judges whether the powder bed is in a state in which a desired model can be additively manufactured based on the state of the powder material for each region of the powder material detected by the detection unit, thereby judging whether the state of the powder bed is good or bad. Other solutions will be described later in the description of the embodiment of the invention. Effect of the Invention

[0009] According to the present disclosure, it is possible to provide an assessment system and a quality assessment method for determining whether a desired object can be manufactured with a high yield. [Brief description of the drawings]

[0010] [Figure 1] FIG. 1 is a block diagram showing an additive manufacturing condition determination device and judgment system of the present disclosure. [Figure 2A] FIG. 13 is a side view illustrating a method for determining unevenness of powder material in a powder bed. [Figure 2B] FIG. 13 is a top view illustrating a method for determining unevenness of powder material in a powder bed. [Figure 3A] FIG. 13 is a side view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Figure 3B] FIG. 13 is a top view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Figure 4] FIG. 13 is a side view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Diagram 5] FIG. 13 is a side view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Figure 6A] FIG. 13 is a side view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Figure 6B] FIG. 13 is a top view illustrating a method for determining evenness of powder material in a powder bed in another embodiment. [Figure 6C] FIG. 6B is an enlarged view of part A in FIG. 6A. [Figure 7A] FIG. 11 is a side view illustrating a method for leveling the powder bed when spreading the powder. [Figure 7B] FIG. 7B is a side view illustrating a method for leveling the powder bed when spreading the powder, and shows the process following FIG. 7A. [Figure 8] FIG. 2 is a block diagram showing a hardware configuration of a determination system according to the present disclosure. [Figure 9] 1 is a flowchart illustrating a determination method of the present disclosure. [Figure 10] FIG. 2 is a perspective view showing the shape of a green part used for evaluation. [Figure 11] FIG. 2 is a perspective view showing the shape of a silver part used for evaluation. [Figure 12A] 12 is a top view of a structure that can be used in place of the green part shown in FIG. 10 and the silver part shown in FIG. 11. [Figure 12B] FIG. 12B is a side view of the structure shown in FIG. 12A. [Figure 13] 1 is a flowchart illustrating an additive manufacturing method of the present disclosure. [Figure 14] 1 is a flowchart showing the first stage of the additive manufacturing condition determination method of the present disclosure. [Figure 15]1 is a flowchart showing the latter stage of the additive manufacturing condition determination method of the present disclosure. [Figure 16A] FIG. 1 is a schematic diagram of an additive manufacturing apparatus of the present disclosure viewed from the side. [Figure 16B] FIG. 1 is a schematic diagram of an additive manufacturing apparatus of the present disclosure viewed from above. [Figure 17A] FIG. 13 is a schematic diagram showing an example of a powder-laying operation. [Figure 17B] FIG. 13 is a schematic diagram showing an example of a powder-laying operation. [Figure 17C] FIG. 13 is a schematic diagram showing an example of a powder-laying operation. [Figure 17D] FIG. 13 is a schematic diagram showing an example of a powder-laying operation. [Figure 18A] FIG. 4 is a schematic diagram showing an example of a coating operation. [Figure 18B] FIG. 4 is a schematic diagram showing an example of a coating operation. [Figure 19] FIG. 11 is a diagram illustrating an example of a drying operation. DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS

[0011] Hereinafter, a form for carrying out the present disclosure (referred to as an embodiment) will be described with reference to the drawings. In the following description of one embodiment, other embodiments applicable to the one embodiment will also be described as appropriate. The present disclosure is not limited to the following one embodiment, and different embodiments can be combined with each other or modified as desired without significantly impairing the effects of the present disclosure. In addition, the same symbols will be used for the same members, and duplicate descriptions will be omitted. Furthermore, the same names will be used for members having the same functions. The contents shown are merely schematic, and for the sake of illustration, changes may be made from the actual configuration within a range that does not significantly impair the effects of the present disclosure, and some members may be omitted or modified between drawings. In addition, the same embodiment does not necessarily need to have all the configurations.

[0012] 1 is a block diagram showing an additive manufacturing condition determination device 100 and a judgment system 110 of the present disclosure. Fig. 1 also shows a binder jet type additive manufacturing device 200 (additive manufacturing device), an input device 300, and an output device 400.

[0013] Additive manufacturing using the additive manufacturing device 200 broadly includes four processes. That is, the processes include lamination (powder laying, binder application, and drying), resin curing, degreasing, and sintering. Green parts, which are intermediate structures, are manufactured by lamination and resin curing. Silver parts, which are the final structures, are manufactured by degreasing and sintering the intermediate structures.

[0014] Among these steps, powder laying, which is the so-called "entrance" of additive manufacturing, is preferably carried out so as to obtain a powder bed 3 (FIGS. 2A and 2B) capable of forming a desired object (structure, product). The powder bed 3 is a powder material (e.g., metal powder) spread in a modeling tank 2 (described later). That is, the powder bed 3 can be formed by spreading the powder material in the modeling tank 2. If the condition of the powder bed 3 is poor, the accuracy, condition, etc. of the object (product) obtained using the powder bed 3 may also be poor.

[0015] In other words, in the layering process during additive manufacturing, the influence of control factors related to powder laying may cause variation in the state (e.g., density) of the powder material at each position in the powder bed 3. Furthermore, binder injection during additive manufacturing is performed on the powder bed 3 formed in the modeling tank 2 to form a three-dimensional shape. For this reason, regardless of the conditions for applying and drying the binder, the state of the powder material in the powder bed 3 may vary.

[0016] The variation is caused, for example, by inappropriate control factors related to the supply of powder material and the method of powder transportation. The variation in state leads to variation in the state of the green parts, and even if the control factors of the subsequent processes are changed, the influence and effect may not be accurately reflected. For this reason, it is preferable to optimize the formation state of the powder bed 3 by the powder spreading process (control factors for spreading the powder material) during the lamination process. By optimizing, the influence of the control factors related to binder application and drying in the lamination process can be evaluated, and the variation of the evaluations in the subsequent processes can also be suppressed.

[0017] Therefore, in the present disclosure, among the above-mentioned steps, the powder-laying step is optimized by the determination system 110. Specifically, the determination system 110 determines whether the state of the powder bed 3 formed in the powder-laying step is good or bad. As a result, a molded object can be obtained using a powder bed 3 in good condition, and the quality of the molded object can be made as desired. A powder bed 3 in good condition can make a molded object obtained by, for example, additive manufacturing have a desired state (quality, physical properties, etc.).

[0018] In the following explanation, for convenience of explanation, the judgment system 110 that optimizes the powder-laying process will be explained, and then the determination device 100 that optimizes each of the above processes other than the powder-laying process will be explained. Note that in the explanation of the judgment system 110, the judgment system 110 itself will be mainly explained. Meanwhile, explanations of each part other than the judgment system 110 (the determination device 100, additive manufacturing device 200, input device 300, output device 400, etc.) will be given while explaining the determination device 100, with reference to FIG. 10 and subsequent figures.

[0019] The judgment system 110 is a system that detects "uneven layering" of the powder material in the powder bed 3 to judge whether the powder bed 3 is suitable for additive manufacturing, i.e., judges the quality of the powder bed 3. The judgment system 110 includes a detection unit 111, a judgment unit 112, and a learning unit 113 including a database 113a.

[0020] The detection unit 111 detects the state of the powder material spread on the powder bed 3 of the additive manufacturing apparatus 200, which affects the shape of the additively manufactured object, for each of the powder material existence regions A and B (described later). If the state of the powder material differs for each of the existence regions A and B of the powder bed 3, the state of each part of the object obtained by injecting the binder into the powder bed 3 will also differ. This may cause the occurrence of shrinkage rates and the like in each process, such as binder degreasing and sintering, to differ, which may cause deformation and cracking of the object. For this reason, it is preferable that the state that affects the shape of the additively manufactured object is uniform (ideally completely the same) throughout the powder bed 3. The term "uniform" here refers to a state in which, even if the state differs for each of the existence regions A and B of the powder bed 3, there is a difference that does not have a significant effect on the shape of the object (dimensions, the presence or absence of large deformation, the presence or absence of cracks, etc.) (desired dimensions are obtained).

[0021] The detection unit 111 detects the state of each existence region A in the powder bed 3 by measuring an index value that affects the shape of the molded object for the powder material in each existence region A and B in the powder bed 3. The state of the powder material in each existence region A and B is a quantifiable index value, such as density, in the existence regions A and B, as will be described in detail later. There are multiple existence regions A on the upper surface 35 of the powder bed 3 (FIGS. 2A and 2B), and they extend in the horizontal direction. On the other hand, there are multiple existence regions B in the height direction of the powder bed 3, and they extend in the vertical direction, as will be described in detail later.

[0022] If the index values ​​(reference values ​​for existence regions A and B) indicating the overall state of the powder material are different in existence regions A and B, which have a certain area, the state of the parts of the molded object corresponding to existence regions A and B may also change. For this reason, it is preferable that the state is uniform as described above, and by detecting using the index value, which is a numerical value, it is possible to detect what may be called "unevenness" (non-uniform state) of the powder material in the powder bed 3.

[0023] FIG. 2A is a side view illustrating a method for determining unevenness in the laying of powder material in the powder bed 3. FIG. 2B is a top view illustrating a method for determining unevenness in the laying of powder material in the powder bed 3. The determination system 110 includes a measurement mechanism 27 and a heating mechanism 23. In the example of the present disclosure, both the measurement mechanism 27 and the heating mechanism 23 are provided in the additive manufacturing apparatus 200. The heating mechanism 23 heats the upper surface 35 of the powder bed 3. The heating mechanism 23 is a heat source that can provide a large amount of energy to the upper surface 35, which is the heated surface, in a short time (instantaneous) without contact. The heating mechanism 23 preferably heats the entire upper surface 35 uniformly, but the heating may be uneven. The heating mechanism 23 is preferably a mechanism that can heat the entire upper surface 35, and is, for example, a halogen heater or a device that irradiates blurred laser light. The measurement mechanism 27 measures the temperature (for example, temperature distribution) of the upper surface 35 heated by the heating mechanism 23. The measuring mechanism 27 is an infrared camera (IR camera) or the like.

[0024] The detection unit 111 (FIG. 1) detects the state of each of the existence regions A in the powder bed 3 by evaluating the temperature drop rate (heat removal rate) on the upper surface 35 obtained using the measurement mechanism 27 after heating by the heating mechanism 23. In the existence regions A containing powder material, the temperature drop rate differs due to differences in the state of, for example, density. Thus, by evaluating the temperature drop rate using the measurement mechanism 27, it is possible to detect the difference in state among the multiple existence regions A. The temperature drop rate can be evaluated, for example, by comparing the temperature of the upper surface 35 immediately after heating by the heating mechanism 23 is completed with the temperature of the upper surface 35 after a predetermined time (for example, 1 second or more and 10 seconds or less) has elapsed since the end of heating.

[0025] It is preferable that the heating mechanism 23 heats the upper surface 35 instantaneously. When heated, heat is transferred downward from the upper surface 35 to the bottom plate 22. In addition, heat tends to be transferred horizontally inside the powder bed 3, which may make it difficult to grasp the state of each of the existence regions A extending in the horizontal direction. Therefore, by heating instantaneously, it is possible to heat only the vicinity of the upper surface 35 as much as possible, making it easier to grasp the state of each of the existence regions A. The driving control of the heating mechanism 23 can be performed by, for example, the detection unit 111.

[0026] The conditions for the instantaneous heating are not particularly limited. For example, the instantaneous heating can be performed by heating the entire upper surface 35 for a time period of 0.1 to 10 seconds. As described above, the preferred heating time period in the present disclosure has a certain range, so the term "instantaneous" is used to include the heating time period in the present disclosure in the general concept of "instantaneous." The heating temperature is not particularly limited, but it is preferable to heat the material at a temperature higher than the drying temperature described below and at a temperature at which the binder does not disappear (for example, 100°C to 300°C).

[0027] FIG. 3A is a side view illustrating a method for determining unevenness in the laying of a powder material in a powder bed 3 in another embodiment. FIG. 3B is a top view illustrating a method for determining unevenness in the laying of a powder material in a powder bed 3 in another embodiment. The determination system 110 illustrated in FIG. 3A and FIG. 3B includes a pressurizing mechanism 24. In the example of the present disclosure, the pressurizing mechanism 24 is provided in the additive manufacturing apparatus 200, and specifically, is configured integrally with the powder laying mechanism 4. The powder laying mechanism 4 lays a powder material on the bottom plate 22 that forms the powder bed 3, and is provided in the additive manufacturing apparatus 200. Although details will be described later, the pressurizing mechanism 24 can be disposed above the powder bed 3 by moving the powder laying mechanism 4 in the horizontal direction.

[0028] The pressurizing mechanism 24 pressurizes the upper surface 35 of the powder bed 3 at a desired position. Pressurization can be performed by pressing with a constant load or stroke. The pressurizing mechanism 24 includes a columnar body 25 (terminal) such as a needle, and the columnar body 25 faces the upper surface 35. A plurality of columnar bodies 25 are provided, for example, so as to be distributed over the entire upper surface 35 at equal intervals. The pressurizing mechanism 24 can pressurize the upper surface 35 by abutting the columnar body 25 against the upper surface 35 while the columnar body 25 is disposed on the upper surface 35 of the powder bed 3.

[0029] The detection unit 111 (FIG. 1) detects the state of each existence region A in the powder bed 3 by evaluating the pressure on the upper surface 35 at the position where pressure is applied. In the example of the present disclosure, the reaction force in the pressurizing mechanism 24 is evaluated, and the pressure distribution on the upper surface 35 is detected by detection. In the existence region A containing the powder material, for example, due to differences in the state of the powder material, the pressure (e.g., reaction force) on the upper surface 35 when the columnar body 25 hits it may differ. Thus, by evaluating the pressure, the difference in state in the multiple existence regions A can be detected.

[0030] FIG. 4 is a side view for explaining a method for determining unevenness in the spreading of powder material on the powder bed 3 in another embodiment. The detection unit 111 (FIG. 1) detects the state of each of the existence regions A on the powder bed 3 in a plurality of existence regions A having a predetermined range on the upper surface 38 of the layer 37 for each layer 37 having a predetermined thickness spread by the powder spreading mechanism 4. The powder spreading mechanism 4 usually spreads the powder material on the powder bed 3 one layer at a time. Therefore, in the powder bed 3, a plurality of layers 37 of the powder material arranged in the horizontal direction are stacked. Therefore, the detection unit 111 detects the state of each existence region A present on the upper surface 38 of the spread layer 37 for each layer spread by the powder spreading mechanism 4. This makes it possible to easily detect the entire powder bed 3.

[0031] FIG. 5 is a side view illustrating a method for determining unevenness in the laying of a powder material on a powder bed 3 in another embodiment. In the determination system 110 shown in FIG. 5, the pressurizing mechanism 24 is independent of the powder laying mechanism 4. In the example of the present disclosure, the pressurizing mechanism 24 is provided in the additive manufacturing apparatus 200, and specifically, is configured to be movable up and down above the powder bed 3 by an actuator (not shown). The detection unit 111 (FIG. 1) can pressurize the upper surface 35 by driving the actuator to lower the pressurizing mechanism 24 so that the pillars 25 abut against the upper surface 35. After pressing, the detection unit 111 can move the pillars 25 away from the upper surface 35 by driving the actuator to raise the pressurizing mechanism 24.

[0032] 2A to 4, the presence region A extends in the horizontal direction on the upper surface 35. Therefore, the detection unit 111 (FIG. 1) detects the state of each of the presence regions A on the powder bed 3 in a plurality of presence regions A having a predetermined range in the horizontal plane of the powder bed 3. However, the presence region in which the powder material exists may be the presence region B extending in the height direction of the powder bed 3.

[0033] Fig. 6A is a side view illustrating a method for determining unevenness in spreading of powder material in the powder bed 3 in another embodiment. Fig. 6B is a top view illustrating a method for determining unevenness in spreading of powder material in the powder bed 3 in another embodiment. In a determination system 110 shown in Figs. 6A, 6B, and Fig. 6C described later, a detection unit 111 (Fig. 1) detects the state of each of a plurality of existence regions B in the powder bed 3, each having a predetermined range in the height direction of the powder bed 3.

[0034] The determination system 110 includes an irradiation mechanism 26 (e.g., a transmitting probe) that irradiates ultrasonic waves onto the powder bed 3 from the side of the powder bed 3. In the example of the present disclosure, the irradiation mechanism 26 is included in the additive manufacturing apparatus 200. In the example of the present disclosure, the irradiation mechanism 26 is, for example, an ultrasonic receiving probe that also has a function of receiving ultrasonic waves reflected inside the powder bed 3. Therefore, the irradiation mechanism 26 is an ultrasonic transmitting and receiving probe. However, the determination system 110 may include a receiving device (not shown, for example, a receiving probe) that receives ultrasonic waves on the opposite side of the irradiation mechanism 26 from the powder bed 3.

[0035] Irradiation mechanism 26 is driven along the extending direction (horizontal direction) of upper surface 35 of powder bed 3. That is, irradiation mechanism 26 reciprocates from one direction to the other in a horizontal plane. Irradiation mechanism 26 is also configured to be movable up and down. This allows ultrasonic waves to be irradiated at different heights in powder bed 3.

[0036] Fig. 6C is an enlarged view of part A in Fig. 6A. Irradiation mechanism 26 is driven as shown by the thick solid arrow in Fig. 6C while emitting ultrasonic waves. That is, irradiation mechanism 26 moves from one direction to another direction in the horizontal plane at the highest part of powder bed 3. After irradiation mechanism 26 reaches the vicinity of the end of powder bed 3, irradiation mechanism 26 lowers its position in the height direction and moves from the other direction to the one direction. Then, irradiation mechanism 26 repeats these operations.

[0037] In this manner, the detection unit 111 (FIG. 1) irradiates ultrasonic waves to the powder bed 3 from the side of the powder bed 3 using the irradiation mechanism 26, and detects the state of each existence region B in the powder bed 3 of the powder material laid on the powder bed 3 using at least one of the transmitted or reflected ultrasonic waves. In the example of the present disclosure, the detection unit 111 uses the reflected ultrasonic waves. If the state of each existence region B in the height direction in the powder bed 3 differs, the state of the transmitted or reflected ultrasonic waves (e.g., reception time, reception intensity, etc.) may differ. For this reason, the detection unit 111 can detect the state of the powder material in each existence region B extending in the height direction by irradiating ultrasonic waves from the side.

[0038] The size (length in the height direction) of the predetermined region B may be the length of one layer of the powder material in the powder bed 3, or may be the length of multiple layers.

[0039] Furthermore, the state evaluation in the predetermined regions A and B may be performed by irradiating the bottom surface of the modeling tank 2 (powder bed 3) with ultrasonic waves from the outside toward the inside of the powder bed 3 to grasp the state inside the powder bed 3 (presence of voids, etc.). Furthermore, in each of the above-described embodiments, X-rays or the like may be used instead of ultrasonic waves.

[0040] Returning to FIG. 1, the index value includes the density of the powder material in the existence regions A and B. However, the index value is not limited to density, and may be void ratio or the like. The existence regions A and B have a certain area or volume. Therefore, if the entire powder bed 3 is completely uniform, the index value such as density has a certain value regardless of the existence regions A and B. However, if the density in the existence regions A and B (density of the existence regions A and B) is different, the density of the part of the molded object corresponding to the existence regions A and B may also change. For this reason, the occurrence conditions of the shrinkage rate in each process such as binder degreasing and sintering may also differ, which may cause deformation and cracking of the molded object. For this reason, by detecting using density, it is possible to detect the so-called "unevenness" of the powder material in the powder bed 3.

[0041] The determination unit 112 determines whether the powder bed 3 is in a state where a molded object having a desired shape can be additively manufactured based on the state of the powder material for each of the existence regions A and B detected by the detection unit 111, thereby determining whether the state of the powder bed 3 is good or bad. As described above, depending on the state of the powder bed 3 (e.g., density, etc.), the shape, dimensions, etc. of the molded object manufactured using the powder material spread on the powder bed 3 may change. Therefore, it is determined whether the state of the powder bed 3 is in a state where a desired molded object can be manufactured. Then, if the state of the powder bed 3 is in a state where a desired molded object can be manufactured, the state of the powder bed 3 is determined to be good. On the other hand, if the state of the powder bed 3 is not in a state where a desired molded object can be manufactured, the state of the powder bed 3 is determined to be bad.

[0042] The judgment unit 112 judges that the condition of the powder bed 3 is good when it judges that the index value of the powder material for each of the existence regions A and B in the powder bed 3 is uniform in a predetermined direction, which is at least one of the horizontal direction and the height direction of the powder bed 3 (or both). The index value is the same type of value measured by the same method regardless of the existence regions A and B. For this reason, it is considered that if index values ​​are obtained for all the existence regions A and B, and the powder bed 3 is uniform based on all the obtained index values, it is possible to additively manufacture a desired shaped object regardless of the existence regions A and B. For this reason, the quality of the condition of the powder bed 3 can be judged by using whether or not it is uniform as a judgment criterion.

[0043] When the difference between the maximum and minimum values ​​of the index values ​​in a predetermined direction is within a predetermined range, the judgment unit 112 judges that the state of the powder material in each of the existence regions A and B in the powder bed 3 is uniform. Since the index value is, for example, a numerical value, the difference between the maximum and minimum values ​​can be calculated. The smaller the difference, the closer to uniformity it is. Ideally, the difference is 0, and in this case, the powder bed 3 is in a completely uniform state. Therefore, by judging whether the difference is within a predetermined range, it is possible to judge whether the powder material is uniform or not.

[0044] The "predetermined range" here is a predetermined range for determining whether the powder bed 3 is uniform or not, and if the powder bed 3 falls within this range, it is considered to be uniform (including approximately uniform). The predetermined range can be determined, for example, based on experiments, simulations, etc.

[0045] The judgment unit 112 makes a judgment using a plurality of existence regions A extending in the horizontal direction and a plurality of existence regions B extending in the height direction of the powder bed 3. For example, in the case of the existence region A shown in FIG. 2B and the like, the judgment unit 112 judges whether the state of the powder bed 3 is good or bad based on each state of the plurality of existence regions A in the horizontal plane of the powder bed 3 detected by the detection unit 111. In this way, the state (variation, unevenness) of the powder bed 3 in the horizontal plane can be judged.

[0046] Also, for example, in the case of presence region B shown in FIG. 6C etc., the judgment unit 112 judges whether the state of the powder bed 3 is good or bad based on each state of the multiple presence regions B in the height direction of the powder bed 3 detected by the detection unit 111. In this manner, the state (variation, unevenness) of the powder bed 3 in the height direction can be judged. Then, when it is judged that both the state in the horizontal direction and the state in the height direction are uniform, it can be judged that the entire powder bed 3 is uniform.

[0047] The learning unit 113 performs machine learning using the formation conditions of the powder bed 3 and at least one (or both) of predetermined information of the index values ​​for each of the existence regions A and B and the quality of the powder bed 3 as features. By providing the learning unit 113, the formation conditions of the powder bed 3 can be predicted based on the predetermined information. The machine learning can be performed, for example, by supervised learning with the predetermined information as an explanatory variable and the formation conditions of the powder bed 3 as an objective variable. However, it is not limited to supervised learning, and either unsupervised learning or reinforcement learning may be used. For example, Bayesian optimization, kernel ridge regression analysis, etc. can be used as the machine learning algorithm. The formation conditions of the powder bed 3 and the predetermined information are associated with each other and stored in the database 113a.

[0048] For example, changing the formation conditions (powder spreading conditions) of the powder bed 3 may change at least one of the index values ​​(e.g., density) for each of the existence regions A and B, or the quality of the powder bed 3. For example, if the supply speed of the powder material to the bottom plate 22 (FIG. 2A, an example of a base) that forms the powder bed 3 changes, the density of the powder material for each of the existence regions A and B may also change. As a result, the quality of the powder bed 3 may also change. Therefore, by performing machine learning using the formation conditions of the powder bed 3 and specified information, it is possible to determine the formation conditions of the powder bed 3 that will improve the state of the powder bed 3.

[0049] The formation conditions of the powder bed 3 used for machine learning preferably include at least one of the following: the layer thickness of each layer made of powder material, the supply amount of powder material to the powder bed 3 (output, time, etc. of the vibration mechanism 11 described later), the supply speed of the powder material to the powder bed 3, the powder spreading speed when spreading the powder material on the powder bed 3 (movement speed of the powder spreading mechanism 4 described later), or the leveling setting when spreading the powder material on the powder bed 3 (the rotation speed of the roller (an example of the leveling mechanism 12), etc.). These are parameters that affect the state of the powder bed 3. For this reason, by including at least one of these, it is possible to determine the formation conditions corresponding to the quality of the powder bed 3.

[0050] 7A is a side view illustrating a method for leveling the powder bed 3 when spreading the powder. A vibration mechanism 11 may be used to promote uniformity of the condition of the powder bed 3 when spreading the powder. The vibration mechanism 11 is provided under the powder bed 3. The vibration mechanism 11 provided in the powder bed 3 vibrates at least the upper surface 35 of the powder bed 3, and is provided in the determination system 110. The vibration mechanism 11 is also provided in the powder supply mechanism 10 that supplies powder.

[0051] The powder bed 3 is formed by moving the powder spreading mechanism 4 in the powder spreading direction while rotating the leveling mechanism 12 such as a roller. By vibrating the entire powder bed 3 with the vibration mechanism 11 while the leveling mechanism 12 is rotating, the voids present inside the powder bed 3 can be filled with powder material. This increases the overall density of the powder bed 3, making it easier to make the entire powder bed 3 uniform.

[0052] FIG. 7B is a side view explaining a method of leveling the powder bed 3 when the powder is laid, and shows the state performed following FIG. 7A. After leveling by the leveling mechanism 12, the vibration mechanism 11 vibrates at least the upper surface 35 of the powder bed 3. Then, the detection unit 111 detects the state of each existence region A in the powder bed 3 by evaluating the behavior of the powder material on the upper surface 35 vibrated by the vibration mechanism 11. If the state of the powder material in the powder bed 3 is different, the state in which the powder material vibrates when the powder bed 3 is vibrated may also be different. Therefore, by evaluating the difference in the possible states, the state of each existence region A in the powder bed 3 can be detected. The state of the vibration can be grasped, for example, by photographing with a camera 39 installed above the powder bed 3.

[0053] 8 is a block diagram showing a hardware configuration of the determination system 110. The determination system 110 is configured to include, for example, a CPU (Central Processing Unit) 1001, a RAM (Random Access Memory) 1002, a ROM (Read Only Memory) 1003, an I / F (Interface) 1004, a bus 1005, and the like. The CPU 1001, the RAM 1002, the ROM 1003, and the I / F 1004 are connected via, for example, the bus 1005. The determination system 110 is realized by a predetermined control program (for example, the determination method of the present disclosure) stored in the ROM 1003 being expanded in the RAM 1002 and executed by the CPU 1001. Signals and information are exchanged between the determination system 110 and various devices (such as servers), an external network, and the like, in terms of hardware, through the I / F 1004.

[0054] FIG. 9 is a flowchart showing the determination method of the present disclosure. The determination method of the present disclosure is a method for determining whether or not the powder bed 3 is suitable for additive manufacturing, that is, the quality of the powder bed 3, by detecting "uneven layering" of the powder material in the powder bed 3. The determination method of the present disclosure includes steps S21 to S29. The determination method of the present disclosure can be executed by a determination system 110. For this reason, in the following description, FIG. 1 will be referred to as appropriate. In the following description, descriptions that overlap with the description of the determination system 110 will be omitted.

[0055] In step S21, the formation conditions of the powder bed 3 are input to the determination system 110. The input can be performed through the input device 300. The formation conditions of the powder bed 3 (powder-laying conditions) are parameters that differ for each additive manufacturing apparatus 200, but some of them are common. Therefore, it is preferable that the input formation conditions are shared parameters. In step S22, a powder-laying operation (powder-laying process) is performed by the additive manufacturing apparatus 200 using the formation conditions input in step S21. The powder-laying operation can be performed using the powder-laying mechanism 4.

[0056] In step S23, it is determined whether or not the powder spreading has been completed to a predetermined thickness. The predetermined thickness may be a thickness corresponding to one layer of metal particles (particle diameter of metal particles), or may be a thickness of multiple layers of metal particles stacked together. If the powder spreading has not been completed to the predetermined thickness in step S23 (No), step S22 is performed again. On the other hand, if the powder spreading has been completed to the predetermined thickness (Yes), step S24 is performed.

[0057] In step S24 (detection step), the state (variation) of the metal particles on the upper surface 35 is detected. Specifically, the state of the powder material spread on the powder bed 3 provided in the binder jetting type additive manufacturing apparatus 200, which affects the shape of the additively manufactured object, is detected for each powder material presence region A. In step S24, detection is performed on the upper surface 35 extending in the horizontal direction of the powder bed 3. The detection can be performed by the detection unit 111.

[0058] In step S25, it is determined whether detection has been completed at all evaluation heights. Therefore, the above step S24 is performed for the entire height direction of the powder bed 3 to be finally used, for example, by dividing it into sections for each existence region B. If detection has not been completed at all evaluation heights (No), that is, if there is an existence region B that has not yet been detected, step S22 and subsequent steps are performed again. On the other hand, if detection has been completed at all evaluation heights (Yes), step S26 and step S28 are performed.

[0059] First, in step S26 (judgment step), it is judged whether the powder bed 3 is in good condition as a whole. That is, based on the state of the powder material for each of the existence regions A and B detected in step S24 (detection step), it is judged whether the powder bed 3 is in a state where a desired object can be additively manufactured, thereby judging whether the state of the powder bed 3 is good or bad. The judgment can be executed by the judgment unit 112. Note that, although the judgment is performed here for all evaluation heights (heights in the powder bed 3 that are finally used), the judgment may be performed each time powder spreading to a predetermined thickness is completed in step S23.

[0060] If it is determined in step S26 that the entire powder bed 3 is in a state where the desired object can be additively manufactured (Yes), the formation conditions for the powder bed 3 input in step S21 are determined as the powder laying conditions (step S27).

[0061] Meanwhile, as described above, step S25 is followed by step S28. In step S28, the forming conditions for the powder bed 3 (step S21), the detection results (e.g., index values), and the existence regions A and B in which the detection results are obtained are stored as learning data in the database 113a.

[0062] In step S29 (learning step), machine learning is performed using, as features, the formation conditions of the powder bed 3 and at least one of predetermined information of index values ​​for each of the existence regions A and B and the quality of the powder bed 3. In the example of FIG. 9, the predetermined information is only the index values ​​for each of the existence regions A and B. The learning can be performed by the learning unit 113. As a result of the machine learning, conditions under which the powder bed 3 is considered to be in a state in which a desired object can be additively manufactured in all of the existence regions A and B are derived as "recommended formation conditions."

[0063] The recommended forming conditions are input in step S21 as the forming conditions of the powder bed 3. That is, powder spreading and detection are performed again under the recommended forming conditions determined by the learning unit 113.

[0064] As described above, by using the judgment system 110 and judgment method disclosed herein for a powder bed 3 on which powder has been laid based on initial powder laying conditions, conditions with less density variation can be selected from the results of evaluating density variation, for example, and then the lamination process including binder application and drying can be evaluated, thereby making it possible to appropriately evaluate the influence of control factors on the quality of the green part.

[0065] Next, a description will be given of the determination device 100. The hardware configuration of the determination device 100 is the same as that of the determination system 110 shown in FIG.

[0066] In additive manufacturing using the additive manufacturing apparatus 200, it is preferable to set appropriate manufacturing conditions (recipe) for additive manufacturing according to the conditions of the material, manufacturing apparatus, etc. As described above, additive manufacturing broadly includes four processes. That is, the processes of lamination (powder laying, binder application, and drying), resin curing, degreasing, and sintering are included. Of these processes, the conditions of the powder laying process (e.g., part of the lamination process), i.e., the conditions for forming the powder bed 3, can be optimized by the method described with reference to Figs. 1 to 9 above. By optimizing, it is possible to suppress the occurrence of abnormalities in the molded object caused by the powder laying process (cracks, deformation, excessive shrinkage, etc.).

[0067] To optimize each process other than the powder-laying process, it is preferable to evaluate the performance of the product (obtained structure) in each process. However, evaluation of the product itself may be difficult due to circumstances such as the state of the product and the convenience of the process. Therefore, in the example of the present disclosure, the dimensions, mass, density, etc. of the green part (intermediate structure), which is the molded product after the resin has hardened and before degreasing, are measured. Furthermore, the dimensions, weight, density, etc. of the silver part (final structure), which is the molded product after the sintering process, are also measured.

[0068] For example, degreasing the green part reduces the density of the green part by about 55% to 65%, and the bond becomes loose. Note that 100% density is defined as a state in which the powder material (e.g., metal powder) used in additive manufacturing is densely packed. The final density of the silver part is improved to about 95% to 99% compared to the green part after degreasing. Thus, during additive manufacturing, a large shrinkage deformation occurs, for example, in the powder metallurgy bonding process. Therefore, the shrinkage rate, anisotropy, and density change caused by, for example, sintering the green part affect the accuracy of the final shape of the silver part. Therefore, as disclosed herein, it is preferable to evaluate the shape and density of the green part manufactured during the manufacturing process of the silver part.

[0069] The determination device 100 determines manufacturing conditions during additive manufacturing in an additive manufacturing device 200 that performs additive manufacturing using a binder jetting method. The manufacturing conditions include first manufacturing conditions during lamination and resin hardening, and second manufacturing conditions during degreasing and sintering. Note that since it is preferable to optimize the conditions during the powder-laying process, which is part of the lamination process (i.e., the conditions for forming the powder bed 3), using the method described above with reference to Figures 1 to 9, it is preferable that the first manufacturing conditions do not include the conditions during the powder-laying process. However, the first manufacturing conditions may include the conditions during the powder-laying process.

[0070] The manufacturing conditions include many parameters (control factors described later). The parameters may also vary depending on the metal material used, the specifications of the additive manufacturing device 200, and the shape and density of the desired final structure (silver part). For this reason, it is not easy to consider appropriate parameters and determine manufacturing conditions each time manufacturing is performed. Therefore, by using the determination device 100, appropriate manufacturing conditions can be searched for and the manufacturing conditions can be determined.

[0071] The determination device 100 includes a first condition determination unit 101, a first learning unit 102, a second condition determination unit 103, and a second learning unit 104.

[0072] The first condition determination unit 101 determines the first manufacturing conditions from the shape and density of the first structure manufactured by lamination and resin curing performed during additive manufacturing using the binder jetting method. The first manufacturing conditions are manufacturing conditions during lamination and resin curing for manufacturing a first structure having a desired shape and density, as described above. The first structure is a structure (intermediate structure) obtained in the manufacturing process of a second structure described below. By providing the first condition determination unit 101, it is possible to determine the manufacturing conditions for the first structure having a desired shape and density.

[0073] The first condition determination unit 101 determines, as the first manufacturing conditions, the manufacturing conditions when manufacturing a first structure in which the dimensional difference and density difference from the desired shape are both within a predetermined range. The shape and density of the actually manufactured first structure are compared with the desired shape and density, and if the difference is within a predetermined range, it can be determined that the manufactured first structure is close to the desired first structure. The difference here is the dimensional difference between corresponding sides in the case of shape, and the density difference in the case of density. Therefore, the first condition determination unit 101 can determine the manufacturing conditions when such a first structure can be manufactured as the first manufacturing conditions when manufacturing a first structure having a desired shape and density.

[0074] The predetermined range here refers to an allowable difference in size and density that is expected due to lamination and resin curing. Specific values ​​of the predetermined range can be determined, for example, by experiment.

[0075] Fig. 10 is a perspective view showing the shape of the green part 30 used for evaluation. The green part 30 is manufactured as a first structure under, for example, initially set first manufacturing conditions, and the first condition determination unit 101 determines the first manufacturing conditions using the manufacturing conditions, the shape and density of the green part 30, etc. Therefore, the green part 30 shown in Fig. 10 is a structure manufactured for determining the first manufacturing conditions.

[0076] The green part 30 (first structure) includes a plurality of unit structures 31, 32, and 33 that are different in volume but have the same shape. Therefore, the unit structures 31, 32, and 33 have a similar relationship. The unit structures 31, 32, and 33 each have three types of block shapes: large, medium, and small. The block shapes have sides that are formed by straight lines. The block shapes make it easier to measure the length of each side, and make it easier to calculate the dimensional difference. In the illustrated example, the unit structures 31, 32, and 33 are all cubes. By using cubes, the dimensions in each direction of the x, y, and z axes can be measured, and the shrinkage in each direction due to lamination, etc. can be evaluated. In addition, the effects of temperature distribution due to differences in volume, thermal stress due to temperature distribution, and the effects of uneven processing can be evaluated. However, the shapes are not limited to cubes.

[0077] The green part 30 further includes a unit structure 34. The unit structure 34 includes a support portion 341 and a pin-shaped portion 342 standing upright from the support portion 341. Thus, the green part 30 includes the pin-shaped portion 342. As described above, the green part 30 is manufactured by stacking powder material and curing the resin. Therefore, if the powder material has an elongated shape like the pin-shaped portion 342, it is difficult to stack the powder material. Even if the powder material can be stacked, it is brittle and easily crumbles, and is therefore easily damaged. Therefore, the pin-shaped portion 342 is included in order to determine and eliminate the first manufacturing condition that causes such a phenomenon.

[0078] The shape of the pin-shaped portion 342 (columnar portion) is not limited, but for example, it can be a columnar shape with a length (width) of, for example, 1 mm to 2 mm in the direction in which the powder material is laid (direction perpendicular to the stacking direction). The column may be a circular cylinder, an elliptical cylinder, or a rectangular cylinder.

[0079] Returning to FIG. 1, the first condition determination unit 101 further determines, as the first manufacturing condition, the manufacturing condition under which the density differences in the unit structures 31, 32, and 33 (FIG. 10) are all within a predetermined range. For example, when the unit structures 31, 32, and 33 are laminated and resin cured, the density of each of the unit structures 31, 32, and 33 decreases below the density when the structure in which the powder material is densely present is taken as 100%. However, the degree of decrease may differ for each of the unit structures 31, 32, and 33 because the unit structures have the same shape but different volumes as described above. Therefore, by determining, as the first manufacturing condition, the manufacturing condition under which the density differences in the unit structures 31, 32, and 33 are all within a predetermined range, it is possible to determine the first manufacturing condition under which the density change is the same regardless of the volume. This makes it possible to suppress the density change regardless of the volume of the first structure and the second structure.

[0080] The predetermined range here is an allowable density difference that is assumed to occur due to lamination and resin hardening. The specific numerical value of the predetermined range can be determined, for example, by experiment. In addition, it is preferable that the density differences of the unit structures 31, 32, and 33 are approximately the same for the reasons described above.

[0081] The first learning unit 102 performs machine learning using the first manufacturing conditions, the shape of the intermediate structure, and the density of the intermediate structure as feature quantities. By providing the first learning unit 102, the evaluation results of the first structure linked to lamination and resin curing can be constructed as a database of learning conditions. The shape here refers to, for example, the dimensions of the sides forming the first structure. The machine learning can perform, for example, supervised learning using the first manufacturing conditions as the objective variable and the shape and density of the first structure as the explanatory variables. However, the machine learning is not limited to supervised learning, and can be either unsupervised learning or reinforcement learning. For example, Bayesian optimization, kernel ridge regression analysis, etc. can be used as the machine learning algorithm.

[0082] The second condition determination unit 103 determines the second manufacturing conditions from the shape and density of the second structure manufactured by debinding and sintering performed during additive manufacturing. The second manufacturing conditions are manufacturing conditions during debinding and sintering for manufacturing a second structure having a desired shape and density using the manufactured desired first structure. The second condition determination unit 103 can determine the manufacturing conditions for manufacturing a second structure (final structure, final product) having a desired shape and density from a first structure having a desired shape and density.

[0083] The second condition determination unit 103, like the first condition determination unit 101, determines the manufacturing conditions when manufacturing a second structure in which the dimensional difference and density difference from the desired shape are both within a predetermined range as the second manufacturing conditions. The shape and density of the actually manufactured second structure are compared with the desired shape and density, and if the difference is within a predetermined range, it can be determined that the shape and density of the manufactured second structure are close to the desired second structure. The difference here is the dimensional difference between corresponding sides in the case of shape, and the density difference in the case of density. Therefore, the second condition determination unit 103 can determine the manufacturing conditions when such a second structure can be manufactured as the second manufacturing conditions when manufacturing a second structure having a desired shape and density.

[0084] The predetermined range here refers to an allowable difference in size and density that is expected to occur due to debinding and sintering. Specific values ​​of the predetermined range can be determined, for example, by experiment.

[0085] Fig. 11 is a perspective view showing the shape of the silver parts 40 used in the evaluation. The silver parts 40 are manufactured as a second structure under, for example, initially set second manufacturing conditions, and the second condition determination unit 103 determines the second manufacturing conditions using the manufacturing conditions, the shape and density of the silver parts 40, etc. Therefore, the silver parts 40 shown in Fig. 11 are structures manufactured for determining the second manufacturing conditions.

[0086] The silver part 40 (second structure) includes a plurality of unit structures 41, 42, 43 that are different in volume but have the same shape. Therefore, the unit structures 41, 42, 43 have a similar relationship. The unit structures 41, 42, 43 each have three types of block shapes, large, medium, and small. Therefore, the above-mentioned green part 30 and silver part 40 have a block shape. The block shape has sides formed by straight lines. The block shape makes it easy to measure the length of each side, and makes it easy to calculate the dimensional difference. In the illustrated example, the unit structures 41, 42, 43 are all cubes. By using a cube, the dimensions in each direction of the x, y, and z axes can be measured, and the shrinkage in each direction due to lamination, etc. can be evaluated. In addition, the effects of temperature distribution due to differences in volume, thermal stress due to temperature distribution, and the effects of uneven processing can be evaluated. However, the shape is not limited to a cube.

[0087] The silver part 40 further includes a unit structure 44. The unit structure 44 includes a main body 441 having an overhanging shape, and a support 442 that supports the lower side (stacking direction) of the main body 441. Therefore, the silver part 40 has an overhanging shape. An overhanging shape is a shape in which there is nothing below in the stacking direction. The bonds of an overhanging shape are weakened by debinding, and the shape is prone to deformation due to its own weight during the process of raising the temperature to the sintering temperature. This can result in a large change in shape. Therefore, by using the unit structure 44, the impact of such changes can be evaluated.

[0088] It is possible that the self-weight deformation of the overhang can be somewhat suppressed by adjusting the heating rate, etc. However, in the example of the present disclosure, it is preferable to consider the case on the premise that the self-weight deformation will occur and to provide supports 442. Therefore, supports 442 are also used in the unit structure 44 used as a prototype.

[0089] Fig. 12A is a top view of a structure 50 that can be used in place of the green part 30 shown in Fig. 10 and the silver part 40 shown in Fig. 11. Fig. 12B is a side view of the structure 50 shown in Fig. 12A. For example, the above description of the green part 30 and the silver part 40 can be similarly applied to the structure 50, except for the matters described below with reference to Figs. 12A and 12B.

[0090] The structure 50 has a cantilever structure. Specifically, the structure 50 includes beams 52 and 53 on two different faces (for example, two opposing faces) 51a of a support 51, which is, for example, a rectangular parallelepiped. The beams 52 and 53 are fixed to the support 51 by being connected to the faces 51a, respectively. Therefore, the beams 52 and 53 are cantilever beams. By having the beams 52 and 53 be cantilever beams, it is possible to evaluate the self-weight deformation of the beams 52 and 53 that occurs during sintering, for example. In the example of the present disclosure, the support 51 and the beams 52 and 53 are integrally molded as a single unit.

[0091] The beams 52, 53 have the same thickness in side view (FIG. 12B) and the same width in top view (FIG. 12A). The beams 52, 53 are disposed at the same height (distance from the lower end surface of the support 51) in side view (FIG. 12B). However, the beams 52, 53 have different lengths (lengths from the surface 51a to the respective ends of the beams 52, 53) in, for example, top view (FIG. 12A) and side view (FIG. 12B). Therefore, the beams 52, 53 have different volumes. By having the different volumes of the beams 52, 53, the difference in behavior of self-weight deformation caused by the different volumes can be evaluated.

[0092] Returning to FIG. 1, the second condition determination unit 103 further determines, as the second manufacturing condition, the manufacturing condition under which the density differences in the unit structures 41, 42, and 43 (FIG. 11) are all within a predetermined range. For example, when the unit structures 41, 42, and 43 are stacked and degreased, the density of each of the unit structures 41, 42, and 43 decreases below the density when the structure in which the powder material is densely present is taken as 100%. However, the degree of decrease may differ for each of the unit structures 41, 42, and 43 because the unit structures have the same shape but different volumes as described above. Therefore, by determining, as the second manufacturing condition, the manufacturing condition under which the density differences in the unit structures 41, 42, and 43 are all within a predetermined range, it is possible to determine the second manufacturing condition under which the density change is the same regardless of the volume. This makes it possible to suppress the density change regardless of the volume of the first structure and the second structure.

[0093] The second learning unit 104 performs machine learning using the second manufacturing conditions, the shape of the second structure, and the density of the second structure as feature quantities. By providing the second learning unit 104, the evaluation results of the second structure linked to debinding and sintering can be constructed as a database of learning conditions. The shape here refers to, for example, the dimensions of the sides forming the second structure. The machine learning can perform, for example, supervised learning using the second manufacturing conditions as the objective variable and the shape and density of the final structure as the explanatory variables. However, the machine learning is not limited to supervised learning, and can be either unsupervised learning or reinforcement learning. For example, Bayesian optimization, kernel ridge regression analysis, etc. can be used as the machine learning algorithm.

[0094] The second learning unit 104 performs machine learning by using, as second manufacturing conditions, manufacturing conditions determined by applying a desired shape and density to a thermal deformation analysis model using the thermal properties of the first structure and the shape and density of the second structure. This makes it possible to calculate the shape and density of the second structure using the thermal deformation analysis model and the thermal properties (e.g., linear expansion coefficient, etc.) of the first structure without actually performing debinding and sintering. Then, machine learning can be performed using the calculated shape and density of the second structure, so that learning opportunities in machine learning can be increased.

[0095] Sintering in particular is a time-consuming process. For this reason, it may take a long time to converge if all conditions are evaluated using experimental methods. Therefore, by performing a thermal deformation analysis using the thermal properties of the first structure and the shape and density of the second structure, which are linked to the sintering conditions obtained by initial learning, and fitting them, it is possible to analyze the second manufacturing conditions and increase the learning opportunities.

[0096] The additive manufacturing apparatus 200 performs additive manufacturing by the binder jetting method as described above. Specific details of the additive manufacturing apparatus 200 will be described later with reference to Fig. 16A and subsequent figures. The additive manufacturing apparatus 200 performs additive manufacturing based on the first manufacturing conditions and the second manufacturing conditions determined by the determination apparatus 100, for example.

[0097] Information regarding the shapes and densities of the first and second structures obtained by additive manufacturing, information regarding the thermal property values ​​of the intermediate structure, etc. are input to the determination device 100 via the input device 300, for example, by a user. The input device 300 is, for example, a keyboard, a mouse, etc. The first and second manufacturing conditions determined by the determination device 100 are output to the output device 400 as appropriate. The output device 400 is, for example, a monitor, a display, a printer, etc.

[0098] The determined first and second manufacturing conditions are manufacturing conditions corresponding to, for example, the properties of the powder material, the properties (functions, specifications, etc.) of the additive manufacturing apparatus 200 used, and the shape and density of the final structure that is the desired second structure. Therefore, by performing additive manufacturing using the additive manufacturing apparatus 200 assumed at the time of determining the conditions under the determined first and second manufacturing conditions, a structure having a desired shape and density can be manufactured.

[0099] 13 is a flowchart showing the additive manufacturing method of the present disclosure. The additive manufacturing method of the present disclosure includes a first condition determination step S101, a second condition determination step S102, and an additive manufacturing step S103.

[0100] The first condition determination step S101 is a step of determining the first manufacturing conditions from the shape and density of the first structure manufactured by lamination and resin curing performed during additive manufacturing by the binder jet method. The first manufacturing conditions are manufacturing conditions during lamination and resin curing for manufacturing a first structure having a desired shape and density, as described above. The first condition determination step S101 can be executed by the first condition determination unit 101 (FIG. 1). The second condition determination step S102 is a step of determining the second manufacturing conditions from the shape and density of the second structure manufactured by debinding and sintering performed during additive manufacturing. The second manufacturing conditions are manufacturing conditions during debinding and sintering for manufacturing a second structure having a desired shape and density using the manufactured desired first structure, as described above. The second condition determination step S102 can be executed by the second condition determination unit 103 (FIG. 1). The first learning step by the first learning unit 102 and the second learning step by the second learning unit 104 may be performed as appropriate.

[0101] The additive manufacturing step S103 is a step of manufacturing a structure using the first manufacturing conditions determined in the first condition determination step S101 and the second manufacturing conditions determined in the second condition determination step S102. The structure referred to here is preferably one that has the conditions used when the first manufacturing conditions and the second manufacturing conditions were determined. In addition, it is preferable that the additive manufacturing device 200 (FIG. 1) used for additive manufacturing also has the conditions used when the first manufacturing conditions and the second manufacturing conditions were determined. According to the manufacturing method of the present disclosure, an additive manufactured product (a layered object, a molded object, a structure) having a desired shape and density can be manufactured.

[0102] The determination of the manufacturing conditions and the additive manufacturing may be performed at the same location or at different locations. In the latter case, for example, the manufacturing conditions may be determined at point A, and additive manufacturing may be performed at point B using the determined first and second manufacturing conditions. In this case, it is preferable to determine the manufacturing conditions at point A after obtaining in advance the specifications, characteristics, etc. of the additive manufacturing apparatus 200 to be used at point B. In this way, for example, the first and second manufacturing conditions that match the specifications, characteristics, etc. of the additive manufacturing apparatus 200 installed at the remote point B can be determined by a user at point A and provided to a manufacturer at point B.

[0103] FIG. 14 is a flow chart showing the first stage of the additive manufacturing condition determination method of the present disclosure. The additive manufacturing condition determination method of the present disclosure includes steps S1 to S9. FIG. 14 mainly shows an example of a method for exploring both the lamination and resin curing processes to evaluate the performance of the green part. The performance of the green part can be controlled by a "first control factor" which will be described below with reference to FIG. 14. Therefore, the first manufacturing condition during lamination and resin curing is an appropriately set first control factor. The flow shown in FIG. 14 can be executed by the first condition determination unit 101 and the first learning unit 102 (both of which are shown in FIG. 1).

[0104] The first control factors during lamination include the quality, particle size, particle size distribution, and binder material of the powder material. However, since these are materials used by end users, it is preferable to make these first control factors fixed factors. Other first control factors include the lamination thickness for each layer, the supply amount of powder material (output, time, etc. of the vibration mechanism 11 described later), the powder spreading speed (movement speed of the powder spreading mechanism 4 described later), leveling settings (rotation speed of the roller (one example of the leveling mechanism 12)), the amount of binder applied (inkjet output from the application mechanism 5 described later, inkjet interval, etc.), drying (output, time, etc. of the heater 13 described later), etc. Among them, the quality of the green parts is particularly affected by these first control factors.

[0105] Some of the first control factors are common to the "formation conditions of powder bed 3" described in the above-mentioned determination system 110. Therefore, the "formation conditions of powder bed 3" determined by the above-mentioned determination system 110 may be used as a part of the first control factors. When used, the "formation conditions of powder bed 3" may be used as a fixed factor in the first control factors. Other first control factors may include the supply rate of the powder material, etc.

[0106] The primary control factors during resin curing include the temperature step of the heat treatment (how many steps the temperature is increased), the rate of temperature increase, the treatment temperature, and the holding time.

[0107] In order to optimize both the lamination and resin curing processes, first, in the lamination process, the first control factors are assigned and allocated, and an initial first control factor group (initialized lamination recipe group) is set (step S1). The initial first control factor group is a group of initial conditions for the predetermined first control factors. The initial first control factor group includes, for example, 10 to 20 manufacturing conditions (parameter sets) in total by changing each of the first control factors during the lamination.

[0108] Next, lamination and drying are performed using the initial first control factor group (steps S2 and S3). Furthermore, in the resin curing process, an initial first control factor group (initial setting curing recipe group) is set by allocating and assigning the first control factors (step S4). The initial first control factor group set here also includes, for example, 10 to 20 manufacturing conditions (parameter sets) in total by changing each of the first control factors during the resin curing. The green part is controlled by curing the resin using each initial first control factor group (step S5).

[0109] The total number of first control factors (number of recipes) during lamination and resin hardening is the product of the number of lamination recipes and the number of hardening recipes, and the number of evaluations may be enormous. Here, in the resin hardening process, the processing temperature is roughly determined by the resin material. Therefore, it is preferable to first fix the heating rate and the number of steps, and roughly determine the processing temperature and processing time. The lamination recipe group may be set using an experimental design method, or by determining the range based on other material recipes that use materials similar to the metal material used when determining the conditions.

[0110] In order to evaluate the quality of the green parts, it is preferable to manufacture and evaluate a plurality of green parts using one recipe, as described above. The plurality of green parts used are, for example, the green parts 30 described with reference to FIG.

[0111] The shape, mass, and density of the green part manufactured using the initially set lamination recipe group and curing recipe group are measured (step S6). The density may be calculated instead of measured. For the manufactured green part, manufacturing conditions close to the target shape are extracted as described with reference to FIG. 10 above, and among these recipes, manufacturing conditions with high density and small density difference due to volume difference (large, medium, and small in FIG. 10 above) are determined as first manufacturing conditions.

[0112] If any of the green parts meet the target in the measurement evaluation, the first manufacturing conditions are determined and both the lamination and resin curing processes can be completed (YES in step S7). However, if the target is not met or if there is a need to check for further improvement (NO in step S7), the initial lamination recipe and curing recipe, as well as the evaluation results of the shape and density of the green parts obtained using these recipes, are stored in a database of learning conditions and machine learning can be performed as described above (step S8). This allows an appropriate combination to be derived.

[0113] After the machine learning in step S7, a new stacking recipe and a new curing recipe are proposed by machine learning (step S9). Step S1 and the following steps are performed again according to the newly proposed stacking recipe and the newly proposed curing recipe.

[0114] The above steps S1 to S9 are first condition determination steps for determining the first manufacturing conditions from the shape and density of the first structure manufactured by lamination and resin curing performed during additive manufacturing using the binder jetting method. The first manufacturing conditions are the manufacturing conditions during lamination and resin curing for manufacturing the first structure having the desired shape and density, as described above.

[0115] FIG. 15 is a flow chart showing the latter part of the additive manufacturing condition determination method of the present disclosure. The additive manufacturing condition determination method of the present disclosure includes steps S11 to S19. The flow shown in FIG. 15 is executed following the flow shown in FIG. 14 above. FIG. 15 mainly shows an example of a method for exploring both degreasing and sintering processes to evaluate the performance of the silver parts. The performance of the silver parts can be controlled by a "second control factor" which will be described below with reference to FIG. 15. Therefore, the second manufacturing conditions during degreasing and sintering are the appropriately set second control factors. The flow shown in FIG. 15 can be executed by the second condition determination unit 103 and the second learning unit 104 (both in FIG. 1).

[0116] The second control factors during debinding include the treatment atmosphere, heating rate, treatment temperature, treatment time (holding time), etc. The second control factors during sintering include the treatment atmosphere, temperature step, heating rate, sintering temperature, holding time, etc.

[0117] In order to optimize both the debinding and sintering processes, first, a green part is produced using the first manufacturing conditions determined with reference to FIG. 14 above, and the produced green part is subjected to simultaneous differential thermal analysis (TG / DTA) (step S11). Data analysis of simultaneous differential thermal analysis reveals the temperature range where mass change occurs due to binder loss, and the debinding process temperature (an example of a debinding recipe) can be determined. By maintaining the temperature range where mass change occurs during debinding, it is possible to suppress damage to the green part caused by a temperature that is too high, and insufficient binder removal caused by a temperature that is too low. This allows the binder to be sufficiently removed, and a brown part can be produced.

[0118] The degreasing process time (an example of a degreasing recipe) can be set, for example, with reference to the binder removal time calculated by a predetermined algorithm based on the record of the thermal mass change. The predetermined algorithm can use, for example, a predetermined formula that associates the process temperature and the rate of change of the temperature (an example of a degreasing recipe) during heating.

[0119] In addition, the thermal property values ​​of the produced green parts, such as the linear expansion coefficient, are measured (step S12). The thermal property values ​​here are physical property values ​​that affect at least one of the shape and density of the sintered silver parts due to the heating caused by debinding and sintering.

[0120] Next, the produced green part is degreased (step S13).

[0121] After debinding, a group of initial second control factors (a group of sintering recipes initially set) is set to optimize sintering (step S14). The group of initial second control factors is a group of initial conditions for predetermined second control factors. The group of initial second control factors includes, for example, 10 to 20 manufacturing conditions (parameter sets) by changing each of the second control factors during sintering. Sintering is performed with the initial second control factors as the second manufacturing conditions (step S15). The shape, mass, and density of each silver part obtained by sintering are measured (step S16). The density may be calculated instead of being measured.

[0122] In order to evaluate the quality of the silver parts, it is preferable to manufacture and evaluate a plurality of silver parts according to one recipe, as described above. The plurality of silver parts used are, for example, the silver parts 40 described with reference to FIG.

[0123] The shape, mass, and density of the silver parts manufactured using the initially set debinding recipe group and sintering recipe group are measured (step S16). The density may be calculated instead of measured. For the manufactured silver parts, manufacturing conditions close to the target shape are extracted as described with reference to FIG. 11 above, and among these recipes, manufacturing conditions with high density and small density difference due to volume difference (large, medium, and small in FIG. 11 above) are determined as second manufacturing conditions.

[0124] If any of the silver parts meet the target in the measurement evaluation, the second manufacturing conditions are determined and both the debinding and sintering processes can be completed (YES in step S17). However, if the target is not met or if there is further prospect for improvement (NO in step S17), the initial debinding recipe and sintering recipe, as well as the evaluation results of the shape and density of the silver parts obtained using these, are stored in a database of learning conditions and machine learning is performed as described above (step S18). This makes it possible to derive an appropriate combination. The data used for the machine learning in step S18 may be data obtained by thermal deformation analysis as described above.

[0125] After the machine learning in step S18, new debinding and sintering recipes are proposed by machine learning (step S19). Steps S12 and after are performed again according to the newly proposed debinding and sintering recipes.

[0126] The above steps S11 to S19 are second condition determination steps for determining second manufacturing conditions from the shape and density of the second structure manufactured by debinding and sintering performed during additive manufacturing. The second manufacturing conditions are manufacturing conditions during debinding and sintering for manufacturing a second structure having a desired shape and density using the desired first structure manufactured as described above.

[0127] According to the above-described determination device 100 and determination method, additive manufacturing is performed under the determined first and second manufacturing conditions, thereby making it possible to manufacture a structure (final structure) having a desired shape and density. In particular, additive manufacturing using a binder jetting method involves multiple manufacturing processes and a wide variety of control factors, such as the first control factor and the second control factor. For this reason, it is not easy to determine suitable control factors. In order to determine suitable manufacturing conditions, a large number of molded objects are manufactured and evaluated. For this reason, constructing a recipe, which is the manufacturing conditions for each process, requires huge costs and time.

[0128] In contrast, in the above-mentioned powder bed fusion, there is a method of using machine learning to compare control factors and modeling results, and performing regression analysis to optimize. However, even if this method is applied to the binder jet method, the binder jet method requires processing time to evaluate the recipe through a series of processes of additive manufacturing conditions. For this reason, it still takes a long time to evaluate, and the number of evaluations increases, and the analysis does not converge to the search for appropriate conditions when the results of the previous process are poor. Therefore, according to the determination device 100 and determination method disclosed herein, the manufacturing conditions are determined by dividing them into first and second manufacturing conditions, so that the time and labor required for determination can be reduced and the determination can be made efficiently. In other words, by dividing the search for additive manufacturing conditions in the binder powder method into multiple parts, the efficiency of optimization can be improved.

[0129] FIG. 16A is a schematic diagram of the additive manufacturing apparatus 200 of the present disclosure, viewed from the side. FIG. 16B is a schematic diagram of the additive manufacturing apparatus 200 of the present disclosure, viewed from above. The additive manufacturing apparatus 200 applies a binder to the powder material (powder bed 3) laid in layers and dries the powder material to bind the powder material. The binder is, for example, a solution containing a resin. The binder may contain additives such as a surfactant and a viscosity modifier in addition to a resin and an organic solvent. The powder material is, for example, a powder of a metal material such as hot work tool steel, copper, a titanium alloy, a nickel alloy, an aluminum alloy, a cobalt chromium alloy, or stainless steel, a powder of a resin material such as polyamide, or a ceramic powder. The additive manufacturing apparatus 200 forms a three-dimensional structure of the powder material bound by the binder in the modeling tank 2 by repeating the formation of the powder bed 3, the application of the binder, and the drying in this order.

[0130] The fluidity of the powder material changes depending on the humidity. For this reason, the moisture can be removed and the humidity can be controlled by heating the parts that come into contact with the powder material, such as the powder supply mechanism 10, the leveling mechanism 12, the powder spreading mechanism 4, and the modeling tank 2, which will be described later.

[0131] A space that functions as a collection section 20 is formed around the modeling tank 2. Excess powder falls into the collection section 20 as the powder spreading mechanism 4 spreads the powder. The powder collected in the collection section 20 is reclassified using a sieve and reused for modeling.

[0132] The additive manufacturing apparatus 200 includes a modeling tank 2, a powder-laying mechanism 4, a coating mechanism 5 that applies a binder, a supply mechanism 6 that supplies the binder to the coating mechanism 5, a chamber 7, and a lifting mechanism 8 that raises and lowers the modeling tank 2. The additive manufacturing apparatus 200 further includes a monitoring device 9 that monitors the operation and control of the additive manufacturing apparatus 200. The monitoring device 9 can have the same hardware configuration as the determination system 110 shown in FIG. 8 above, for example. The chamber 7 houses each mechanism of the additive manufacturing apparatus 200 except for the monitoring device 9 and the supply mechanism 6. However, the chamber 7 may also house the monitoring device 9 and the supply mechanism 6.

[0133] The modeling area 21 is composed of a modeling tank 2 and a bottom plate 22. The bottom plate 22 is fixed to a lifting mechanism 8, and the vertical position of the bottom plate 22 changes according to the operation of the lifting mechanism 8.

[0134] The powder-laying mechanism 4, all of which will be described later with reference to Figure 17A etc., comprises a powder supplying mechanism 10, a vibration mechanism 11, a leveling mechanism 12, a heater 13, and a drive mechanism (not shown) that moves the powder-laying mechanism 4. The heater 13 does not necessarily have to be provided.

[0135] The application mechanism 5 includes a binder head 51, which will be described later, and a drive mechanism (not shown) that moves the binder head 51. The binder is supplied from the supply mechanism 6 to the binder head 51 through a pipe (not shown), and the supplied binder is discharged from the binder head 51.

[0136] Figures 17A, 17B, 17C, and 17D are schematic diagrams showing an example of the powder-laying operation. The powder-laying operation is performed in the order of the operations shown in Figures 17A, 17B, 17C, and 17D. The powder-laying operation is a part of the lamination process, and is preferably performed under conditions included in the first manufacturing conditions.

[0137] First, the powder-laying mechanism 4 will be outlined. The powder-laying mechanism 4 vibrates the powder supply mechanism 10 with the vibration mechanism 11 to drop the powder into the modeling tank 2. Furthermore, the powder-laying mechanism 4 moves, and the powder material is laid in the modeling tank 2 via the leveling mechanism 12, forming the powder bed 3. Although not shown, the powder supply mechanism 10 may be provided in a concave shape on the inner bottom of the chamber 7 (FIG. 16A) independent of the powder-laying mechanism 4, with the top open and an opening at the top end. In this case, it is preferable that the powder supply mechanism 10 has a stage that can be moved up and down for placing and supplying the powder material. This stage constitutes the bottom wall of the powder supply mechanism 10, and is provided so as to be able to be raised and lowered at a predetermined pitch by an appropriate lifting mechanism (not shown).

[0138] Next, the powder spreading operation using the powder spreading mechanism 4 will be described. The powder spreading operation is started by moving the powder spreading mechanism 4 above the modeling tank 2, for example, as shown by the thick solid arrow in Figures 17A to 17D. As shown in Figure 17A, the powder spreading mechanism 4 vibrates the powder supply mechanism 10 by the vibration of the vibration mechanism 11 above the modeling tank 2, thereby dropping the powder material to be supplied. As shown in Figure 17B, the powder spreading mechanism 4 drops the powder material while moving, so that the powder material is spread on the modeling tank 2. As shown in Figure 17C, the leveling mechanism 12 acts on the spread powder material by the movement of the powder spreading mechanism 4, and the powder bed 3 is formed by leveling the ground. The leveling mechanism 12 is a blade, roller, etc. that adjusts the height level of the powder bed 3. In the case of a roller, the powder bed 3 can be formed by compressing it by moving the powder spreading mechanism 4 while rotating the roller. As shown in Figure 17D, the excess powder material falls from the modeling tank 2.

[0139] 18A and 18B are schematic diagrams showing an example of a coating operation. The coating operation is a part of the lamination process, and is preferably performed under conditions included in the first manufacturing conditions. As shown in FIG. 18A, the binder is applied to a powder bed 3 formed on a modeling tank 2 in accordance with a two-dimensional planar shape corresponding to one layer of a modeled product (green part). The coating is performed continuously by inkjet from a coating mechanism 5 that moves as indicated by the thick solid arrow in FIG. 18A and FIG. 18B.

[0140] Fig. 19 is a diagram showing an example of the drying operation. The drying operation is a part of the lamination process, and is preferably performed under conditions included in the first manufacturing conditions. The applied binder can be dried by, for example, heating the heater 13 moving in the direction shown by the thick solid arrow and passing it over the powder bed 3. Alternatively, drying may be performed simultaneously with the application of the binder by continuously heating the modeling tank 2 or the bottom surface of the modeling tank 2 with the heater 13.

[0141] 17A to 19 are repeated to form a three-dimensional shape of a model in the modeling tank 2 (FIG. 16A). The modeling tank 2 is then placed in a thermostatic tank (not shown) and held at a temperature range of about 100°C to 300°C for several hours to several tens of hours, whereby the binder hardens. This allows a green part to be taken out of the powder in the modeling tank 2.

[0142] The removed green parts are degreased by being held in a constant temperature bath or heat treatment furnace (neither shown) at a temperature range of about 300°C to 600°C for several hours to several tens of hours. This removes the binder and gives brown parts. Degreasing is preferably carried out under conditions included in the second manufacturing conditions. In the brown parts, the bonds between the powder materials have been weakened by the removal of the binder, so the shape is easily distorted and care must be taken when handling them.

[0143] The brown parts are sintered by being held in a vacuum furnace or a heat treatment furnace in a vacuum atmosphere (neither shown) at a temperature range from below the melting point to approximately 80% of the melting point for several hours to several tens of hours. The powder particles of the brown parts are metallically bonded together by sintering, resulting in a sintered body, a silver part. The sintering is preferably performed under conditions included in the second manufacturing conditions. Note that degreasing and sintering may be performed collectively in the same furnace.

[0144] Brown parts may deform under their own weight when exposed to high temperatures during sintering. Therefore, in the parts with the above-mentioned overhanging shapes, supports or other supports are manufactured at the same time, as described above, to suppress deformation during sintering. It is preferable that the supports are installed with a small gap between them and the molded object during sintering, and a release agent is applied to the interface to prevent deformation contact and bonding during sintering.

[0145] The present disclosure encompasses the following technical ideas.

[0146] (Appendix 1) A first condition determination unit that determines a first manufacturing condition for lamination and resin hardening to manufacture a first structure having a desired shape and density from the shape and density of the first structure manufactured by lamination and resin hardening performed during additive manufacturing using a binder jetting method; A second condition determination unit that determines second manufacturing conditions for the debinding and sintering from the shape and density of the second structure manufactured by the debinding and sintering performed during the additive manufacturing, in order to manufacture the second structure having a desired shape and density using the desired manufactured first structure. An apparatus for determining additive manufacturing conditions. (Appendix 2) The first condition determination unit determines, as the first manufacturing conditions, manufacturing conditions for manufacturing the first structure, in which a dimensional difference and a density difference from a desired shape are both within a predetermined range. 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 3) the first structure includes a plurality of unit structures having different volumes but the same shape, The first condition determination unit further determines, as the first manufacturing condition, a manufacturing condition under which each density difference in the unit structures falls within a predetermined range. 3. An additive manufacturing condition determination device as described in claim 2. (Appendix 4) a first learning unit that performs machine learning using the first manufacturing condition, the shape of the first structure, and the density of the first structure as feature quantities; 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 5) The first structure and the second structure each have a block shape. 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 6) The first structure includes a pin-shaped portion. 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 7) The second condition determination unit determines, as the second manufacturing conditions, manufacturing conditions for manufacturing the second structure, under which a dimensional difference and a density difference from a desired shape are both within a predetermined range. 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 8) the second structure includes a plurality of unit structures having different volumes but the same shape, The second condition determination unit further determines, as the second manufacturing condition, a manufacturing condition under which each density difference in the unit structures falls within a predetermined range. 8. An additive manufacturing condition determination device as described in claim 7. (Appendix 9) a second learning unit that performs machine learning using the second manufacturing conditions, the shape of the second structure, and the density of the second structure as feature quantities; 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 10) The second learning unit performs the machine learning by using manufacturing conditions determined by applying a desired shape and density to a thermal deformation analysis model using the thermal properties of the first structure and the shape and density of the second structure as the second manufacturing conditions. 10. An additive manufacturing condition determination device as described in claim 9. (Appendix 11) The second structure has an overhang shape. 2. An additive manufacturing condition determination device as described in claim 1. (Appendix 12) A first condition determination step of determining a first manufacturing condition for lamination and resin curing for manufacturing a first structure having a desired shape and density from the shape and density of the first structure manufactured by lamination and resin curing performed during additive manufacturing by a binder jetting method; A second condition determination step of determining second manufacturing conditions for the debinding and sintering performed during the additive manufacturing process from the shape and density of the second structure manufactured by the debinding and sintering, in order to manufacture the second structure having a desired shape and density using the desired first structure manufactured; A method for determining additive manufacturing conditions, comprising: (Appendix 13) A first condition determination step of determining a first manufacturing condition for lamination and resin curing for manufacturing a first structure having a desired shape and density from the shape and density of the first structure manufactured by lamination and resin curing performed during additive manufacturing by a binder jetting method; A second condition determination step of determining second manufacturing conditions for the debinding and sintering performed during the additive manufacturing process from the shape and density of the second structure manufactured by the debinding and sintering, in order to manufacture the second structure having a desired shape and density using the desired first structure manufactured; An additive product is manufactured using the first manufacturing conditions determined in the first condition determination step and the second manufacturing conditions determined in the second condition determination step. A method for producing an additive product, comprising: [Explanation of symbols]

[0147] 10 Powder feeding mechanism 100 decision device 11 Vibration mechanism 110 Judgment System 111 Detection unit 112 Judgment section 113 Learning Department 113a Database 12 Ground leveling mechanism 13 Heater 2 Modeling tank 20 Collection Department 200 Additive Manufacturing Equipment 21 Printing area 22 Bottom plate 23 Heating mechanism 24 Pressurizing mechanism 25 columnar body 26 Irradiation mechanism 27 Measuring mechanism 3 powder bed 300 Input Device 35 Top 4 Flour spreading mechanism 400 Output Device S24 step (detection step) S26 step (judgment step) S29 Step (Learning Step)

Claims

1. A detection unit that detects a state of a powder material spread on a powder bed provided in a binder jet additive manufacturing apparatus, the state affecting a shape of a model to be additively manufactured, for each area where the powder material is present; a determination unit that determines whether the powder bed is in a state where a desired object can be additively manufactured based on the state of the powder material for each of the existence regions detected by the detection unit, thereby determining whether the state of the powder bed is good or bad. A determination system comprising:

2. the detection unit detects a state of each of the presence regions in the powder bed by measuring an index value that affects a shape of the object, for the powder material in each of the presence regions in the powder bed; and The determining unit determines that the condition of the powder bed is good when it determines that the index value of the powder material for each presence region in the powder bed is uniform in a predetermined direction that is at least one of a horizontal direction and a height direction of the powder bed. The determination system according to claim 1 .

3. The determination unit determines that the state of the powder material in each of the presence regions in the powder bed is uniform when a difference between a maximum value and a minimum value of the index values ​​in the predetermined direction is within a predetermined range. The determination system according to claim 2 .

4. The index value includes a density of the powder material in the presence region. The determination system according to claim 2 .

5. A learning unit that performs machine learning using the formation conditions of the powder bed and at least one of the index values ​​for each of the existence regions and the quality of the powder bed as features. The determination system according to claim 2 .

6. The conditions for forming the powder bed are: A lamination thickness of each layer composed of the powder material; A supply rate of the powder material to the powder bed; the feed rate of the powder material to the powder bed; The powder spreading speed when the powder material is spread on the powder bed, or Setting a leveling area when laying the powder material on the powder bed; Contains at least one of The determination system according to claim 5 .

7. a heating mechanism for heating an upper surface of the powder bed; a measurement mechanism for measuring a temperature of the upper surface heated by the heating mechanism, The detection unit detects the state of each of the regions in the powder bed by evaluating a temperature drop rate on the upper surface obtained using the measurement mechanism after heating by the heating mechanism. The determination system according to claim 1 .

8. The heating mechanism instantaneously heats the upper surface. The determination system according to claim 7 .

9. a pressing mechanism for pressing the upper surface of the powder bed at a desired position; The detection unit detects the state of each of the regions in the powder bed by evaluating the pressure on the upper surface at the position where the pressure is applied. The determination system according to claim 1 .

10. a vibration mechanism for vibrating at least an upper surface of the powder bed; The detection unit detects a state of each of the regions in the powder bed by evaluating a behavior of the powder material on the upper surface vibrated by the vibration mechanism. The determination system according to claim 1 .

11. The detection unit detects a state of each of a plurality of presence regions in the powder bed, the presence regions having a predetermined range in a height direction of the powder bed, The determination unit determines whether a state of the powder bed is good or bad based on each state in a plurality of presence regions in a height direction of the powder bed detected by the detection unit. The determination system according to claim 1 .

12. The additive manufacturing apparatus includes a powder-laying mechanism that lays the powder material on a table that forms the powder bed; The detection unit detects the state of each of a plurality of presence regions having a predetermined range on the top surface of each layer having a predetermined thickness laid by the powder laying mechanism in the powder bed. The determination system according to claim 11 .

13. an irradiation mechanism for irradiating the powder bed with ultrasonic waves from a side of the powder bed; The detection unit irradiates ultrasonic waves onto the powder bed from a side of the powder bed using the irradiation mechanism, and detects a state of each area of ​​the powder material laid on the powder bed by using at least one of the transmitted and reflected ultrasonic waves. The determination system according to claim 11 .

14. The detection unit detects a state of each of a plurality of existence regions in the powder bed, the existence regions having a predetermined range in a horizontal plane of the powder bed, The determination unit determines whether a state of the powder bed is good or bad based on each state of a plurality of existence regions in a horizontal plane of the powder bed detected by the detection unit. The determination system according to claim 1 .

15. A detection step of detecting a state of a powder material spread on a powder bed provided in a binder jet additive manufacturing apparatus, the state affecting a shape of a model to be additively manufactured, for each area where the powder material is present; and a determining step of determining whether the powder bed is in a state in which a desired object can be additively manufactured based on the state of the powder material for each of the existence regions detected in the detecting step, thereby determining whether the state of the powder bed is good or bad. A determination method comprising:

Images