JP2023081811A - Molding apparatus, molding system and control method - Google Patents

Abstract

To provide an apparatus, system and method allowing for improving the quality of a three-dimensional molded object.SOLUTION: A molding apparatus 10 for molding a three-dimensional molded object includes formation means that forms a layer comprising powder, application means that applies a molding liquid to the layer, acquisition means that acquires surface information of the layer, and control means that controls the operation of at least the formation means or the application means based on the surface information. The acquisition means acquires the surface information after the formation means formed the layer, and the control means changes the operation of the formation means for the next layer depending upon the surface information acquired.SELECTED DRAWING: Figure 3

Description

The present invention relates to a modeling apparatus, a modeling system, and a method for controlling modeling of a three-dimensional object.

As a method for modeling a three-dimensional object, there are a SLS (Selective Laser Sintering) method that selectively irradiates a laser, an EBM (Electron Beam Melting) method that irradiates an electron beam, and a BJ (Binder Jetting) method that applies a modeling liquid. Are known.

For example, when forming a powder layer and applying liquid based on the shape data of a three-dimensional model as a modeling apparatus that performs modeling by the BJ method, when repeating a series of operations, the dry state of the area of the powder layer to which the liquid is applied A device for controlling a series of operations to be performed next has been proposed in response to the above (see, for example, Patent Document 1).

However, since the conventional modeling apparatus controls the next series of operations only based on the drying state of the liquid, there is a problem that the shape and strength of the three-dimensional modeled object vary, and the quality of the three-dimensional modeled object deteriorates.

An object of the present invention is to solve the above-described problems, and to provide an apparatus, system, and method capable of improving the quality of a three-dimensional object.

According to the present invention, a modeling apparatus for modeling a three-dimensional object,
Forming means for forming a layer containing powder;
an application means for applying the modeling liquid to the layer;
Acquisition means for acquiring surface information of the layer;
and control means for controlling operation of at least the forming means or the applying means based on the surface information.

ADVANTAGE OF THE INVENTION According to this invention, the quality of a three-dimensional molded article can be improved.

The figure which showed the structural example of a modeling system. The figure explaining a BJ system. The figure which showed the structural example of a modeling apparatus. The figure which showed an example of the functional structure of a control circuit. The figure which showed an example of the surface information of a powder layer. The figure which showed an example of the picked-up image which has an unevenness, a crack, and an aggregate. FIG. 5 is a diagram showing an example of surface information when scanning speed is changed as an operation for forming a powder layer; The figure which showed the example of an operation|movement at the time of changing the mechanism which forms a powder layer. The figure explaining the cause which controls the operation|movement which apply|coats modeling liquid, and affects the quality of sintering. The figure which showed the example in which sintering did not fully progress. The figure explaining the method of detecting the permeation|permeation range of modeling liquid. The figure explaining the method of detecting the height difference of a powder surface. The figure which showed the place where modeling liquid was apply|coated. The figure which showed an example of the image which image|photographed the shadow by curvature. The figure explaining the method of estimating a density from the height difference of a powder surface. The figure which showed the Example for making it easy to detect abnormality of a powder surface by a powder surface sensor. The figure which showed an example of the picked-up image. The figure which compared the surface information after powder layer formation, and the surface information after modeling liquid application. 4 is a flow chart showing the flow of controlling the operation of the modeling apparatus; A flowchart showing the flow of calibration before modeling. The figure explaining the example of particle size distribution measurement of powder. 4 is a flowchart showing the flow of recoating optimization; 5 is a flow chart showing the flow of calibration of the density of the modeling liquid; 4 is a flowchart showing the flow of recoat adjustment; 4 is a flowchart showing the flow of powder bed analysis; 4 is a flowchart showing the flow of adjusting the density of the modeling liquid; The flowchart which showed the flow of modeling layer analysis. The figure which showed another example of a structure of a modeling system.

Although the present invention will be described below with reference to embodiments, the present invention is not limited to the embodiments described later.

FIG. 1 is a diagram showing a configuration example of a modeling system that models a three-dimensional object. Here, the three-dimensional object is a three-dimensional object (finished product) formed by the modeling system, and the term “modeled object” simply means an assembly of multiple modeling layers in the process of being modeled by the modeling system. . A single layer in which powder is simply spread is a powder layer, and a layer in which a modeling liquid is applied to the powder layer and the powder is combined with the modeling liquid to form a desired shape is a modeling layer. . When simply referred to as a layer, it refers to either the powder layer or the shaped layer or both.

The modeling system includes a modeling device 10 that models a three-dimensional object. The modeling system may be composed only of the modeling apparatus 10, but may include an information processing apparatus 11 that communicates with the modeling apparatus 10 and transmits shape data of a three-dimensional object to be modeled to the modeling apparatus 10. can be done. When the modeling system is composed only of the modeling apparatus 10 , the functions of the information processing apparatus 11 can be incorporated into the modeling apparatus 10 .

The information processing device 11 is a PC (Personal Computer), a tablet terminal, a smart phone, or the like, and is equipped with an application for creating shape data of a three-dimensional object, and transmits the shape data created using the application to the modeling device 10. Send. Note that the information processing apparatus 11 does not implement the above application, stores shape data created by another device, reads out the stored shape data upon receiving an instruction from the user, and transmits the stored shape data to the modeling apparatus 10 . good too. The modeling apparatus 10 and the information processing apparatus 11 may be connected by a cable or the like, or may be wirelessly connected. Further, the modeling apparatus 10 and the information processing apparatus 11 may be connected via a network.

The shape data is created as data for modeling each layer obtained by dividing the three-dimensional object to be modeled by the unit of the layer pitch.

The modeling apparatus 10 is, for example, a BJ-type modeling apparatus. As shown in FIG. 2, the modeling apparatus 10 spreads the powder 20 to a predetermined thickness corresponding to the layer pitch to form a powder layer. Then, based on the shape data received from the information processing device 11, the modeling apparatus 10 applies the modeling liquid 21 by ejecting it from an inkjet print head 22 as a coating unit, and adheres the powder 20 to each other to form a solid. to form a modeling layer. The modeling apparatus 10 repeats the formation of the powder layer and the application of the modeling liquid to form a green body as a modeled object in which particles are adhered to each other by the modeling liquid in layers. The modeling apparatus 10 degreases the green body to remove excess components of the modeling liquid, and sinters the green body to bond the particles together to obtain a desired three-dimensional shaped object.

Powdered metals, ceramics, gypsum and the like can be used as the powder. As the metal, maraging steel, stainless steel, inconel, cobalt chromium, aluminum alloy, titanium alloy and the like can be used. The modeling liquid contains an adhesive that bonds powder particles together. As the adhesive, for example, phenolic resin, furan resin, cement, water glass, or the like can be used.

The modeling apparatus 10 will be described with reference to FIG. 3 . The modeling apparatus 10 includes a powder head 23, a leveling roller 24, a powder feed piston 25, and a build piston 26 as forming means for forming a powder layer, an inkjet print head 22 as applying means, and an acquiring means. and a control circuit 29 as control means.

The forming means forms a layer containing powder. Moreover, an application|coating means applies|coats a modeling liquid to a layer. As an example, the modeling apparatus 10 that operates in the order of forming means and coating means will be described. Also, part of the means included in the modeling apparatus 10 may be included in an apparatus other than the modeling apparatus 10 and configured as a modeling system.

The powder head 23 includes a tank with an open top filled with powder, and moves (scans) in a predetermined direction while the leveling roller 24 rotates to form a layer of powder adjacent to the forming area. , and a powder layer is formed while flattening the surface. The powder feed piston 25 functions as a first elevating means for moving the bottom plate at the bottom of the powder head 23 up and down. The build piston 26 functions as a second elevating means for moving the bottom plate at the bottom of the modeling bath 27 up and down. With the powder head 23 filled with powder, the powder feed piston 25 is raised by one step by the stacking pitch, the build piston 26 is lowered by the stacking pitch by one step, and the leveling roller 24 is formed. is formed by scanning from left to right in FIG. Although the leveling roller 24 has been described as an example of the flattening means, various rollers, plates, etc. other than the leveling roller 24 may be applied.

After one layer of powder layer is formed, the inkjet print head 22 is scanned over the surface of the powder layer (powder surface), and ejects the modeling liquid only to a desired area on the powder surface. The ink jet print head 22 is connected by a tube to an ink tank filled with a modeling liquid, and is installed on two axes extending in the horizontal direction, for example, the x direction, with respect to the surface of the table on which the modeling apparatus 10 is installed. The two axes can be moved in the y-direction. Thus, the inkjet printhead 22 can scan in the x-direction by sliding along the two axes in the x-direction and in the y-direction by moving the two axes in the y-direction. . The ejection location of the modeling liquid is determined based on the shape data received from the information processing device 11 . In this way, a modeling layer is formed by selectively coating the modeling liquid on the powder surface and solidifying only the desired regions. Note that the modeling apparatus 10 is in a temporarily fixed state only by applying the modeling liquid.

The modeling apparatus 10 is provided with an infrared lamp, a heater, or the like as heating means in order to apply heat to a desired region coated with the modeling liquid to dry it or promote adhesion. By repeating the formation of the powder layer and the application of the modeling liquid, the modeling apparatus 10 can model a three-dimensional object having a desired three-dimensional shape as a modeled object in which a plurality of modeling layers are laminated.

In the BJ type molding machine, the density of the green body varies due to variations in the characteristics due to differences in the composition of the powder, and variations in the forming and coating methods. There was a problem that deposits were formed on the surface. To solve this problem, it is necessary to form a uniform and dense green body. Therefore, as a result of intensive studies, the inventor found that it is possible to form a uniform and dense green body by controlling at least the operation of the forming means or the coating means based on the surface information acquired by the acquiring means. Found it. As a result, it is possible to improve the quality of the three-dimensional object because it is not affected by the difference in the properties of the powder and the variations in the forming means and coating means.

Therefore, the modeling apparatus 10 is provided with the powder surface sensor 28 as acquisition means for acquiring surface information of the layer. The powder surface sensor 28 is, for example, a camera that photographs the powder surface, is installed at a position facing the powder surface with a gap therebetween, and obtains surface information by photographing the powder surface. The surface information of the layer may be surface information of the powder layer, surface information of the modeling layer, or surface information of both. The powder surface sensor 28 detects not the entire powder surface of the powder spread in the modeling tank 27, but the molding area on the molding table because the object is molded on the molding table installed at the bottom of the molding tank 27. It is sufficient to be able to photograph the powder surface of

As another example of the powder surface sensor 28, a camera for photographing the powder surface may be provided so as to be in contact with the inkjet printhead 22 so that scanning can be performed integrally with the inkjet printhead. Since it is possible to shoot close to the powder surface, even the details of the powder surface can be magnified and photographed. There are advantages. On the other hand, there is also a drawback that it is difficult to photograph the entire powder surface.

Further, as another example of the powder surface sensor 28 as an acquisition means, a sensor (optical section type 3D scanner) that irradiates a laser and acquires depth information from the position of the bright line, a sensor that irradiates an infrared ray and detects changes in reflected light. An acquiring sensor or the like may be applied.

After the applying means applies the modeling liquid, the surface information obtained by applying the modeling liquid to the layer containing the powder may be acquired, or before the applying means applies the modeling liquid, the surface information including the powder may be acquired. can be obtained, or a combination thereof. By obtaining the surface state before application, the powder filling state by recoating can be obtained. As a result, recoating parameters such as the amount of powder supplied, the scanning speed of the recoater, and the number of revolutions can be adjusted. Means for acquiring the surface state on which the modeling liquid is applied can be the same as the powder surface sensor 28 for acquiring surface information of the modeling layer, or another means can be used.

The modeling apparatus 10 has a control circuit as control means for controlling the operations of the leveling roller 24, the powder feed piston 25, the build piston 26, and the inkjet print head 22 based on the surface information acquired by the powder surface sensor 28. 29. Here, the modeling apparatus 10 has the control circuit 29, but the information processing apparatus 11 may have the function as the control means.

The control means analyzes the image as the surface information acquired by the acquisition means, and controls at least one of the forming means, coating means, and heating means. As for the analysis of the image, extraction of feature amounts and the like are performed by texture analysis. The powder filling density and ink excess/deficiency are determined from the feature values. Alternatively, the area or the like is determined by analyzing the difference in brightness, and the variation in the powder or the size of the shadow of the ink is determined. The control means changes the speed of the forming means, the density of the coating means, the amount of powder supplied, and the heating time of the heating means according to the analysis results of these images. These means may be changed singly or in combination.

Control based on surface information may be performed on the next layer, or may not be performed on all layers according to the surface information. In other words, the necessity of control may be determined according to the surface information, and control may not be performed on the next layer. Further, the surface information of the n layer may be used for the control of the n+1 layer, or the surface information of the first layer to the n−1 layer before the n layer may be accumulated and reflected in the control of the n+1 layer. may All the surface information of the 1st layer to the n-1th layer prior to the n layer may be reflected in the control of the operation of the n+1 layer. Alternatively, surface information similar to the n-th layer is detected from the surface information of the first layer to the n−1 layer, and the same control as the control performed from the similar surface information is reflected in the n+1 layer. may Control may be determined in consideration of modeling data as well as approximate surface information.

The control circuit 29 includes a storage device that stores a program for controlling the operation of the leveling roller 24 and the like, a CPU (Central Processing Unit) that executes the program, and a RAM (Random Access Memory), an input/output I/F for controlling input of surface information from the powder surface sensor 28, and output of control information to the leveling roller 24 and the like.

Control circuit 29 may implement a machine learning program. The control circuit 29 learns the correspondence between the modeling data, its surface information, and the optimum control of the modeling apparatus, so that the similar modeling data can be optimally controlled in advance. The machine learning program can cause the control circuit 29 to function as data storage means 30, learning means 31, and estimation means 32, as shown in FIG. The data storage means 30 receives modeling data, surface information, and operation information of the modeling apparatus, and stores them as data. The learning means 31 extracts feature amounts from the data accumulated in the data accumulation means 30 and learns their relationship. Based on the results learned by the learning means 31, the estimating means 32 estimates the surface state and suitable operation of the molding apparatus from arbitrary input molding data, and the control means controls the operation of the molding apparatus based on this estimation. Control.

Next, control by the control circuit 29 will be described. The control circuit 29 controls the operation of the leveling rollers 24 and the like based on the surface information acquired by the powder surface sensor 28 . Therefore, what kind of control is performed in the case of what kind of surface information will be described in detail below.

FIG. 5 is a diagram showing an example of surface information acquired by the powder surface sensor 28. As shown in FIG. FIG. 5 is a diagram showing that there is an abnormality in the powder surface. The unevenness is caused, for example, by a mixture of portions where the layer is thick and portions where the layer is thin. Abnormalities include unevenness, cracks, and the like. Such an abnormality can be detected by photographing the powder surface 33 from above with a camera as the powder surface sensor 28 and analyzing the photographed image.

When the powder surface 33 is normal, the pixel values of the pixels forming the captured image are substantially constant. However, if there is an abnormality in the powder surface 33, color shading will occur, and pixel values will differ by a certain amount or more between the abnormal portion and the non-abnormal portion. Therefore, by setting a certain range for pixel values and determining whether each pixel value is within the certain range, it is possible to identify the presence or absence of an abnormality, and if there is an abnormality, which part has the abnormality. . Note that this method is only an example, and other methods may be used to identify the presence or absence of an abnormality and the location of the abnormality.

When the powder surface 33 is photographed at the stage where the powder is discharged from the powder head 23 and the powder layer is formed, and it is detected that the powder surface 33 has an abnormality such as unevenness or cracks, further powder layer formation is performed. Continuing the formation does not eliminate unevenness and cracks. If unevenness and cracks remain, problems arise in that the shape of the three-dimensional modeled object changes, or that the strength is low and the product is fragile. Therefore, the control circuit 29 can be controlled to stop the operation of the forming means for forming the powder layer, such as the powder head 23 . Such control can prevent waste of powder and modeling liquid.

Moreover, as shown in FIG. 6, when unevenness, cracks, and aggregates 34 are small, the influence on the shape and strength of the three-dimensional object is small. However, if even small abnormalities are accumulated, there is a possibility that the strength of the entire model will be insufficient. Therefore, by recording even small abnormalities that can be detected, it can be used as data for guaranteeing the quality of the modeled object. Therefore, the control circuit 29 can be controlled so as to stop the operation of the mechanism for forming the powder layer when the scale of the unevenness or cracks is greater than or equal to a certain scale. The scale of unevenness and cracks can be determined from the area of the uneven portion, the width of the crack, the length of the crack, and the like. Such unevenness and cracks can be dealt with quickly when an abnormality occurs by taking an image of each layer and performing detection processing.

FIG. 7 is a diagram showing an example of surface information when the scanning speed is changed as the operation of forming the powder layer. One of the operations for forming the powder layer is the scanning of the leveling rollers 24 . Changing the scanning speed of the leveling rollers 24 changes the packing density of the powder. Therefore, the control means may control at least the formation means or the application means based on the powder packing density obtained from the surface information. FIG. 7(a) shows an image of the powder surface taken as surface information when the scanning speed of the leveling roller 24 is low, and FIG. 7(b) is an image obtained when the scanning speed of the leveling roller 24 is high. An image of a powder surface is shown as surface information. Scan speed is also called recoat speed.

When the scanning speed shown in FIG. 7(a) is low, the particles of the powder are uniformly arranged. On the other hand, when the scanning speed shown in FIG. 7B is high, the powder particles are arranged irregularly.

In general, it is desirable to be able to spread the powder at a high speed because the amount of modeling per unit time can be increased and the productivity is improved. However, it is known that when the powder is moved at high speed, the particles of the powder agglomerate, making it difficult to disperse and distribute them uniformly (for example, “Effect of Recycling on the Spreadability of Metallic Powder”, June 3, 2020, AZO MATERIALS, <URL: https://www.azom.com/article.aspx?ArticleID=19313>).

Apart from the scanning speed of the leveling roller 24, the state of the powder in the powder layer also changes depending on the rotation speed of the leveling roller 24. FIG. When the rotation speed is high, the force to roll up the powder becomes stronger and the force to push the powder becomes weaker, so that the packing density of the powder becomes lower. On the other hand, when the rotation speed is slow, the force to roll up the powder becomes weaker and the force to push the powder becomes stronger, so that the packing density of the powder becomes higher. The higher the packing density of the powder, the higher the density and strength of the molded object after condensation, so it is generally desirable. It will be different depending on On the other hand, the powder is evenly filled by increasing the rotation speed and increasing the power to roll up the powder. It is desirable to set an appropriate rotation speed in consideration of density and uniformity.

The speed at which the powder can be uniformly layered depends on whether the powder is, for example, a new material or a recycled material, and also depends on the composition and properties of the powder. The properties are, for example, the particle size and shape of the powder.

The control means may vary the amount of powder supplied by the forming means. Alternatively, the control means may change at least one of the rotation speed or scanning speed of the leveling rollers as a change in the operation of the forming means. For example, as shown in FIG. 7, by judging the uniformity of the arrangement of particles on the powder surface from an image taken by a camera, it is possible to determine whether or not the powder is evenly layered. If this determination reveals non-uniformity, the control means adjusts at least one of the rotational speed or scanning speed of the leveling rollers 24 . Alternatively, the control means adjusts the powder supply amount. This enables uniform lamination. When the leveling roller 24 rotates or scans at a set value that is too high or too low due to variations in the characteristics of the powder material being used, the powder surface becomes uneven. The rotation speed and scanning speed of the roller 24 can be changed. If the powder supply amount is small and uneven, the control circuit 29 can increase the powder supply amount by, for example, raising the bottom plate of the powder head 23 by means of the powder feed piston 25 . Such surface information is acquired, and the scanning speed of the leveling rollers 24 and the rotation speed of the rollers are adjusted at a frequency of about once every several tens of layers before or during modeling.

The amount of powder supplied, the rotation speed of the leveling roller 24, and the scanning speed may be changed in their entirety or only in part.

In FIG. 2, a powder feed piston 25 and a build piston 26 are used as forming means for forming a powder layer to form a step, and a leveling roller 24 conveys the powder on the step to a modeling tank 27 to flatten it. to form a powder layer. The forming means is not limited to such a system, and may include supply means for supplying powder and vibration imparting means for imparting vibration to the supply means. FIG. 8 shows a hopper feeder 40 as feeding means. The powder may be dropped from the hopper feeder 40 into the modeling tank 27, and the surface thereof may be smoothed by the roller 41 to form a powder layer. A predetermined amount of the powder can be caused to drop downward by applying vibration to the hopper feeder 40 by means of vibration applying means. The method shown in FIG. 8 is called a hopper method.

With reference to FIG. 8, the operation of forming a model when the mechanism for forming the powder layer is replaced with a hopper system will be described. In the hopper method, the forming means of the modeling apparatus 10 includes a hopper feeder 40 as feeding means and rollers 41 as flattening means. The application means of the modeling apparatus 10 is a print head 42, and in addition, an infrared lamp 43 as a heating means is included. The roller 41 is provided, for example, at the lower end of the hopper feeder 40 and behind the openable and closable hole 40a provided at the bottom of the hopper feeder 40 in the scanning direction, and smoothes the powder that has fallen from the hole 40a. The infrared lamp 43 is provided, for example, in front of the hole 40a in the scanning direction. Hopper feeder 40 and printhead 42 are slidably mounted on a shaft extending in the x-direction and the shaft is movable in the y-direction. Thus, it is possible to scan in the x direction by sliding the hopper feeder 40 and print head 42 and to scan in the y direction by moving the axis in the y direction. Here, the axis extends in the x direction, but this is not a limitation, and the axis may extend in the y direction and be movable in the x direction.

As shown in FIG. 8(a), the bottom plate at the bottom of the modeling tank 44 can be moved up and down by the build piston 45, and the height of the bottom plate is adjusted so that one layer of powder can be accommodated. be. As a result, it becomes possible to form a powder layer for one layer (one stage) by spreading the powder up to the height of the top of the modeling tank 44 . At this time, the height position of the surface of the powder is one step below the top of the modeling tank 44 . The hopper feeder 40 has a hole 40a at the bottom thereof, and the powder contained therein drops when vibration is applied by the vibration applying means. By scanning in the direction indicated by the arrow, the entire surface of the powder layered up to that point can be filled with the powder forming the next powder layer.

Merely dropping the powder from the hopper feeder 40 causes unevenness on the surface, and the powder is not evenly spread. Therefore, as shown in FIG. 8B, by rotating and scanning the roller 41, the surface is flattened and the powder is evenly spread. Excess powder falls into a powder recovery groove 46 provided around the modeling tank 44 and is recovered.

After uniformly spreading the powder to form a powder layer, as shown in FIG. 8C, the printing head 42 discharges the modeling liquid to apply the modeling liquid to a predetermined area of the powder surface. The area to which the modeling liquid is applied is determined based on the shape data received from the information processing device 11 . A modeling layer is formed by applying the modeling liquid, and a green body 47 is formed as a modeled object in which the modeling layers are laminated as shown in FIG. 8(d). In the example shown in FIG. 8(d), the green body 47 is a cup with a handle.

Thereafter, as shown in FIG. 8(e), the powder surface is heated by an infrared lamp 43 to dry the applied modeling liquid or accelerate the chemical reaction.

Even in the hopper-type modeling apparatus 10, even if the roller 41 is used to level the powder, the powder may not be filled uniformly when the scanning speed is high. Therefore, by changing the operation of the hopper feeder 40 based on the surface information acquired by the powder surface sensor 28, the powder can be uniformly layered. Specifically, the control circuit changes the amount of powder to be dropped as the supply amount of powder. Alternatively, it is necessary to correct the fact that the amount of powder is uneven depending on the location. Therefore, the control means changes at least the vibration frequency or the vibration time of the vibration imparted by the vibration imparting means.

So far, we have discussed controlling the operation of forming a powder bed. A modeled object is modeled by applying a modeling liquid and bonding particles of the powder through the progress of chemical reactions such as drying and adhesion. For this reason, it is conceivable that the quality of sintering is affected not only by the formation of the powder layer, but also by the application of the modeling liquid, heating, and the like. Therefore, the control means may control the heating time or heating temperature by the heating means according to the surface information.

FIG. 9 is a diagram for explaining the factors that control the operation of applying the modeling liquid and affect the quality of sintering. Gaps 51 are formed between the particles of the powder 50 in terms of particle size, no matter how densely the powder layer is laid. Therefore, when the modeling liquid 52 is applied to the powder layer, the modeling liquid 52 penetrates into the gaps 51 . The modeling liquid 52 permeated into the gap 51 is interposed between the particles to bond the particles together, and the particles are attracted to each other by the capillary force of the modeling liquid 52, thereby improving the density of the powder. After that, when heated, sintering progresses from the point where the particles are in close contact with each other, densification progresses, and the powder is united to form a modeled object.

In general, when the density of the green body 47 before sintering is low, there is a problem that there are many defects even if sintered, or sintering does not progress in the first place. To obtain a high-quality sintered body, it is desirable to increase the density of the green body 47 as much as possible.

The density of the green bodies 47 can be improved by the capillary force of the modeling liquid 52 . However, when the amount of the modeling liquid 52 is large, the capillary force does not act and the density is not improved. On the other hand, if the amount of the modeling liquid 52 is too small, the strength of the green body 47 will be low and the green body 47 may be broken due to the weak adhesion between the particles. Therefore, in order to improve the density of the green body 47, it is necessary to appropriately control the amount of the modeling liquid 52. FIG. Therefore, based on the surface information, the control means changes at least the amount of the modeling liquid or corrects the region to which the modeling liquid is applied, as a change in the operation of the applying means.

Where the powder layer is dense and where it is sparse can be determined by obtaining surface information from the powder surface sensor 28 and analyzing it. The density of the green body 47 can be improved by reducing the amount of the modeling liquid 52 in the dense locations and increasing the amount of the modeling liquid 52 in the sparse locations and changing the amount depending on the location. Therefore, the control circuit 29 can improve the density of the green body 47 by correcting the amount of the modeling liquid 52 and the area to be applied based on the surface information.

FIG. 10 is a diagram showing an example in which sintering did not proceed sufficiently. This is an example in which sintering proceeded only partially due to variations in the density of the green body 47 . In FIG. 10, the black portion is the portion 53 where the sintering did not proceed and the powder remained.

The control means may control the applying means based on at least the height difference of the surface or the permeation range of the modeling liquid as the surface information. As an example of the permeation range, FIG. 11 is a diagram illustrating a method of detecting the permeation range of the modeling liquid from the surface information. FIG. 11(a) is an image taken by a camera of a powder surface on which the modeling liquid is applied with intervals of one dot each. FIG. 11(b) is an image of the powder surface photographed with a camera after applying the modeling liquid and removing the powder from the unapplied area.

From the image shown in FIG. 11A, it is possible to detect the range 54 in which the modeling liquid permeates based on the shade of color. From the image shown in FIG. 11(b), it is possible to detect a range 54 in which the modeling liquid permeates depending on whether the image is in focus. By the way, in the image shown in FIG. 11(a), the dark-colored portions shown as dots at substantially equal intervals indicate the range 54 in which the modeling liquid permeates, and in the image shown in FIG. 11(b), the focused A portion shown in FIG.

When the modeling liquid is applied, the droplets of the modeling liquid that land on the powder surface spread and permeate with the passage of time. In order to perform stable modeling, it is necessary to apply an appropriate amount of modeling liquid, and for that reason, it is important how far the modeling liquid penetrates when it is applied. However, the degree of permeation is affected by variations in the composition of the powder and variations in the properties of the modeling liquid, so the permeation range may be wider or narrower than the presumed permeation range.

The powder surface sensor 28 can photograph the powder surface into which the modeling liquid permeates, and acquire the permeation range of the modeling liquid as surface information. The control circuit 29 adjusts the application amount so as to obtain an appropriate amount of modeling liquid based on the obtained permeation range. If the permeation range of the modeling liquid is wide, it can be adjusted so that the modeling liquid is applied to a range of a shape that is one size smaller than the target shape. Also, considering the influence of penetration, reduce the amount of modeling liquid for areas that are easy to penetrate, and increase the amount of modeling liquid for areas that are difficult to penetrate. can be changed.

As an example of the height difference, FIG. 12 is a diagram illustrating a method of detecting the height difference of the powder surface. FIG. 12(a) is a diagram showing the result of measuring the height difference of the powder surface on which the modeling liquid is not applied, using a three-dimensional measuring device. FIG. 12(b) is a diagram showing the result of measuring the height difference of the powder surface coated with the modeling liquid with a three-dimensional measuring device. In this way, the modeled object warps due to the time difference in contraction during modeling. Here, the height difference of the powder surface is measured by a three-dimensional measuring device, but the height difference of the powder surface can be obtained from the image taken by the camera as the powder surface sensor 28 .

This will be described with reference to FIGS. 13(a) and 13(b). FIG. 13(a) shows the case where the modeled article does not warp, and FIG. 13(b) shows the case where the modeled article warps. 55 shows the area|region which applied the modeling liquid among the modeling tanks 27. FIG. 56 is a light source.

In the image captured by the camera as the powder surface sensor 28, no shadow appears when no warp occurs as shown in FIG. 13(a). On the other hand, when warp occurs as shown in FIG. The size of the height difference is proportional to the size of the shadow, and since the control circuit 29 knows what kind of shape has been formed in advance, the size of the warp is estimated based on the size of the shadow 57. can do. FIG. 14 is an example of an image of a shadow 57 caused by warping as shown in FIG. 13(b).

If the size of the warp can be estimated from the size of the shadow in this way, the temperature of heating by the infrared lamp and the waiting time for heating can be adjusted. Since the modeled object shrinks over time due to the application of the modeling liquid, warping occurs due to the difference in the degree of shrinkage that accompanies the progress of modeling. If the deformation has progressed sufficiently for each layer, it will not deform further and no warping will occur. The contractile force accumulates, and at a certain point, it can no longer withstand the support around it, and it deforms greatly. Therefore, by estimating the magnitude of warpage based on the size of the shadow for each layer, it is possible to determine whether a sufficiently small deformation has been completed for each layer. If a sufficiently small deformation is not completed, the temperature of the heating by the infrared lamp is raised, or the heating standby time is increased so that the small deformation is completed, and large warping can be prevented.

The powder surface to which the modeling liquid is not applied maintains a flat state as shown in FIG. On the other hand, as shown in FIG. 12B, the surface of the powder coated with the modeling liquid becomes lower due to the application of the modeling liquid. It is thought that this is because the particles are attracted to each other and contract under the influence of capillary force generated by the permeation of the modeling liquid between the particles, which causes dents on the powder surface and lowers the powder surface. Therefore, by detecting the height difference of the powder surface, it is possible to detect the degree to which the modeling liquid permeates and the density is improved.

It takes time for the modeling liquid to permeate, dry, and solidify. After forming the powder layer and applying the modeling liquid, providing a waiting time is effective for stabilizing the modeling. This is because the particles of the powder are sufficiently attracted to each other by capillary force, and the density is improved. However, if the standby time is provided, the time required to form one layer is increased by that amount. The control means changes the waiting time according to the surface information when determining that waiting is necessary based on the acquired surface information.

The powder surface sensor 28 acquires the state in which the depression as surface information progresses as a height difference. The control circuit 29 adjusts the necessity of waiting and the length of the waiting time based on the height difference acquired by the powder surface sensor 28 .

The state in which the dents progress can be monitored by observing the height difference, in which the modeling liquid permeates and the particles are attracted to each other and contracted due to capillary force. For this reason, the molding liquid is applied in advance several times, and the degree of shrinkage of the powder is grasped by observing the height difference of the powder surface. By adjusting the number of times of application according to the grasped progress of shrinkage, it is possible to adjust the amount of the modeling liquid to an optimum amount.

The control means estimates the filling density from at least surface unevenness or shadow as the surface information. As an example of unevenness of the surface information, FIG. 15 is a diagram illustrating a method of estimating the density from the height difference of the powder surface. The density of the powder bed depends on how tightly the particles of powder 50 are packed. The degree of particle clogging depends on the roughness of the powder surface (surface roughness), and the coarser the powder surface, the lower the density tends to be. Therefore, the powder surface sensor 28 can acquire the roughness of the powder surface as surface information according to whether the powder surface has many irregularities and whether the height difference is large. The control circuit 29 can estimate the density from the roughness of the powder surface acquired by the powder surface sensor 28 .

To estimate the density, the density corresponding to the roughness is measured in advance according to the composition of the powder, etc., and accumulated as data, and the accumulated data is used based on the detected roughness of the powder surface. can do. Since this method is an example, it is not limited to this.

The density of the powder layer can be estimated not only from the roughness of the powder surface but also from the shade of the powder surface. As an example of the shadow of the surface information, a method of estimating the density from the shadow of the powder surface will be described with reference to FIGS. 13(b) and 14. FIG. As shown in FIG. 13(b), by applying light at an angle (shallow angle) that is nearly parallel to the powder surface, the dark-colored portion that becomes the shadow 57 stands out. After making the shadow 57 stand out in this way, the image is captured by the camera as the powder surface sensor 28, and the distribution (texture) of the shadow 57 as shown in FIG. 14 is obtained. The control circuit 29 estimates the density based on the obtained shadow 57 distribution.

FIG. 16 shows an embodiment for facilitating detection of an abnormality in the powder surface by the powder surface sensor 28. In FIG. By illuminating the modeling tank 27 with the light source 56 at a shallow angle and photographing it with the powder surface sensor 28, it becomes easy to detect the abnormality of the powder surface. An example of a captured image is shown in FIG. If the powder surface is substantially flat, the photographed image will be an image with uniform brightness. If there are cracks on the powder surface or if there are agglomerates, they will be shadowed and can be easily detected.

When the density of the powder layer is low, the particles of the powder are sparse and there are many gaps, and the surface of the powder is also uneven. Therefore, when light is applied at a shallow angle, the shadow 57 is clearly visible. On the other hand, when the density of the powder layer is high, the particles of the powder are densely packed and the irregularities of the powder surface are small. Therefore, when light is applied at a shallow angle, fine shadows 57 appear scattered.

Thus, the density can be estimated from the difference in appearance of the shadow 57 . With this method as well, it is possible to measure the density according to the shadow distribution in advance, store it as data, and use the stored data to estimate the density based on the detected shadow distribution.

FIG. 18 is a diagram illustrating surface information after forming a powder layer and surface information after applying a modeling liquid. FIG. 18(a) is an image captured by a camera as the powder surface sensor 28 as surface information after forming the powder layer. FIG. 18(b) is an image captured by a camera as the powder surface sensor 28 as surface information after applying the modeling liquid. The dark-colored area on the right side is the area 55 to which the modeling liquid is applied.

The timing of acquiring the surface information by the camera as the powder surface sensor 28 is after forming the powder layer before applying the modeling liquid, and after forming the powder layer, and modeling is performed on a predetermined area of the formed powder layer. After applying the liquid. In the case of estimating the packing density of powder from the irregularities and shadows on the surface described with reference to FIGS. to get as 11 and 12, when detecting the permeation range of the modeling liquid and changes in surface unevenness due to the application of the modeling liquid, the powder surface after applying the modeling liquid is photographed with a camera, and the obtained This image is obtained as surface information.

With reference to FIG. 19, the method of controlling modeling of a three-dimensional molded article is summarized. The control of the control circuit 29 starts from step 100 when the shape data is transmitted to the molding device 10 and the molding of the three-dimensional object is instructed. In step 101, modeling-related parameters are calibrated prior to modeling. In step 102, the control circuit 29 instructs the powder head 23, the leveling roller 24, the powder feed piston 25, and the build piston 26 as forming means to form a powder layer. The head 23 and the like form one powder layer. At this time, the leveling rollers 24 and the like operate at the set rotational speed and scanning speed.

At step 103, it is determined whether recoating adjustment (adjustment of the rotational speed and scanning speed of the leveling roller 24) is to be performed. For example, recoating adjustment may be performed every several tens of layers. Therefore, when the specified number of layers are formed, the process proceeds to step 104 to perform recoating adjustment.

In step 105, the powder surface sensor 28 as acquisition means acquires surface information of the formed powder layer. Powder layer analysis is performed based on the acquired surface information. As a result, for example, if it is found that the powder is insufficiently supplied, the amount of powder supplied is increased, and if any other abnormality is found, the data of the occurrence of the abnormality is saved as a history.

In step 106, for example, when the result of step 105 shows that the occurrence of abnormality exceeds the allowable range, it is determined to stop modeling. Whether or not to stop modeling may be determined, for example, by the number of occurrences of anomalies or by the size of an anomaly.

In step 107, the control circuit 29 instructs the inkjet print head 22 as the coating means to apply the modeling liquid based on the data indicating the corresponding modeling layer in the shape data, and the inkjet print head 22 A modeling liquid is applied to a desired region of the powder surface of the formed powder layer. When forming the second-layer modeling tank, the data indicating the second-layer modeling tank is used as the data indicating the corresponding modeling tank. As an example, a method in which the forming means and the applying means operate in this order has been described, but the applying means and the forming means may operate in this order, or a predetermined process may be performed between the forming means and the applying means.

At step 108, it is determined whether or not to adjust the density of the modeling liquid. For example, the density of the modeling liquid may be adjusted every several tens of layers. Therefore, when the designated number of layers are formed, the process proceeds to step 109 to adjust the density of the modeling liquid.

At step 110, a build layer analysis is performed. The powder surface sensor 28 acquires surface information of the powder surface to which the modeling liquid is applied, and adjusts the infrared heating temperature and standby time based on the acquired surface information.

In step 111, heating is performed by an infrared lamp as a heating means. In addition, the control circuit 29 can change the setting values of the heating time and the heating temperature by the infrared lamp based on the acquired surface information, thereby making it possible to increase the density and improve the quality. Therefore, the control circuit 29 can also control the operation of the infrared lamp.

At step 112, the control circuit 29 refers to the shape data and determines whether the last modeling layer has been formed. If there is a build layer to form, return to step 102 to form the next build layer. On the other hand, if the final modeling layer has been formed, the process proceeds to step 113 and the control ends.

In step 106, if there is unevenness or cracks in the powder layer, it is determined that the operation of forming the powder layer should be stopped. End control.

Next, with reference to FIG. 20, calibration before modeling will be described. When calibration before modeling is started at step 120 , powder particle size distribution is measured at step 121 , recoating optimization is performed at step 122 , and modeling liquid concentration is calibrated at step 123 .

An example of powder particle size distribution measurement will be described with reference to FIG. In step 131, a small amount of powder is sprinkled on a predetermined place. In step 132, this is photographed by the powder surface sensor 28. FIG. In step 133, the diameter of each powder particle reflected in the image is calculated. At step 134, the distribution of diameters of the particles in the image is determined. A plurality of images may be taken, or only one image may be used. The distribution of powder particles is an important characteristic of powder, and at the same time recording this, it is possible to adjust the mixing ratio of recycled powder, the scanning speed and rotation speed of the leveling roller, and the density of the modeling liquid. do.

The recoating optimization procedure will be described with reference to FIG. In step 141, the control circuit 29 instructs the powder head 23, the leveling roller 24, the powder feed piston 25, and the build piston 26 as forming means to form a powder layer. 23 and the like form a plurality of powder layers. At step 142, the powder surface sensor 28 takes an image of the powder surface.

Texture analysis is performed in step 143 . In texture analysis, for example, co-occurrence matrices are calculated. For each pixel in the acquired image, this is a calculation of what kind of luminance is highly probable for each distance and angle of adjacent pixels, and different textures indicate different values. Other known texture analyses, such as those using higher-order autocorrelation, those using Fourier analysis, and those using fractal structures. Based on these multiple analysis results, the packing density of the powder is estimated in step 144 . That is, for powder surfaces with different packing densities, the correspondence between an image of the powder surface and various texture analysis results is checked in advance, and the powder packing density is estimated based on these relationships.

At step 145, the estimate of the flour packing density is checked to see if it is acceptable. If the powder packing density is within the acceptable range, proceed to step 148 and exit. If the powder packing density is lower than the acceptable range, step 146 reduces the scan speed or rotation speed of the leveling rollers to increase the packing density. After that, the process returns to step 141 to form the powder layer again. On the other hand, if the estimated powder packing density is higher than the acceptable range, step 147 increases the scanning speed or rotation speed of the leveling rollers to increase the packing density. In general, the higher the packing density of the powder, the better, but if the packing density does not fall below the allowable range, focusing on productivity, the scanning speed of the leveling rollers is increased to shorten the molding time. After that, the process returns to step 141 again to form a powder layer.

A procedure for calibrating the density of the modeling liquid will be described with reference to FIG. First, in step 151, modeling is performed in a normal sequence using shape data for checking (equivalent to steps 102 to 113 in FIG. 19). Then, step 152 removes excess flour. For example, by blowing compressed air provided in the device, the powder in the area where the modeling liquid is not applied is blown off. In step 153, the powder surface sensor 28 takes an image of the powder surface. At this time, the area to which the modeling liquid is applied is focused and photographed. In steps 154 and 155, the area coated with the modeling liquid is extracted. Since the image taken in step 153 is in focus on the area where the modeling liquid is applied, the area where the powder was blown off in step 152 without applying the modeling liquid is out of focus ( Figure 11).

Therefore, first, in step 154, image averaging processing is performed to calculate the luminance average value of the peripheral pixels. Next, in step 155, the difference between the brightness value of each pixel and the brightness average value of the surrounding pixels is calculated, and the permeation area of the modeling liquid can be determined by identifying the difference based on the size. By comparing the result with the shape data for checking in step 156, the permeation range of the modeling liquid can be obtained, and the permeation range of the modeling liquid is indexed. In step 157, the application density of the modeling liquid is adjusted based on the permeation range of the modeling liquid.

The recoating adjustment procedure in step 104 shown in FIG. 19 will be described with reference to FIG. In step 161, the powder surface sensor 28 is used to take an image of the powder surface of the powder bed. In step 162, a plurality of texture analysis results are performed to determine the feature amount of the texture of the powder surface. This uses a technique similar to step 143 shown in FIG. In step 163, the packing density of the powder is estimated from the feature amount of the texture. This uses a technique similar to step 144 shown in FIG. At step 164, if the estimated value of the packing density is within the allowable range, the process proceeds to step 167 and terminates. If it is lower than the allowable range, the process proceeds to step 165 to reduce the scanning speed of the leveling rollers for the next and subsequent layers. Conversely, if the estimated packing density is higher than the acceptable range, the process proceeds to step 166 to increase the scanning speed of the leveling rollers for the next and subsequent layers. Then, the process proceeds to step 167 and terminates.

The procedure of powder bed analysis in step 105 shown in FIG. 19 will be described with reference to FIG. In step 171, the powder surface sensor 28 is used to take an image of the powder surface of the powder bed. This may use the same image as step 161 . At step 172, a calculation is performed for all pixels to determine the luminance difference between the target pixel of the image and the surrounding pixels. In step 173, areas where the luminance difference is equal to or greater than the threshold are extracted and divided into connected clusters. At steps 174 to 178, classification of each mass is performed for each mass that has been divided. The area of the extracted mass is determined in step 174, and if the area is less than or equal to a threshold, then in step 175 it is classified as noise or a minor nonproblematic anomaly. If the area is greater than or equal to the threshold, then in step 176 it is determined whether the extracted mass is a step in the downstream area scanned by the leveling roller. is classified as an anomaly associated with a lack of Otherwise, go to step 178 and classify it as other anomaly. At step 179, it is determined whether or not all clumps have been classified. If not, the process returns to step 174. If so, the process proceeds to step 180 to end the powder analysis.

Referring to FIG. 26, the procedure for adjusting the density of the modeling liquid at step 109 shown in FIG. 19 will be described. In step 191, the powder surface sensor 28 is used to take an image of the powder surface of the modeling layer. Texture analysis of the image is performed at step 192 . The target of texture analysis is the area of the image obtained in step 191 to which the modeling fluid is applied. The content of the texture analysis may be the same as in steps 143 and 162, or may be different. At step 193, it is determined from the results of the texture analysis whether more or less modeling fluid has been applied. The result of texture analysis of the image of the powder surface when the amount of the modeling liquid is changed and applied is stored in advance, and the amount of the modeling liquid is estimated by comparison with the result. If the estimated value of the amount of the modeling liquid is appropriate, the process proceeds to step 196 without doing anything and terminates. If the estimated amount of the modeling liquid is insufficient, the process proceeds to step 194 to increase the concentration of the modeling liquid in subsequent layers. If the estimated amount of the modeling liquid is insufficient, the process proceeds to step 195 to reduce the concentration of the modeling liquid in subsequent layers. Either goes to step 196 and exits.

The procedure of modeling layer analysis in step 110 shown in FIG. 19 will be described with reference to FIG. In step 201, the powder surface sensor 28 is used to capture an image of the powder surface of the modeling layer. The same image as in step 191 may be used for this. In step 202, the difference between each pixel and the average brightness of surrounding pixels is calculated. In step 203, a shadow is extracted as a region where the difference in brightness is equal to or greater than the threshold and is located on the boundary between the region to which the modeling liquid is applied and the region to which the modeling liquid is not applied.

At step 204, it is determined whether the shadow size is within an acceptable range. Based on the shape data, the relationship between the size of the warp and the size of the shadow is obtained in advance. can determine if there is If the magnitude of the warp is not within the allowable range, the process proceeds to step 205 to make adjustments such as increasing the heating temperature in step 110 or lengthening the heating/waiting time.

FIG. 28 is a diagram showing another configuration example of the modeling system. The modeling system shown in FIG. 28 includes a modeling apparatus 10 and an information processing device 11 as another embodiment of the modeling system shown in FIG. cloud) 12. The modeling apparatus 10, the information processing apparatus 11, and the information management apparatus 12 are communicably connected to each other via a cable or network. The network may be a wireless network or a wired network. Further, the information management device 12 is not limited to the information processing device 11 and may have a part of the functions of the modeling device 10 .

As described above, the surface information is acquired each time a layer is formed, and based on the acquired surface information, the three-dimensional modeling is performed by controlling the operation of forming the powder layer, the operation of applying the modeling liquid, etc. Variations in the shape and strength of objects can be reduced, and the quality of three-dimensional objects can be improved.

Although one embodiment of the present invention has been described so far, the present invention is not limited to the above-described embodiment, and the constituent elements of the present embodiment may be changed or deleted, or the constituent elements of the present embodiment may be changed. It can be modified within the range that a person skilled in the art can conceive, such as adding other components, and as long as the effects of the present invention are exhibited in any aspect, it is included in the scope of the present invention. .

DESCRIPTION OF SYMBOLS 10... Modeling apparatus 11... Information processing apparatus 20... Powder 21... Modeling liquid 22... Ink jet print head 23... Powder head 24... Leveling roller 25... Powder feed piston 26... Build piston 27... Modeling layer 28... Powder Surface sensor 29 Control circuit 30 Data accumulation means 31 Learning means 32 Estimation means 33 Powder surface 34 Unevenness, cracks, aggregates 40 Hopper feeder 40a Hole 41 Roller 42 Print head 43 Infrared lamp 44... Modeling tank 45... Build piston 46... Powder recovery groove 47... Green body 50... Powder 51... Gap 52... Modeling liquid 53... Portion where powder remains 54... Range 55 where modeling liquid permeates ... area 56 applied with the modeling liquid ... light source 57 ... shadow

Japanese Patent Application Laid-Open No. 2020-142420

Claims (15)

A modeling apparatus for modeling a three-dimensional object,
Forming means for forming a layer containing powder;
an application means for applying a modeling liquid to the layer;
Acquisition means for acquiring surface information of the layer;
and a control means for controlling the operation of at least the forming means or the applying means based on the surface information. The acquiring means acquires the surface information after the forming means forms the layer,
2. The modeling apparatus according to claim 1, wherein said control means changes the operation of said forming means for the next layer according to said acquired surface information. 3. The modeling apparatus according to claim 2, wherein said control means changes the amount of powder supplied by said forming means as the change in the operation of said forming means. the forming means comprises a roller;
4. The modeling apparatus according to claim 2, wherein said control means changes at least the rotational speed or scanning speed of said roller as the change in the operation of said forming means. The forming means includes a container for supplying powder and vibration imparting means for imparting vibration to the container,
4. The modeling apparatus according to claim 2, wherein said control means changes at least the vibration frequency or the vibration time of said vibration as the change of the operation of said forming means. The acquiring means acquires the surface information after the coating means has coated the modeling liquid,
The modeling apparatus according to any one of claims 1 to 5, wherein said control means changes the operation of said application means according to said acquired surface information. 7. The modeling apparatus according to claim 6, wherein said control means changes at least an amount of said modeling liquid or corrects an area to which said modeling liquid is applied, as a change in the operation of said applying means. The modeling according to any one of claims 1 to 7, wherein the control means controls at least the operation of the forming means or the applying means based on the powder packing density obtained from the acquired surface information. Device. 9. The modeling apparatus according to claim 8, wherein, as the surface information, the control means estimates the packing density from at least the shadow of powder particles on the surface. 8. The modeling apparatus according to claim 6, wherein said control means controls said coating means based on at least a surface height difference or a permeation range of said modeling liquid as said surface information. including heating means for heating the layer after the modeling liquid is applied,
The modeling apparatus according to any one of claims 1 to 10, wherein said control means changes at least the heating time or heating temperature by said heating means according to said acquired surface information. 12. The modeling apparatus according to any one of claims 1 to 11, wherein said control means changes a waiting time according to said surface information when determining that waiting is necessary based on said acquired surface information. . data accumulation means for accumulating modeling data of the three-dimensional object, surface information of the three-dimensional object, and data including the operation of the modeling apparatus;
a learning means for extracting a feature amount based on the data and learning based on the feature amount;
an estimating means for estimating the surface information and the operation of the modeling apparatus based on the learning by the learning means;
2. The modeling apparatus according to claim 1, wherein said control means controls at least the operation of said forming means or said coating means based on the result estimated by said estimation means. A modeling system for modeling a three-dimensional object,
Forming means for forming a layer containing powder;
an application means for applying a modeling liquid to the layer;
Acquisition means for acquiring surface information of the layer;
and control means for controlling the operation of at least the forming means or the applying means based on the surface information. A method for controlling modeling of a three-dimensional object,
forming a layer comprising powder;
applying a modeling fluid to the layer;
obtaining surface information of the layer;
and controlling operations in at least the forming step or the applying step based on the surface information.

Images