Configuration for forming a three-dimensional structure and related forming method

Abstract

An embodiment relates to a method for forming a three-dimensional structure. The method includes determining one or more locations for positioning a sensing device based on structure coordinate information for the three-dimensional structure to be formed. The method further includes forming a portion of the three-dimensional structure based on the structure coordinate information. The method further includes positioning the sensing device at one of the one or more locations.

Description

Technical Field

[0001] (Cross - reference to related applications) This application claims the priority of European Patent Application No. 20155109.0 filed with the European Patent Office on February 3, 2020, and the entire content thereof is incorporated herein by reference for all purposes.

[0002] The embodiments described herein relate to a configuration for forming a three - dimensional structure and a forming method related thereto.

Background Art

[0003] The manufacturing of a three-dimensional structure can be monitored using indirect or direct measurement of printing errors. In indirect measurement, printing parameters as well as environmental variables are measured, while in direct measurement, defects of the printed structure itself are monitored. For example, in fused deposition modeling (FDM) printing, various sensors may be used for indirect measurement, as many measurable variables can represent the printing state. For example, the acoustic emission (AE) method can be used to determine an abnormal printing state and detect filament breakage. In some applications, hidden semi-Markov models and k-means methods may be used to distinguish printing states. In some applications, monitoring of nozzle clogging may be performed. In some applications, the deposition state of the material can be determined by measuring the current of the filament supply motor. In some applications, a two-dimensional laser triangulation method may be used to scan the size of the extruded track. In some applications, ultrasonic excitation can be used to detect bonding defects during printing. In some applications, fiber Bragg grating sensors and thermocouples can be embedded in the sample to monitor the residual stress and temperature profile generated during printing. Thermocouples may be used to evaluate the temperature conditions during the printing process. In some applications, by applying compressive sensing, it may be possible to reconstruct the temperature field of a melt filament processed sample using four temperature measurement values in one step. In some applications, a sensor array with multiple sensors can monitor the printing process. Classification of the states of various processes (normal operation, abnormal operation, build failure) can be performed by analyzing sensor data using a nonparametric Bayesian model.

[0004] For direct measurement, the image of the printed model can be compared with the CAD model by using augmented reality-based technology. In some application examples, a single camera system and a dual camera system can be used to detect clogged nozzles, incomplete protrusions, and filament damage. In some application examples, a three-dimensional image of the part during printing can be reconstructed using two cameras to find the difference between the point clouds of the ideal model and the printed part. In some application examples, the printed part can be imaged with a two-dimensional camera. The center of the printed part can be detected and compared with the center of the ideal geometric arrangement. In some application examples, the border signature method is applied to detect the deviation of the geometric arrangement in the geometric arrangement outside the simple solid, and the ideal outer shape can be compared with the outer shape in the printed layer direction. In some application examples, a USB (registered trademark) microscope video camera can be used to measure and control the deviation.

[0005] The on-line inspection method can be used for other printing technologies. For example, a computed tomography (CT) scan of a part manufactured by a metal-based powder bed fusion method can be performed, and two-dimensional images of all layers can be analyzed by multifractal analysis. In some applications, fringe projection can be used to measure the surface topography of an AM-manufactured part. Furthermore, on-line signature analysis can be performed on the layer of the ceramic sensor. In some applications, the signature of the extracted track can be compared with the template signature to detect defects. In some applications, in order to image the error in selective laser melting, the intensity change of a single pixel in consecutive images from a video frame can be captured by a statistical method, and pixels with abnormal intensity profiles over time can be detected. Moreover, specific points of the printed part or label printed on the object to be printed can be recognized to capture the appearance of the additive manufactured part.

Summary of the Invention

[0006] Various embodiments relate to a method and configuration for monitoring the quality of a three-dimensional structure during a manufacturing process. The method and configuration provide an improved process for classifying defects and their causes and identifying their locations during printing.

[0007] Various embodiments relate to a method of forming a three-dimensional structure. The method includes determining one or more locations for positioning a sensing device based on coordinate information of the structure with respect to the three-dimensional structure to be formed. The method further includes forming a portion of the three-dimensional structure based on the coordinate information of the structure. The method further includes positioning the sensing device at one of the one or more locations.

[0008] Various embodiments relate to a configuration for forming a three-dimensional structure. The configuration includes a forming device for forming a three-dimensional structure based on coordinate information of the structure with respect to the three-dimensional structure. The configuration further includes a movable sensing device. The configuration includes a processor configured to determine one or more locations for positioning the sensing device based on coordinate information of the structure with respect to the three-dimensional structure to be formed and to control the positioning of the movable sensing device to the one or more locations.

[0009] Various embodiments relate to a configuration for forming a three-dimensional structure. The configuration includes a forming apparatus for forming a three-dimensional structure including a plurality of layers. The configuration further includes a movable sensing device. The configuration further includes a processor configured to determine at least one location for positioning the sensing device for one layer set consisting of a plurality of layers. The processor is further configured to control the formation of the layer set by the apparatus based on the coordinate information of the structure regarding the three-dimensional structure. The processor is further configured to control the positioning of the movable sensing device to at least one location associated with the layer set from forming the layer set until forming a further layer set. The processor is further configured to determine a process state based on the acquired structural data of the formed layer set, and the acquired structural data is acquired by the sensing device at at least one location associated with the formed layer set.

[0010] The foregoing and other features of the present disclosure will become more fully apparent from the following description and the appended claims, taken in conjunction with the accompanying drawings. The accompanying drawings merely illustrate several embodiments according to the present disclosure and thus are not to be considered as limiting its scope. The present disclosure will be described more specifically and in detail with reference to the accompanying drawings so that the advantages of the present disclosure can be more easily confirmed.

Brief Description of the Drawings

[0011]

Figure 1A

Figure 1B

Figure 2A

Figure 2B

Figure 3

Figure 4A

Figure 4B

Figure 5A

Figure 5B

Figure 6

Figure 7A

Figure 7B

Figure 7C

Figure 8

Figure 9A

Figure 9B

Figure 9C

Figure 9D

Figure 10A

Figure 10B

Figure 10C

Figure 10D

Figure 10E

Figure 10F

Figure 11A

Figure 11B

Figure 11C

Figure 11D

Figure 11E

Figure 11F

Figure 12A

Figure 12B

Figure 12C

Figure 12D

Figure 12E

Figure 12F

Figure 13A

Figure 13B

Figure 13C

Figure 14A

Figure 14B

Figure 14C

Figure 15A

Figure 15B

Figure 16

DETAILED DESCRIPTION OF THE INVENTION

[0012] In the following detailed description, reference is made to the accompanying drawings, which illustrate, for purposes of illustration, specific embodiments in which the subject matter of the claims can be practiced. These embodiments are described in sufficient detail to enable those skilled in the art to practice the subject matter. It is to be understood that the various embodiments are different but not necessarily mutually exclusive. It should be understood that the terms "embodiment", "example of an embodiment", "exemplary embodiment", and "the present embodiment" do not necessarily refer to a single embodiment, and that they and examples of various embodiments can be readily combined and / or interchanged without departing from the scope or spirit of the examples of embodiments. For example, the specific features, structures, or characteristics described herein in connection with one embodiment can be practiced in other embodiments without departing from the spirit and scope of the subject matter recited in the claims. When the term "one embodiment" or "an embodiment" is used herein, it means that a specific element, structure, or feature described in connection with the embodiment is included in at least one implementation encompassed by this description. Thus, the use of the phrases "one embodiment" or "an embodiment" does not necessarily refer to the same embodiment. Moreover, it should be understood that the location or configuration of individual elements in each disclosed embodiment can be changed without departing from the spirit and scope of the subject matter recited in the claims. Accordingly, the following detailed description is not to be taken in a limiting sense, and the scope of the subject matter is defined only by the properly construed appended claims, along with the full scope of equivalents to which the claims are entitled. In the drawings, like numbers refer to the same or similar elements or functions throughout the several views, and the elements depicted in the figures are not necessarily to scale with each other, but rather, the individual elements may be enlarged or reduced to more readily include the element within the context of the description.

[0013] As used herein, the terms "above", "to", "between", and "on" may refer to the relative position of one layer with respect to another layer. One layer "above" or "on" another layer may be in direct contact with the other layer or may have one or more intervening layers. One layer "between" layers may be in direct contact with the layers or may have one or more intervening layers.

[0014] The terms "a", "an", and "the" may refer to the singular and plural. Further, as used in this disclosure and the appended claims, the term "and / or" refers to any and all combinations of one or more of the listed related items and can include them. The phrase "A and / or B" as used in this disclosure means (A), (B), or (A and B). The phrase "A, B, and / or C" as used in this disclosure means (A), (B), (C), (A and B), (A and C), (B and C), or (A, B, and C). The term "or" as used in this disclosure, when used in the phrase "A, B, or C", means that (A) does not exclude (B) and (C), (B) does not exclude (A) and (C), and (C) does not exclude (A) and (B).

[0015] FIG. 1A shows a flowchart of a method 100 for forming a three-dimensional structure.

[0016] Method 100 includes a step 110 of determining one or more locations for positioning a sensing device based on coordinate information of a structure related to the three-dimensional structure to be formed. Method 100 further includes a step 120 of forming a portion of the three-dimensional structure based on the coordinate information of the structure. Method 100 further includes a step 130 of positioning the sensing device at one of the one or more locations.

[0017] The coordinate information of the structure may include or may be coordinate information related to the layout of the three-dimensional structure to be formed. The coordinate information of the structure may include coordinate information related to the physical structure and / or the location, position, orientation, dimensions, shape and / or form of the layout of the three-dimensional structure to be formed. The coordinate information of the structure may be information based on a coordinate system that uniquely defines and / or describes the location, position, orientation, dimensions, shape and / or form of the features of the three-dimensional structure to be formed using scales or axis numerical values. Optionally, the coordinate information of the structure may be based on a three-dimensional Cartesian coordinate system having three mutually perpendicular axes (e.g., the x-axis, y-axis and z-axis) and three mutually orthogonal planes. The coordinate information of the structure may include information related to the internal structure and / or external structure of the three-dimensional structure printed with reference to the three-dimensional Cartesian coordinates. Alternatively, the coordinate information of the structure may be based on any other coordinate system such as a spherical coordinate system, a polar coordinate system, an elliptical coordinate system or a cylindrical coordinate system, or any coordinate system whose coordinates are convertible between Cartesian coordinates.

[0018] The coordinate information of the structure may (or may be) include computer-aided design (CAD) information related to the construction, mapping, form or layout of the three-dimensional structure. Moreover, alternatively or optionally, the coordinate information of the structure may include computer-aided manufacturing (CAM) information for controlling a forming device (e.g., a three-dimensional printing device or configuration). Moreover, alternatively or optionally, the coordinate information of the structure may (or may be) include tool path commands for controlling a forming device that forms the three-dimensional structure. For example, the coordinate information of the structure may (or may be) include numerical code commands such as G-code commands. The numerical code tool path commands and / or G-code commands may be based on and / or may include coordinate information related to the structure and / or construction of the three-dimensional structure to be formed. Such commands, when executed by a processor, can control the operation and / or path of a device (e.g., a feeding device) for forming the three-dimensional structure.

[0019] The coordinate information of the structure may include information related to the internal structure and / or external structure of the three-dimensional structure. The three-dimensional structure to be formed may include, for example, but is not limited to, a three-dimensional scaffold structure and / or a mesh structure. Alternatively, or optionally thereon, the three-dimensional structure may include filling lines (also referred to herein as strands), or may be any structure including a lattice network of strands or lines. Instead of filling lines, the strands may also include outfill lines, which form the outer boundary of the structure.

[0020] The three-dimensional structure formed (such as a scaffold structure and / or a mesh structure) may include a plurality of layers or a plurality of layer sets. In the three-dimensional forming (or printing) process (or during it), the three-dimensional structure forms a first layer and then continuously forms layers on top of each other such that the continuous layers are stacked perpendicular to each other in the printing direction (which may be referred to as the z-direction in this specification). Optionally, such a printed layer may be understood as, or referred to as, a lateral layer, a horizontal layer, a planar layer, or a flat layer. By these terms, the layer is understood to be a layer in the x-y plane or within the x-y plane. Such a printed layer may have dimensions extending along the x-axis (x-direction) and the y-axis (y-direction) that are larger than the dimension in the z-direction (for example, at least 10 times, or for example, at least 50 times, or for example, at least 100 times). The terms lateral layer and horizontal layer as used in this specification can be understood to refer to a direction perpendicular to the printing direction, which can be considered the z-direction or the vertical direction. Once the printing of one lateral (or horizontal) layer is completed, the printer prints a continuous lateral (or horizontal) layer on top of (or, for example, above, or, for example, on top of, or, for example, covering) the previous layer, such that the lateral layers can progress to be stacked perpendicular to each other (for example, vertically stacked) in the printing direction (for example, in the case of three-dimensional printing). Optionally, one layer set (or layer) may include a first sublayer (or a first group of sublayers) of strands directed in a first direction and a second sublayer of strands directed in a second direction different from the first direction. Optionally, one layer set is not limited to including only the first group of sublayers or the second group of sublayers, but may include any possible number of sublayers. Each group of sublayers may include (or refer to) one or more sublayers.Optionally, intersecting (or, for example, cross-intersecting) lines or strands may form a plurality of repeating unit cells of one layer set.

[0021] The outer structure of the three-dimensional structure can be the region of the outer surface of the three-dimensional structure. The region of the outer surface can refer to (or can be) the outermost surface, the outermost layer, and / or the outermost contour of the three-dimensional structure. The outermost surface and the outermost contour can be formed from one or more layers or strands. The region of the outer surface can be referred to as (or can be) the outermost layer group of the three-dimensional structure (for example, the outermost single layer, or, for example, the outermost plurality of layers, or one boundary layer). The region of the outer surface can refer to the surface of the three-dimensional structure facing the outside (or, for example, the outer side).

[0022] The internal structure of the three-dimensional structure may be any part of the three-dimensional structure other than the outermost boundary of the three-dimensional structure. Additionally, or alternatively, or optionally, a unit cell whose all sides are surrounded by other unit cells within one layer set can be regarded as the internal structure of the three-dimensional structure. Conversely, a unit cell at the boundary that may include at least one side not surrounded by another unit cell can be regarded as part of the external structure of the three-dimensional structure. The internal structure of the three-dimensional structure may be any part of the three-dimensional structure that is arranged more than twice the desired strand width or thickness of the strand to be printed from the outermost surface. The desired width or thickness of the strand may be, for example, the ideal thickness of the strand of the three-dimensional structure to be printed. Optionally, the average thickness (or width, or diameter) of the strand may be between 0.001 mm and 30 mm (or, for example, between 0.001 mm and 1 mm, or, for example, between 0.01 mm and 0.5 mm), but is not limited thereto. Optionally, the separation distance between adjacent lines may be between 0.001 mm and 30 cm (or, for example, between 0.001 mm and 50 mm, or, for example, between 0.01 mm and 50 mm), but is not limited thereto. Similarly, each sublayer oriented in the second direction can be separated from an adjacent (or consecutive) sublayer oriented in the (same) second direction by a separation distance between 1% and 100% of the average thickness of the strand.

[0023] Method 100 includes determining one or more locations (or, for example, a plurality of locations) for positioning a sensing device based on the coordinate information of the structure related to the three-dimensional structure to be formed. The processor can be configured to determine (or calculate) one or more locations from the coordinate information of the structure or based on the coordinate information of the structure. The one or more locations can be determined from the coordinate information of the structure, for example, (as shown in FIGS. 2A - 2B) by calculating or determining the intersection regions or intersections between the strands (or strand segments) of the three-dimensional structure.

[0024] Step 120 of forming a portion of the three-dimensional structure may include the step of printing a portion of the three-dimensional structure by a three-dimensional printing apparatus or configuration. The three-dimensional structure can be formed by a three-dimensional (3D) printing process. The three-dimensional printing process may include at least one of stereolithography (SLA), digital light processing (DLP), fused deposition modeling (FDM), selective laser sintering (SLS), selective laser melting (SLM), electron beam melting (EBM), laminated object manufacturing (LOM), binder jetting (BJ), and material jetting (MJ). The processor can execute numerical code commands based on the coordinate information of the structure to control the operation and / or path of at least a portion of the forming apparatus for forming the three-dimensional structure. In some examples, a portion of the apparatus whose operation is controlled by the coordinate information of the structure may be a supply portion for supplying a printing material for forming a three-dimensional structure (e.g., a strand for a three-dimensional structure), such as in the fused deposition modeling method. In other examples, a portion of the apparatus may be controlled by a laser, and the coordinate information of the structure controls a selective location that can generate at least one of sintering, curing, melting, bonding, laminating, and / or resin curing on the material for forming the three-dimensional structure. It can be understood that a portion of the apparatus whose operation is controlled by the numerical code commands forms (e.g., forms a strand of the three-dimensional structure) the three-dimensional structure at a location defined and selected by the coordinate information of the numerical code commands and is a portion of the apparatus that plays a role. Alternatively or optionally, the three-dimensional structure can be formed by a printing process based on an operation beyond three dimensions, such as a five-dimensional (5D) printing process or a six-dimensional (6D) printing process.

[0025] Since method 100 is related to an in-situ monitoring process, it can be understood that the formed part (formed at 120) referred to herein may be smaller than the entire three-dimensional structure to be formed. For example, a part of the three-dimensional structure may be a certain layer set (e.g., one layer set, or for example more than one layer set) consisting of a plurality of layers (and / or a plurality of layer sets) of the three-dimensional structure. The formed part may be or may include an internal structure of the three-dimensional structure. For example, the formed part may include one or more layer sets including a plurality of intersecting lines or strands forming a lattice of unit cells within the three-dimensional structure. A defect detection process for a plurality of layers will be further described in relation to FIG. 1B.

[0026] Method 100 may include determining a group consisting of locations for positioning a sensing device in relation to the formed part of the three-dimensional structure, before starting the forming process or during the forming process. For example, the group consisting of locations can be determined before and after the formation of a part of the three-dimensional structure. Method 100 may include sequentially positioning the sensing device at a location among the group of locations determined in relation to the already formed part, before forming a further (or for example the next) part of the three-dimensional structure. The group consisting of locations can include several locations. Some of those locations can be random, consecutive, array-based and / or determinable by the user. The group consisting of locations may refer to one or more (e.g., a plurality of) locations (or may be so, or may include them).

[0027] At an individual location of the group consisting of locations (e.g., each location), method 100 may include obtaining data of at least a portion of the portion to be formed by a sensing device. The sensing device may be any sensing device capable of obtaining data including the structural information of the portion to be formed. For example, the sensing device may be at least one device from a group of devices, and the group of devices may consist of an imaging device, a camera, a thermal camera, a microscope, a laser scanner, and a three-dimensional scanning device. Optionally, the acquired data may be a two-dimensional image or a three-dimensional image of the portion to be formed obtained at each location of the group consisting of locations.

[0028] The acquired data may include visual data, and the visual data may be related to one or more structural features of the portion to be formed of the three-dimensional structure and / or may include or be such structural data about the structural features. The structural information obtained from and / or determined from the acquired data may include information related to the physical structural features and / or the location, position, orientation, dimensions, shape, and / or form of the layout of the portion to be formed of the three-dimensional structure. For example, the structural features may be lines, line segments, strands, strand segments, holes, and / or walls of the portion to be formed of the three-dimensional structure.

[0029] Furthermore or optionally, method 100 may further include determining a process state based on the acquired data related to at least a portion of the portion to be formed of the three-dimensional structure. The process state can be determined after obtaining data related to the portion to be formed (e.g., a layer set) and (optionally) before forming a further portion (e.g., a further layer set) of the three-dimensional structure. The further portion referred to herein may be the portion to be formed next, but is not limited thereto. For example, the forming device may continue the printing process until the process state is determined, which may mean that one or more further portions of the three-dimensional structure may already be printed depending on the time required to determine the process state.

[0030] The step of determining the process state may include the step of determining a parameter value of a structural feature of the portion to be formed of the three-dimensional structure based on the acquired data. The parameter value of the structural feature may be at least one of length, width or diameter, height, roughness, color, homogeneity, thickness and tilt angle of the structural feature of the portion to be formed of the three-dimensional structure. Optionally, the parameter value of the structural feature may be determined based on edge detection of the strand segment (for example, by detecting the edge of the strand segment). For example, the parameter value may be the width of the strand segment between two detected edges of the strand segment. The edge of the strand segment can be detected, for example, by performing an artificial neural network process or a tilt-based detection process. Optionally, the parameter value of the structural feature may be determined based on the difference between the structural feature and a predetermined baseline or a comparison structure (for example, by subtracting image values), and / or by determining the deviation between the structural feature and a predetermined baseline or a comparison structure.

[0031] Optionally, the step of determining the process state may include the step of determining (for example, calculating, or for example, generating) an ideal axis of a strand segment of a portion of the three-dimensional structure based on the coordinate information of the structure. The step of determining the process state may further include the step of determining (for example, measuring, or for example, calculating) an actual axis of the strand segment of the portion to be formed based on the acquired data of the portion to be formed. The process state can be determined based on the comparison between the ideal axis and the actual axis. A defective process state can be determined when the difference between the ideal axis and the actual axis exceeds a threshold value.

[0032] Optionally, the step of determining the process state may include comparing the ideal parameter values of the strand segments of a portion of the three-dimensional structure with the determined parameter values of the strand segments of the portion to be formed. The determined parameter values of the strand segments may be determined based on the acquired data of at least a portion of the portion to be formed of the three-dimensional structure. The ideal parameter values of the strand segments may be determined based on at least one of the coordinate information of the structure and the input value. The input value may be the input value of one or more users and / or one or more values of one data set or database (for example, one set of data for several prints or data of a previous process). When the difference between the ideal parameter value and the actual parameter value exceeds a threshold value, a defective process state may be determined.

[0033] Furthermore, optionally or alternatively, the step of determining the process state may include determining a plurality of parameter values associated with a plurality of strand segments of the portion to be formed. Statistical parameters (for example, at least one of standard deviation, variance, median, mode, range, correlation, frequency, maximum, minimum, quartile, mean, error) of the plurality of parameter values may be determined. When the difference between the statistical parameter and the comparison parameter exceeds a threshold value, a defective process state may be determined.

[0034] Method 100 may further include the step of adapting process parameters for forming a three-dimensional structure based on a determined process state. The process parameters may include at least one process parameter from the group consisting of a forming temperature, an ambient temperature, a cooling process, a layer gap, a nozzle cleanliness, a flow rate of a printing material, a printing speed, a tool path, and an ambient humidity, and / or a laser output. The ambient temperature may be a room temperature and / or a chamber temperature or a temperature of an enclosed space in which the three-dimensional structure is formed. The forming temperature (e.g., a printing temperature) may be a temperature of a cartridge holding a material (e.g., a polymer) for forming a supplied three-dimensional structure. The laser output may be a laser output and / or energy for sintering, curing, melting, bonding, laminating, and / or resin curing of a material for forming a three-dimensional structure.

[0035] FIG. 1B shows a flowchart of a method 160 for forming a three-dimensional structure. Method 160 may include one or more or all of the features already described in connection with FIG. 1A. FIG. 1B shows a process of a defect detection system and a method of on-site monitoring of a process of forming a three-dimensional structure.

[0036] As described in connection with FIG. 1A, method 160 may include the step of determining respective groups of locations for positioning a sensing device for each one layer set (or, for example, each respective layer set) consisting of a plurality of layers of a three-dimensional structure to be printed. The group of locations may be determinable based on coordinate information of the structure of the three-dimensional structure to be printed, and the group of locations may optionally be determinable before the start of the formation process (e.g., the printing process) of the three-dimensional structure. Alternatively, the processor may be configured to determine the group of locations during the printing process as long as the group of locations for each respective part of the three-dimensional structure is determined before or at the time when each respective part begins to be formed or until it is completely formed. Optionally or alternatively, the processor may be configured to determine the group of locations for each respective part after each respective part is formed.

[0037] Method 160 may further include generating numerical code instructions adapted to form a three-dimensional structure. For example, the adapted numerical code instructions may include instructions for controlling the operation of the forming device and the operation of the movable sensing device. The adapted numerical code instructions may include the coordinate information of the structure related to the three-dimensional structure to be formed and the information related to the step of positioning the sensing device at one or more locations. For example, the information related to the step of positioning the sensing device may be information for controlling the positioning or operation of the movable sensing device to one or more locations. Further, the adapted numerical code instructions can adapt the timing sequence or operation sequence of the forming device in consideration of the operation of the movable sensing device to one or more locations.

[0038] By the processor executing the adapted numerical code instructions, method 160 may include step 120 of forming a portion of the three-dimensional structure based on the coordinate information of the structure (as described in relation to method 100). Step 120 of forming a portion of the three-dimensional structure may include or mean forming (e.g., printing) one layer set out of a plurality of layer sets (e.g., the first layer set, or for example, any first layer set).

[0039] After forming the (first) layer set and before forming a further (or for example, the next, for example, the second, or for example, any second) layer set, method 160 may include step 140 of determining a process state (for example, a manufacturing state) in which the three-dimensional structure is formed. Step 140 of determining the process state may include sequentially positioning a sensing device at locations in a group of locations determined for the layer set to be formed. Step 140 of determining the process state may further include step 130 of obtaining data (for example, by acquiring an image) related to the layer set to be formed at each of the locations in the group of locations determined for the layer set to be formed. Step 140 of determining the process state may further include step 141 of processing the acquired data (for example, by processing each acquired image).

[0040] Step 140 of determining the process state may include step 142 of determining a defective process state (for example, determining whether a significant printing defect has occurred) based on a comparison between an ideal parameter value and an acquired parameter value. If the processor determines a defective process state, the processor may execute command 180 to stop the forming process or command 170 to adapt the printing parameters. If a defective process state is not determined, the processor executes a command to continue the forming process and the next layer set may be printed.

[0041] Method 160 may include alternately repeating a step of forming one layer set (or one or more layer sets) out of a plurality of layer sets, and a step of sequentially determining a process state from when the layer set is formed until a further layer set is formed (which may include positioning a sensing device at a location of a group consisting of locations of the formed layer sets). Optionally, the latter alternately repeating process is executable until the printing process is completed and a three-dimensional structure is formed, or alternatively, until a critical defect is determined for which the processor executes a command to stop printing. Optionally, even if no defect is determined, the alternately repeating process may be executed until a particular three-dimensional structure (e.g., a predetermined portion) is printed. Such a predetermined portion may be a specified or particular number of layers out of the total number of layers of the structure, or may be a specified or particular location or portion within the three-dimensional structure. Such a predetermined portion may be a portion of the three-dimensional structure that is more or less complex than other portions of the three-dimensional structure and may be selectable according to the requirements of the manufacturing process.

[0042] A sensing device (e.g., a commercially available digital microscope) can be used to capture images (photos) of the region of interest in the layer direction. These images can be processed to detect defects. FIG. 1B shows the features regarding the layer of detection method 160. Each time a layer is printed, the microscope can be automatically positioned above the specified region of interest by the axis of the printer. Depending on the field of view and the size of the monitored features, one or more images can be captured and processed. For the automation of the camera positioning and image capture, it may be necessary to know the image capture time and the optimal camera position. This information is derived from a numerical code (e.g., G-code), which includes all the operations of the printer used to print the three-dimensional scaffold structure. The numerical code can be modified or adapted or used to determine the potential camera positions and can be changed with commands for positioning the camera. Optionally, the sensing device may be arranged on an individual or different axis system from the printing device (e.g., print head). Using the adapted numerical code, two individual axis systems, namely, the axis system of the sensing device and the axis system of the printing device, can be controlled. For example, the adapted numerical code may be sent to a plurality of different processors or controllers that control the different axis systems respectively.

[0043] The following image processing can be utilized for measuring the diameter of the strand. If the measurement is within a predetermined tolerance range, the printing can continue. If there are significant deviations, the printing can be stopped (step 180). If the deviation is within a predetermined tolerance (e.g., a deviation that is not large enough to stop the printing but may interfere with subsequent layers), the printing parameters may be adjusted (step 170). After the adjustment 170, the printing process may proceed. Methods 100 and 160 can be used, for example, to detect the diameter of the strand and, thereon, or alternatively, optionally, to detect breakage or geometric deviations.

[0044] Figures 2A and 2B respectively show examples of different layer sets described with respect to the method of FIGS. 1A-1B.

[0045] Figures 2A and 2B show diagrams of layer sets of unit cells. The layer set may include a first sublayer (or a first group of sublayers) including strands 228 oriented in a first direction and a second sublayer (or a second group of sublayers) including strands 229 oriented in a second direction different from the first direction.

[0046] Optionally, the strands of each sublayer may be a portion of the strands of a continuous sublayer that extends continuously from a starting point S of the sublayer to an ending point E of the sublayer. For example, the strand 228 of the first sublayer may be a portion of the strands of a continuous sublayer that meanders continuously from the starting point S to the ending point E of the first sublayer. For example, the strand of the second sublayer may be a portion of the continuous sublayer strands that meanders continuously from the starting point S to the ending point E of the second sublayer. When the terms line or strand are used in the specification, they may include or refer to not only straight lines but also curved or free-form strands. For example, the intersecting strands forming the unit cell may be straight lines, or alternatively, the strands may be sinusoidal lines or curved, and the unit cell may take on a "free form" shape. Alternatively or optionally, the layer may be printed as a path having more than one starting and ending point. For example, there may be a gap between each layer, or the layer may include different regions, or the layer may include different structures formed simultaneously.

[0047] Optionally, some within each sublayer of strands 228, 229 may be parallel to each other (e.g., the acute angles between strands within a sublayer or between strands that best fit the sinusoidal strands may be within the range of + / -5°). Alternatively, the strands may be curved, or may extend in random directions rather than being parallel to each other.

[0048] Optionally, a plurality of strands 228 of a first sublayer and a plurality of strands 229 of a second sublayer may be intersected at intersections or intersection regions so as to form a two-dimensional lattice configuration of two-dimensional unit cells of the layer. The two-dimensional unit cells of each layer of the stacked layer configuration can form a three-dimensional lattice structure. Each unit cell of the layer set can include or be formed from intersecting strands 228, 229 from adjacent sublayers that define the pore size of the unit cell. For example, two adjacent (and parallel) strands 228 of a first sublayer can be made to intersect two adjacent (and parallel) strands 229 of a second (adjacent) sublayer. The region of the unit cell surrounded by the intersecting strands may be a rhomboid unit cell, a polygonal unit cell, a triangular unit cell, a diamond-shaped unit cell, a free-form shaped unit cell, a square unit cell, a parallelogram unit cell, and / or a hexagonal unit cell.

[0049] One layer set may include at least a first sublayer and at least a second sublayer. A particular location among a plurality of locations may be determined from the coordinate information of the structure based on an intersection 231 between the first sublayer and the second sublayer (e.g., an intersection between strand 228 of the first sublayer and strand 229 of the second sublayer). For example, a particular location 232 among a plurality of locations 232 may be between two intersections 231 of the first sublayer and the second sublayer.

[0050] After a printing process (e.g., an FDM process), a layer-related method may follow. The structure of the support material S can be represented by a layer set L.

Number

Number

[0051]

Number

[0052] The geometric description of the support material enables the definition of the appropriate camera location (or position) or the region of interest of the layer of the structure. The region of interest may be the region where the connection points between the current layer and the previous layer occur. In the case of a symmetric nozzle, the center of these connection points can be represented by the intersection point (j > 1) of two consecutive layers L j and L j-1 projected onto the (x, y) plane. The intersection point can be calculated by the following formula.

Number

[0053] There can be various methods for calculating these points and determining the corresponding intersection lines. If it is assumed that the number of intersection points is considerably smaller than the square of the number of lines, it may be appropriate to apply the Bentley-Ottmann algorithm. Before calculating the intersection points, the lines can be filtered to exclude the lines in the boundary region for shortening the calculation time. By this operation, the lines can be filtered by their length and adapted to the aperture diameter of the printed part. For example, lines shorter than the minimum aperture diameter of the structure to be printed may be excluded. By calculating the intersection points, the position of the camera can be calculated according to the field of view of the camera used. If the field of view is wide enough to capture the entire layer, the center of all the intersection points can be selected as the camera spot. The position of the camera above a narrower region where the intersection points are arranged can also be calculated when the field of view is too narrow to monitor the entire layer. In this case, the layer can be scanned sequentially by overlapping the scan regions or providing boundaries in order to evaluate the entire layer. Furthermore, the layer may be monitored partially. Based on randomly selected intersection points, the layer can be imaged partially. The strand portion between two intersection points may be regarded as the region of interest. This technique may be appropriate since errors in the diameter of the strand usually occur at the bridges between the intersection points. Optionally, the position of the camera may be calculated to be between (e.g., at the center of) two consecutive intersection points. To calculate these positions, all the lines of layer j

Number

Number

Number

Number

[0054] Figure 2A shows a layer set with two consecutive sub-layers having several intersection points and a number of possible camera locations. Of all the possible camera locations, for example, five points (identified by X in Figure 2A) can be randomly selected. Figure 2A shows the results of structure separation and position calculation. In this specification, five monitor positions can be found at a distance from two intersection points to the boundary. Figure 2B shows a layer set with two sub-layers having fewer intersection points than in Figure 2A. For example, one camera location (identified by X in Figure 2B) can be randomly selected from among the even fewer possible camera locations. Alternatively, since the distance from the camera position to the outer boundary is set, a single monitor position can be found for that layer. The x-axis and y-axis indicate the coordinates of the printer operation. This example shows that the number of locations can be random, consecutive, array-based, and / or determinable by the user.

[0055] FIG. 3 shows a flowchart of at least a portion of a method 300 for forming a three-dimensional structure. FIG. 3 shows a portion of the image processing process 140 described in connection with the methods of FIGS. 1A-2B. The image processing process 140 can be further described in connection with FIGS. 4A-15B. Moreover, the method 300 can include one or more or all of the features already described in connection with FIGS. 1A-2B.

[0056] The method 300 can include performing the image processing process 140 until one layer set is formed and then forming a further layer set. (For example, it can be performed by a processor implementing an image processing algorithm) The image processing process 140 can include or incorporate some of the processes 301-312 (e.g., consecutive processes) described in FIG. 3.

[0057] The image processing process 140 can include a step 301 of processing the coordinate information of the structure (e.g., G-code or numerical code) to determine an ideal axis 302 based on the coordinate information of the structure. For example, the image processing process 140 can include a step 302 of determining the ideal axis of the strand or a step of extracting the ideal axis. Optionally, the ideal axis can be obtained by extraction of the structure by a numerical code (e.g., G-code) that can be represented by a starting point P s and an end point P e and.

[0058] After performing the determination 302 of the ideal axis, the image processing process 140 can include a step 303 of fitting the determined ideal axis to the acquired data. If the acquired data is an image, since the image of the strand can be taken at a specific position and a specific scale, the extracted ideal axis can be fitted onto the captured image. The fitting can be performed, for example, by a step 303 of rotating the image so that the extracted ideal axis can be fitted on the image. This can be done since the position of the camera and the size of the pixels are known.

[0059] After scaling and positioning the ideal axis as in 303, the image processing process 140 may include step 302 of evaluating whether the axis is on the strand to avoid detecting incorrect edges. If there is no strand to be printed, if it is printed at an incorrect position, or if it is severely curled, it may be evaluated that the ideal axis is not on the strand. Edges that do not belong to the strand to be monitored may be detected in this case.

[0060] Figures 4A and 4B show examples of strands with incorrect positioning.

[0061] In Figure 4A, the strand is not located on the ideal axis 415 because the strand is rotated with respect to the ideal axis.

[0062] In Figure 4B, the strand is not located on the ideal axis 415 because the strand is translated parallel to the ideal axis. Edges 416, 417 immediately adjacent to the ideal axis are not the actual edges of the strand and thus cannot be used, for example, to measure the dimensions of the strand.

[0063] Method 300 may further include step 304 of converting the image to grayscale so that these cases can be recognized. The characteristics of the grayscale values along the ideal axis can be analyzed at the start of image processing. These characteristics may be the difference in the average grayscale values of consecutive axis portions and the complexity of the function of the grayscale values along the axis. Both represent the color uniformity of the image along the axis. If the strand is not located on the ideal axis, the dimensions of the strand cannot be measured based on the ideal axis. In the present application, a misaligned strand may be defined as a defect on the print. A misaligned strand may mean that there is no ideal axis within any part of the strand where it is expected to be located.

[0064] Method 300 may include step 305 of detecting the edges of the strand after converting the image to grayscale when the ideal axis of the strand segment is disposed on the actually printed axis (e.g., when the ideal axis is between the edges of the actually printed strand). For example, step 305 of edge detection may include rotating and / or translating the image of the actual strand to obtain the required position adjustment when it is found that the ideal axis is between the edges of the actually printed strand. For example, in the case of FIGS. 4A and 4B, since the ideal axis is not between the edges of the actually printed strand, edge detection is not performed any further. FIG. 8 shows a case where it is found that the ideal axis is between the edges of the actually printed strand. However, the actual axis is not position-adjusted in the required direction. The required position adjustment may depend on the image sensing and image processing techniques used. Since the angle of the ideal axis is known from the G-code, the image can be rotated so that the ideal axis is position-adjusted and / or horizontal (e.g., parallel to a direction such as the x-direction of the axis system). Optionally, the required position adjustment may be such that it positions the edges of the strand (e.g., positions them parallel and / or horizontal to the axis system). The term position-adjusted herein refers to the case where it is made horizontal or parallel in the required direction, such as with respect to a predetermined axis. It can be understood that the term position-adjusted used in connection with the required position adjustment may refer to any angle predefined or required based on the required manufacturing tolerances.

[0065] After rotating the image, the direction of the ideal axis can be defined as the x-direction, and the y-direction can be orthogonal to the ideal axis. The edge detection process 305 can be performed after rotating the image and converting the image to grayscale. The edge detection process 305 may include performing a gradient detection process (V1 * ), or an artificial neural network process such as a convolutional neural network (CNN) (V2 * ).

[0066] Gradient-based edge detector (V1 * ) can be sensitive to frequently occurring features such as edges. By taking the first derivative of an image, features can be detected as large points. For this reason, the image can be convolved with a filter kernel. A simple gradient filter can be the Prewitt operator. The kernels for filtering in the x and y directions can be described as follows.

Number

[0067] Noise in an image can also have high-frequency characteristics and thus can be detected as an edge. To avoid detecting false edges, the image can be preprocessed before the edge detection operation. For example, the image can be blurred with a filter to attenuate the noise. An example of such an operator is the median filter, which can be an edge protection filter and can attenuate the noise while preserving the edges. A local non-linear contour protection filter can be performed before edge detection to improve the operation result of the gradient-based Prewitt filter. The protection filter can remove noise that may be detected as an edge. If the direction of the edge is known, the filter may be sufficient to detect the edge in only one direction. Compared to the size of the image, the detected edges may be relatively long. For this reason, the kernel of the Prewitt operator can be expanded horizontally.

Number

[0068] For example, the kernel can be expanded to a 3×9 kernel with nine elements per row. The filter size can be adapted according to the edge length. Next, the filtered image is scanned along the positive and negative y-directions in the selection range. The y-value of the ideal axis can be used as the starting point. If an edge with a specific thickness is found, the y-value can be noted or detected and saved. This process can be performed for each x-value of the ideal axis.

[0069] Alternatively, a convolutional neural network (CNN) (V2 * ) can be used to process multi-array input data and classify images. Compared with a fully-connected multi-layer network, the pre-processing of the input data can be integrated into the network, and the network itself can perform the feature extraction process. With a CNN, the local group correlations in the image can be considered. Since different regions of the image can share weights in a CNN, features can be detected regardless of their location.

[0070] Figures 5A and 5B show images of training data for a CNN, and the CNN can be trained while being supervised by a portion of the training images obtained during the printing process of the previous three-dimensional structure. The training images show a portion of a strand or an edge. Accordingly, the training images can be labeled as "strand" or "edge". After training the network, the trained network can be used to classify the edges of unknown images obtained from one or more locations determined for the current three-dimensional structure in the monitor.

[0071] Figure 5A shows an example of training data that can be labeled or identified as an edge.

[0072] Figure 5B shows an example of a training image that can be labeled as a strand.

[0073] Figure 6 shows a diagram of edge detection 631 using a CNN.

[0074] An image obtained based on the formed part can be cut along the y-direction to generate a cut image 633. For example, the entire image taken during printing at one of the determined locations can be divided (or cut) into small images. These small cut images 633 are sent to the trained CNN 631, and the trained CNN can classify the images according to whether the images contain edges or not.

[0075] Similar to the gradient-based method, edges can also be searched using the ideal axis as the starting point. At a constant (same) x coordinate, starting from the y coordinate (y0) of the ideal axis, the entire image can be cut into small images 633 along the positive y-direction and the negative y-direction. FIG. 6 shows a plurality of images 633 cut corresponding to the same x coordinate, and each of the cut images can correspond to a different y coordinate. The CNN 631 can analyze the regions within the cut images. As an example, starting from the cut image I(x, y0) of the ideal axis, the CNN can determine that there are no edges in the cut image I(x, y0). Among the images 633 cut in the positive y-direction, the cut image I(x, y k ) can be the first cut image (or the first region) capable of identifying an edge. Therefore, the position of the center point of the first region (or the cut image I(x, y k )) containing the edge can be saved. Similarly, among the images 633 cut in the negative y-direction, the cut image I(x, y n ) can be the first cut image (or the first region) capable of identifying an edge. Therefore, the position of the center point of the first region (or the cut image I(x, y n )) containing the edge can be saved. This process can be repeated for each x coordinate of the ideal axis. In both cases of gradient-based edge detection and CNN-based edge detection, the results of edge detection can be all the edge points found above and below the ideal axis.

[0076] The method 300 (shown in FIGS. 7A-7C) may further include a step 306 of identifying (or finding) one or more main edges, and a step 307 of performing an edge clustering process after the step 305 of identifying (or finding) the edges, since the found edge or edge points may include correct edges and / or incorrect edges.

[0077] FIG. 7A shows edge points 721 found by an edge clustering process that include correct edges and incorrect edges. Incorrect edges 721 may be found due to reflections, noise, and adjacent strands. For this reason, it may be necessary to further analyze the found edges.

[0078] FIG. 7B shows the result of determining (306) the main edges using edge clustering. To determine the main edges of the upper and lower boundaries, the angle between adjacent edge points can be calculated. The points can be clustered in the order of their x values as long as the angle between two adjacent points is below a threshold. If the threshold is exceeded, a new cluster can be created. The main edge can be defined as the largest cluster, assuming that the longest continuous edge belongs to the actual boundary of the strand. The main edge can be the entire boundary of the strand or a segment of the boundary. The result of this clustering is illustrated in FIG. 7B. The longest cluster of continuous edge points can form the main edges 722, e.g., the main upper edge and the main lower edge. Points 1 and 2 can be the starting points of a further (or subsequent) clustering process.

[0079] Figure 7C shows the clustered upper edge 723 and a further (or final) clustering process 307 that identifies the clustered lower edge 723. In the further clustering process 307, the remaining edge points can be collected starting from points outside the main edge. This clustering is based on the nearest neighbor method. The nearest neighbor method can add to a cluster if the difference in x and y between neighbors is below a set threshold. By this method, outliers can be removed by using the main edge as a reference during the clustering process. Thus, the actual edges 723 of the target strand (e.g., the actual upper edge and the actual lower edge) can be identified.

[0080] Method 300 may further include step 308 of approximating the boundaries of the strand after performing edge clustering processes 306, 307. The edge function can be approximated by minimizing the root mean square error using the clustered edge points (step 308). The characteristics of the edge function may be analyzed to detect incorrect approximations.

[0081] When both boundaries of the strand are found, the diameter of the strand can be measured. To make the measurement process appropriate, a measurement axis of the strand (e.g., the actual axis) can be determined. The actual axis of the strand can be calculated by the approximated edge function. The ideal axis of the strand may result in incorrect measurements because the strand may be displaced by translation and rotation during the manufacturing process.

[0082] Figure 8 shows that a method of measuring with respect to the ideal axis 415 can result in incorrect measurements by rotating the printed strand by an angle α with respect to the ideal axis (e.g., the angle between the ideal axis 415 and the actual axis 834 is α, and optionally 0° < α < 180°). The distance between the upper edge and the lower edge can be measured orthogonal to the reference axis, and Figure 8 shows that the error is the distance l measured with the actual axis as a reference realand the distance l measured with respect to an ideal axis as a reference ideal which indicates what can occur due to a difference therebetween.

[0083] Method 300 may further include step 309 of rotating the image for measurement after step 308 of approximating the boundary of the strand. For example, method 300 can rotate the actual axis to conform to the required direction.

[0084] Method 300 may further include step 311 of measuring the dimensions of the strand. For example, the diameter of the strand can be measured by calculating the distance between the upper edge 723 and the lower edge 723 in a direction orthogonal to the actual axis (step 311).

[0085] Figs. 9A to 9D show various printable three-dimensional structures. Various structures including breast implants and cell culture meshes can be printed and monitored during the manufacturing process. The structures can be manufactured with various parameters. For example, the structures can be printed with various nozzle diameters, pore sizes, and various numbers of layers.

[0086] Fig. 9A shows a breast implant scaffold material (50 ml volume). At the bottom, the average pore diameter can be 6 mm. Since the pore diameter decreases as the height increases, the average pore diameter reaches 3 mm at the top. To avoid the formation of a barrier that hinders tissue growth, the layers can be printed with an offset with respect to the previous layer. With this technique, inclined channels and a characteristic overall shape can be formed. The outer boundary develops from a circular structure, and the inner structure includes only straight strands that intersect orthogonally. The structure can be printed with a nozzle diameter of 350 μm.

[0087] Fig. 9B shows a cell culture mesh with a strand width of 350 μm, a pore diameter of 700 μm, and an offset of the strand orientation of 36° for each layer. The structure can be printed with a nozzle diameter of 350 μm.

[0088] FIG. 9C shows a cell culture mesh with a strand width of 150 μm, a pore diameter of 400 μm, and an offset of 36° in the strand direction. The structure may be printed with a nozzle diameter of 150 μm.

[0089] FIG. 9D shows a cell culture mesh with a strand width of 150 μm, a pore diameter of 400 μm, and an offset of 90° in the strand direction. The structure may be printed with a nozzle diameter of 150 μm.

[0090] To obtain a processable image, the structure (in FIGS. 9A - 9D) can be illuminated from below using an LED light, which can be placed under the substrate on which the structure is printed. To automate the camera positioning, the numerical code of each structure can be changed to include information about one or more locations, as described above. The camera can be moved to the calculated position and paused there until an image is captured and saved by image capture software. As an example, a magnification of 40 can be used to measure the diameter of the strands of a 50 ml breast implant. As an example, 4 strands per layer can be monitored during the printing process. While printing the mesh, 10 strands per layer can be monitored using a magnification between 140 and 200 depending on the pore diameter of the mesh. In other words, the number of locations can be determined by the user but may vary depending on the strictness of the manufacturing requirements.

[0091] FIGS. 10A - 10F show images of a 50 ml breast scaffold (described in FIG. 9A) visually monitored using gradient - based edge detection.

[0092] FIG. 10A shows overlapping vertical strands in the third layer of the three - dimensional structure.

[0093] FIG. 10B shows horizontal strands in the fifteenth layer of the three - dimensional structure.

[0094] Figure 10C shows the horizontal strands overlapping in the 43rd layer of the three-dimensional structure.

[0095] Figure 10D shows the state where the horizontal strands adjacent to the boundary are missing in the 53rd layer of the three-dimensional structure.

[0096] Figure 10E shows the horizontal strands in the 75th layer of the three-dimensional structure.

[0097] Figure 10F shows the vertical strands in the 18th layer of the three-dimensional structure.

[0098] Gradient-based techniques can be used to detect strand boundaries. Figure 10A shows layers of different heights starting from the third layer. In the image of the third layer (Figure 10A), the structure of the paper piece placed under the substrate is visible, but the focus is lost due to the growth of the scaffolding material. Also, the lighting conditions can change as the structure grows. This is because the scaffolding material may scatter light depending on its structure. During the construction of the implant, strand overlap may occur in the print, which means that a portion of the strands in the current layer is on top of the strands in the previous layer. These layers are not physically connected and are separated by a gap between them, although they appear to be on top of each other in plan view. This phenomenon is occurring in Figures 10A and 10C. Nevertheless, the strand boundary (line 1023) can be correctly detected. In Figure 10D, the strands to be observed (e.g., expected to be observed based on the coordinate information of the structure) are missing. Some dragged strands can be seen in the background of this image. The axis evaluation process correctly indicates the error in the location of the axis, which means that the strands were not printed on their ideal axis. This misalignment can be recorded as a defect. In this image, the ideal axis is represented by line 1024. The images in Figures 10E and 10F were taken at even higher layers of the scaffolding material. In these layers, the pore size can be significantly smaller than that of the lower layers.

[0099] In the processed image, for example, the values of the minimum (min) or smallest diameter or width and the maximum (max) or largest diameter or width of the strands, and their locations, can be visualized and measured. Optionally, the average diameter or width (avg) in μm can be determined by measuring multiple diameters along the strand segment being monitored.

[0100] Figures 11A to 11F show images visually monitoring the strands in the mesh for cell culture depicted in FIGS. 9B to 9D. To detect the boundaries, a gradient-based method can be used. Since the structure to be measured is smaller compared to the observation of the scaffold material (in FIGS. 10A to 10F), a higher magnification may be set.

[0101] Figures 11A and 11B show strands with a diameter of 350 μm in the mesh, where the pore size is 700 μm and the direction of the strands changes by 36° for each layer. The strands of consecutive layers can intersect each other at an angle of 36°. The images were taken at a magnification of 140.

[0102] The following four images visualize the measurement of the strands in the mesh having a smaller structure. The required pore size of the mesh was 400 μm, and the required diameter of the strands was 150 μm. These images were taken at a magnification of 200.

[0103] Figures 11C and 11D show strands with a diameter of 150 μm in the mesh, where the pore size is 400 μm and the direction of the strands changes by 36° for each layer.

[0104] Figures 11E and 11F show strands with a diameter of 150 μm in the mesh, where the pore size is 400 μm and the direction of the strands changes by 90° for each layer.

[0105] Figures 12A to 12F show the results of image processing. In the figures, using a CNN, edges or boundaries were detected. Figures 12A to 12F are raw images taken during the manufacturing process of 50 ml of formwork material and show the same respective strands as in the case of the gradient-based method described in relation to Figures 10A to 10F. Some deviation can be observed between the measurements using gradient-based edge detection and the measurements using CNN-based edge detection, but similar results were obtained. These deviations can be on the scale up to 30 μm. Due to these deviations, the measured minimum, maximum, and average diameters may differ between the CNN-based edge detection method and the gradient-based method. These differences may be due to the size of the cropped image that can be sent to the CNN network. Since the location of the edge on the classified image is not known exactly, the deviation may be of the same order as the size of the cropped image. This phenomenon becomes particularly apparent by observing the approximated edge function ripped in Figure 12F. In most cases, the edges were detected much earlier (as soon as they appeared at the image boundary), so the deviations that occurred were smaller than the size of the cropped image. Depending on the size of the image sent to the network, the CNN-based technique shown in this specification may seem to be less accurate than the gradient-based method that achieved a resolution of 5 μm at a magnification of 40. However, if the resolution needs to be improved, smaller parts of the image can be sent to the CNN as needed. To achieve this, the CNN network can be trained with similarly smaller images. Further, if necessary, the resolution can be further improved by increasing the training data and / or training the network using specific images that may contain or include edges that were previously indistinguishable by the network.

[0106] In CNN-based edge detection, edges can be found or observed along the entire ideal axis. This can be observed by comparing the images in FIGS. 10A to 10F and FIGS. 12A to 12F. When comparing, in gradient-based edge detection, edges may not be found or observed in the first and last parts of the ideal axis. However, these edges can be detected by the CNN. For example, in FIG. 10C, the edges are not found in the last part of the ideal axis by the gradient-based method, while on the other hand, these edges can be detected using the CNN (see FIG. 12C). Considering the processing time, the gradient-based technique may be able to process faster than the CNN-based method. However, whether the consumption time of the CNN-based method becomes long depends greatly on its application. On the other hand, a trained CNN can better classify the edges in low-contrast regions than the gradient-based method. Depending on the size of the images used in the CNN-based method, the gradient-based method may be more accurate. Scaffolding materials with a pore size as large as 6 mm and as small as 400 μm may be monitorable. However, it is also possible to change the magnification of the digital microscope to compensate for the difference in the size of the features.

[0107] In data analysis, the average diameter of the strands and the average value of the total average diameter of one layer may be variables that are meaningful to measure. Insufficient strand diameter may limit the mechanical properties of the scaffolding material. On the other hand, overly thick strands can cure the print, reduce the pore size, and thus potentially limit the tissue growth process in medical applications. If the average diameter is not within the range of a predetermined tolerance, it may be necessary to adjust the printing parameters. Insufficient average diameter can be compensated for by increasing the material flow rate or decreasing the printing speed. Conversely, the diameter can be reduced by decreasing the flow rate or increasing the speed. Clogged nozzles that require cleaning can also cause insufficient diameter. Further important and measurable variables include the maximum (largest) diameter or width and the minimum (smallest) diameter or width of the strands and their locations, as well as the variation of the diameter along the strands. This information can be used to classify defects in some cases. If the average diameter is within the tolerance range but the diameter varies along the strands, it can be judged to be due to constriction phenomena caused by insufficient layer gaps or inappropriate temperature conditions. In this case, the printing conditions can be improved by widening the layer gaps, extending the cooling process, and / or adjusting the ambient temperature. Furthermore, the stability of the printing process can be evaluated. To test the stability, the standard deviation of the average diameter, minimum diameter, or maximum diameter of one layer or one or all of the entire printed part can be determined. If the standard deviation is small, the printing process is consistent and stable, which can be a good starting point for optimizing the printing parameters. On the other hand, a large standard deviation can indicate an inconsistent process. In such cases, it may be effective to make the process reliable and consistent before optimizing the dimensional accuracy of the strand diameter. Additionally, unmeasurable diameters may indicate geometric problems such as strand misalignment. Reasons for such errors include warping, clogged nozzles, dragged strands, or inaccurate operation of the printer axis.

[0108] Some of these evaluation methods can be explained in relation to the results extracted from the monitoring process of the breast implant scaffold material described in connection with FIG. 9A. Three various analysis levels are illustrated. At the first level, the entire printed portion can be verified, taking into account all the measured diameters. Thereby, the overall quality of the printed object can be evaluated. At the second level, an evaluation regarding the layer that can be used to determine whether a single layer meets the quality requirements can be performed. At the third level, the dimensions of a single strand can be analyzed. Using this evaluation level, defects can be classified and the printing parameters can be systematically optimized.

[0109] Figures 13A - 13C show the first - level evaluation using histograms that measure the diameters (minimum, maximum, and average diameters) of all the strands of the breast scaffold material. To evaluate the quality of the entire printed portion or all the printed strands at a specific time, the frequency of the measured diameters can be displayed in a histogram. This is an appropriate technique because the histogram can show all ranges of the measured diameters. Moreover, the variation in the measured diameters can be visualized. This type of evaluation needs to be performed not only after the printing is completed but also during the printing process.

[0110] Figure 13A shows a histogram representing the variation of the average diameter of the strands for an example of a breast scaffold material against the frequency. As shown in Figure 13A, the difference between the ideal diameter of 350 μm and the average value (μ, unit: μm) of all the measured average diameters does not exceed a difference of 10 μm. The standard deviation (σ = 9.5) is low.

[0111] Figure 13B shows a histogram representing the variation of the maximum (extreme or largest) diameter of the strands for an example of a breast scaffold material against the frequency.

[0112] Figure 13C shows a histogram representing the variation of the minimum (extreme or smallest) diameter of the strands for an example of a breast scaffold material against the frequency.

[0113] Regarding the measured maximum and minimum diameters, the standard deviations can be high (σ = 21.8, σ = 19.8). The allowed tolerance determines the frequency of strand diameters below the allowed minimum diameter or above the allowed maximum diameter.

[0114] Figures 14A - 14C show a second - level evaluation such as the evaluation regarding the layer of the 50 - ml scaffolding material. Figures 14A - 14C show the measured diameters displayed for the layer, Figure 14A shows the minimum diameter, Figure 14B shows the maximum diameter, and Figure 14C shows the average diameter. With these graphs, it is easy to determine which layer has diameters outside a specific tolerance range. Depending on the quality requirements, a layer can be labeled as a defective layer if a specific number of diameters within this layer are outside the required tolerance range. In the example shown in Figures 14A - 14C, the user can determine various thresholds. For example, the minimum diameter of the strand should not be less than 225 μm. The maximum diameter should not exceed 475 μm. The average diameter should be within the range of 310 μm - 390 μm. Strands with diameters outside the tolerance range are considered defective and can be determined to be in a defective process state. The determination of whether a single layer or the entire structure meets the required quality depends on the desired quality. The diameter of a strand outside the allowable range is determined not to meet specific quality requirements.

[0115] For defect classification and optimization, it can be useful to evaluate the variation of the diameter along a single strand.

[0116] Figures 15A - 15B show a third - level evaluation that can perform an evaluation of the strand direction of the strands of the 50 - ml scaffolding material.

[0117] Figure 15A shows the visual result of the measurement process.

[0118] Figure 15B shows a graph of all deviations from the ideal strand diameter (350 μm) along the length of strand L. The tolerance range may be set, for example, at ±50 μm. Measurements that are less than the ideal diameter or are ±50 μm are determined to meet specific quality requirements, while deviations exceeding ±50 μm may be determined to be non-conforming. The characteristics of the displayed graph can vary depending on the type of defect, and this can sometimes allow the defect to be classified. If the cause of the classified defect is known, this classification can be used to systematically adjust the print parameters. In Figure 15B, the location of important portions of the strand being inspected can be identified. The most important portion of the strand may have the smallest diameter, and this minimum diameter deviates by more than 50 μm from the ideal diameter. Since the minimum width of the strand is located outside its center, this defect can be classified as a constriction phenomenon, which may be caused by an inappropriate cooling process or an insufficient layer gap.

[0119] Thus, it can be understood that whether to stop the printing process (e.g., perform the print stop step 180) or not (e.g., instead perform the step 170 of adapting the print parameters) can be based on the severity required by the evaluation process. For example, the decision 180 to stop printing depends on the frequency of the deviation of the determined parameter value from the ideal parameter value and / or the magnitude of the deviation of the determined parameter value from the ideal parameter value. These deviation thresholds can be determined by the user.

[0120] Figure 16 shows a diagram of configuration 150 for forming a three-dimensional structure. Configuration 150 can be configured to perform or execute the methods described in connection with Figures 1A - 15B.

[0121] Configuration 150 includes an apparatus 101 for forming a three-dimensional structure 104 based on coordinate information of the structure regarding the three-dimensional structure 104. Configuration 150 further includes a movable sensing device 102. Configuration 150 further includes a processor 103. The processor 103 is configured to determine one or more locations for positioning the sensing device 102 based on the coordinate information of the structure related to the three-dimensional structure to be formed, and to control the positioning of the movable sensing device 102 to the one or more locations.

[0122] The apparatus 101 for forming a three-dimensional (3D) structure may be a stereolithography (SLA) apparatus, a digital light processing (DLP) apparatus, a fused deposition modeling (FDM) apparatus, a selective laser sintering (SLS) apparatus, a selective laser melting (SLM) apparatus, an electron beam melting (EBM) apparatus, a laminated object manufacturing (LOM) apparatus, a binder jetting (BJ) apparatus, and / or a material jetting (MJ) apparatus.

[0123] The sensing device 102 may be, for example, an imaging device, a camera, or a three-dimensional scanning device, a thermal camera, a digital microscope, an AE sensor, and / or a CT device.

[0124] The processor 103 may be any computer or machine capable of executing instructions of a computer-readable storage medium. Such a computer-readable storage medium may include instructions that, when executed by the computer (or processor) 103, cause the computer 103 to execute the methods described in connection with FIGS. 1A-15B.

[0125] Configuration 150 may be a vision-based monitoring system for performing the methods described in connection with FIGS. 1A-15B. The sensing device 102 (e.g., a commercially available digital microscope) may be configured to acquire layer-wise data (e.g., an image). The acquired data may be processed to detect defects. After one (or each) layer is printed, the sensing device 102 may be positioned at one or more locations by the axis of the printer. Depending on the field of view and the size of the monitored feature, one or more images may be taken and processed.

[0126] For the automation of camera positioning and image capture, it may be necessary to know the image capture time and the optimal camera position. This information is derived from a numerical code that includes all the operations of all printers used to print a three-dimensional scaffold structure. The numerical code may be sorted to determine candidate camera positions and modified with commands for positioning the camera. Thus, the processor 103 can be configured to process the layout of the three-dimensional structure to be formed and / or the coordinate information of the structure related to the internal structure and / or the external structure. For example, the coordinate information of the structure may include CAD-based information or numerical code information. Based on the coordinate information of the structure, the processor 103 can be configured to generate numerical code commands adapted to form the three-dimensional structure 104. The adapted numerical code commands may include the coordinate information of the structure related to the three-dimensional structure to be formed and the information related to the step of positioning the sensing device at one or more locations.

[0127] Since the device 101, the movable sensing device 102, and the processor 103 can be interconnected, a part of the device 101 whose operation is controlled by the numerical code command can be a part of the device 101 (e.g., a print head, a dispenser, a nozzle head, an extruder, or a laser) that forms a three-dimensional structure at a location defined and selected by the coordinate information of the structure of the numerical code command.

[0128] The processor 103 can be further configured to determine the process state based on the acquired data related to at least a part of the portion of the three-dimensional structure to be formed. The acquired data can be acquired or generated by the sensing device 102 at the determined location among one or more locations.

[0129] In some embodiments, it can be understood that the configuration 150 includes a forming device 101 for forming a three-dimensional structure including a plurality of layers. The configuration 150 includes a movable sensing device 102. The configuration 150 includes a processor configured as follows. (a) Determine at least one location for positioning a sensing device for one layer set consisting of a plurality of layers, (b) Control the formation of the layer set by a device based on the coordinate information of the structure regarding the three-dimensional structure, (c) Control the positioning of a movable sensing device to at least one location associated with the layer set from when the layer set is formed until a further layer set is formed, (d) Determine a process state based on the acquired data of the formed layer set, the acquired data being acquired by the sensing device at at least one location associated with the formed layer set.

[0130] The various embodiments described herein may relate to a vision-based system for detecting in-situ defects during an additive manufacturing process of a porous scaffold material. The various embodiments described herein provide a complete process of a defect detection system based on the measurement of the diameter of printed strands. Using a digital microscope with adjustable magnification as a sensor and for vision-based data processing, small defects in the structure can be directly detected. The various concepts may include automated camera positioning using numerical codes and the processing of captured images in which the diameter of the printed strands is measured. The image processing can be implemented using at least two different edge detection methods, such as gradient-based edge detection or CNN-based edge detection.

[0131] The various embodiments and examples can be integrated in an existing printing environment that can detect and measure the strands to be monitored. The various embodiments can provide one or more options for analyzing the measurement values. For example, based on quality control requirements, the quality of a single strand, layer, or the entire structure can be inspected.

[0132] Various embodiments may be implemented at different levels of automation. It may be suitable for manual quality control. For example, a user may manually place the structure under a microscope and take an image of the top layer. Computer-executable instructions may be configured to draw by specifying the axis of the strand measured manually with the start and end points of the axis. The strand may be measurable by computer-executable instructions and the output may be displayed and saved. In this case, no change to the numerical code is required. Optionally or alternatively, automated control of the monitoring process may be performed with automated camera positioning, image acquisition, and processing, but without closed-loop control. In this specification, the user can access the print status at any time during or after the manufacturing process and intervene if necessary. A fully automated closed-loop system may also be implemented. In this case, the print can be made to continue as long as the quality requirements are met, otherwise it can be stopped or the print parameters can be automatically adjusted for compensation.

[0133] Various embodiments can be used to monitor a three-dimensional structure by any three-dimensional printing method. In particular, various embodiments can monitor the quality of strands printed by FDM onto a porous scaffold material during the manufacturing process. The step of measuring the diameter of the strands can be suitable for classifying and locating defects and their causes. A manufacturer may be able to analyze the quality of the printed object on-site. FDM printed porous structures, such as cell culture meshes and custom-made implants, have become popular in applications of tissue engineering. Regarding the production of the scaffold material, PCL may be used as a material because of its excellent rheology and viscoelasticity. PCL has excellent thermal stability, which makes it suitable for the FDM printing process. In particular, sophisticated quality control processes are required for applications with a medical background. Although sensors have become an essential component in manufacturing, FDM, one of the most popular 3D printing technologies, lacks a monitoring system and an open-loop control system. The lack of sensing in the AM process makes on-site quality control very difficult. Post-print quality control is also complicated because it is difficult or impossible to access the internal structure of the printed part.

[0134] The methods and configurations described in relation to FIGS. 1A - 16 may relate to a vision - based inspection process for on - site monitoring of additively - manufactured porous scaffolds. The methods and configurations may include, for example, evaluating the quality of the internal structure of the printed scaffold by monitoring geometric dimensions. A commercially available digital microscope with adjustable magnification can be used to capture selected internal regions of the printed object. The captured images can be analyzed by computer vision methods. At least one of two different edge - detection techniques can be used to detect boundaries in selected regions of the image. The first edge - detection technique may include or be a gradient - based approach. The second edge - detection technique may be based on the classification of a convolutional neural network (CNN). The methods and configurations may be based on the sorting and correction of numerical codes and / or may include using automated camera positioning that can be controlled.

[0135] Structures such as (e.g., biodegradable) breast implants and / or meshes for cell culture can be monitored during their manufacturing process. The method can be integrated at various stages of the quality control process. The method can be part of a closed - loop system that automatically compensates for defects or can stop the manufacturing process if a critical defect occurs.

[0136] Sensors can be used in the manufacturing process to improve process reliability, reproducibility, and automatic control. Desirable effects of monitoring the printing process are detecting errors during printing, avoiding further defects by adjusting printing parameters, automatically stopping the printing process if the detected errors are critical, and detecting errors that are difficult to monitor in the manufactured product (e.g., defects in internal features). This effectively reduces printing time, decreases material waste, and improves the quality control and documentation of the manufacturing process.

[0137] The various embodiments described herein may relate to on-site monitoring of a printed (e.g., FDM printed) porous structure. Existing technologies mainly relate to solid print objects as they monitor the dimensions of the outer boundaries rather than the internal structure. Generally, existing technologies have not reached a resolution of less than 1 millimeter. Since the quality and mechanical properties of a porous structure are scientifically influenced by the appearance of the internal structure, the various embodiments described herein relate to monitoring systems, methods, and configurations capable of monitoring the internal structure as well. In order to obtain a meaningful evaluation of print quality, measurements in the sub-millimeter range can be made.

[0138] In some embodiments, instead of using CAD data, a numerical code can be used as a reference to obtain the ideal structure of the object to be printed. In such embodiments, no CAD program is used. Since the operation of all machines and the diameter of the print strands are known from the code, it is possible to obtain a description of the ideal appearance using the numerical code.

[0139] The various embodiments described herein may relate to monitoring the diameter of the strands. The quality characteristics of the manufactured structure are the dimensions of the printed strands. A high-quality product may be characterized by a strand diameter that is constant at a specified value. The consistency of the strand diameter may also indicate that the printing process is stable and reliable. On the other hand, an inconsistent strand diameter and deviation from the required strand diameter suggest that the manufacturing process is unstable. Therefore, monitoring the strand diameter during printing enables not only reliable on-site quality control but also optimization of the manufacturing process. For tissue engineering, the strand diameter is an important variable as variations in the pore size, which can be caused by diameter deviations, can prevent proper tissue growth and change the mechanical properties.

[0140] The present invention is further characterized by the following items. Item 1: A method for forming a three-dimensional structure, the method comprising: determining one or more locations for positioning a sensing device based on the coordinate information of the structure regarding the three-dimensional structure to be formed; forming a part of the three-dimensional structure based on the coordinate information of the structure; positioning the sensing device at one of the one or more locations.

[0141] Item 2: The method according to Item 1, wherein the coordinate information of the structure includes information regarding the layout of the three-dimensional structure.

[0142] Item 3: The method according to Item 1 or 2, wherein the coordinate information of the structure includes information regarding the internal structure of the three-dimensional structure.

[0143] Item 4: The method according to any one of Items 2 or 3, wherein the coordinate information of the structure includes a tool path command for controlling a forming device for forming the three-dimensional structure.

[0144] Item 5: The method according to any one of Items 1 to 4, further comprising determining a process state based on acquired data regarding at least a part of the portion of the three-dimensional structure to be formed, wherein the acquired data is acquired by the sensing device at the determined location.

[0145] Item 6: The method according to Item 5, wherein the step of determining the process state includes determining a parameter value of a structural feature of the portion of the three-dimensional structure to be formed based on the acquired data.

[0146] Item 7: The method according to Item 6, wherein the parameter value is at least one of length, width or diameter, height, roughness, color, uniformity, thickness and tilt angle of the structural feature of the portion of the three-dimensional structure to be formed.

[0147] Item 8: The method according to Item 6 or 7, wherein the structural feature includes a strand segment of the internal structure of the three-dimensional structure.

[0148] Item 9: The method according to item 8, wherein the parameter value of the above feature is determined by detecting the edge of the strand segment.

[0149] Item 10: The method according to item 9, wherein the edge of the above strand segment is detected by performing an artificial neural network process or a gradient-based detection process.

[0150] Item 11: The step of determining the above process state includes determining an ideal axis of a strand segment of a part of the three-dimensional structure based on the coordinate information of the structure, and determining an actual axis of the strand segment of the formed part based on the acquired data of the formed part, and determining the process state based on a comparison between the above ideal axis and the above actual axis, and the method according to any one of items 5 to 9.

[0151] Item 12: The method according to item 11, including determining a defective process state when the difference between the above ideal axis and the above actual axis exceeds a threshold value.

[0152] Item 13: The step of determining the above process state includes comparing an ideal parameter value of a strand segment of a part of the three-dimensional structure with a determined parameter value of the strand segment of the formed part, wherein the determined parameter value of the above strand segment is determined based on the acquired data of at least a part of the formed part of the three-dimensional structure, wherein the ideal parameter value of the above strand segment is determined based on at least one of the coordinate information of the structure and the input value, and the method according to any one of items 5 to 12.

[0153] Item 14: The method according to any one of claims 5 to 13, comprising the step of determining a defective process state when the difference between the ideal parameter value and the determined parameter value exceeds a threshold value.

[0154] Item 15: The step of determining the process state comprises the step of determining a plurality of parameter values associated with a plurality of strand segments of the formed part, and the step of determining a statistical parameter of the plurality of parameter values, and the method according to any one of items 5 to 14, comprising the step of determining a defective process state when the difference between the statistical parameter and the comparison parameter exceeds a threshold value.

[0155] Item 16: The method according to any one of items 1 to 15, further comprising the step of adapting process parameters for forming a three-dimensional structure based on the determined process state.

[0156] Item 17: A part of the three-dimensional structure comprises one layer set of a plurality of layer sets of the three-dimensional structure, the method according to any one of items 1 to 16.

[0157] Item 18: For one layer set composed of a plurality of layers, the method according to item 17, further comprising the step of determining each group of locations for positioning a sensing device, and the step of forming one layer set of the plurality of layer sets, and the step of sequentially positioning the sensing device at the locations of the group of locations from the time of forming the layer set until creating the next layer set, and acquiring data of the formed layer set at each location.

[0158] Item 19: The layer set includes at least a first sub-layer and at least a second sub-layer, The method according to item 17 or 18, wherein the location among the plurality of locations is determined based on an intersection between a first sublayer and a second sublayer.

[0159] Item 20: The method according to item 18 or 19, wherein the location among the plurality of locations is between two intersections of a first sublayer and a second sublayer.

[0160] Item 21: The method according to any one of items 1 to 20, wherein one or more locations include several locations, and some of those locations are random, consecutive, array-based, or determinable by a user.

[0161] Item 22: The method repeatedly and alternately forming one layer set of the plurality of layer sets, and sequentially positioning a sensing device at the locations of the location group of the layer set from when the layer set is formed until a further layer set is formed, the method according to any one of items 17 to 21.

[0162] Item 23: A computer-readable storage medium that, when executed by a computer, includes instructions for the computer to execute the method according to any one of items 1 to 22.

[0163] Item 24: A configuration for forming a three-dimensional structure, the configuration including a forming device for forming a three-dimensional structure based on structure coordinate information regarding the three-dimensional structure, a movable sensing device, and a processor, wherein the processor determines one or more locations for positioning the sensing device based on structure coordinate information regarding the three-dimensional structure to be formed, and is configured to control the positioning of the movable sensing device to the one or more locations.

[0164] Item 25: The processor is further configured to generate a numerical code command adapted to form a three-dimensional structure, the adapted numerical code command including coordinate information of the structure regarding the three-dimensional structure to be formed and information regarding positioning a sensing device at one or more locations, according to the configuration of Item 24.

[0165] Item 26: It is further configured to determine a process state based on acquired data regarding at least a part of the portion to be formed of the three-dimensional structure, the acquired data being acquired by a sensing device at one of the one or more locations, according to the configuration of Item 24.

[0166] Item 27: A configuration for forming a three-dimensional structure, the configuration including a forming device for forming a three-dimensional structure including a plurality of layers, a movable sensing device, a processor, and the processor determines at least one location for positioning the sensing device for one set of layers composed of a plurality of layers, controls the formation of the set of layers by the device based on the coordinate information of the structure regarding the three-dimensional structure, controls the positioning of the movable sensing device to at least one location associated with the set of layers from the formation of the set of layers until the formation of a further set of layers, is configured to determine a process state based on the acquired data of the formed set of layers, the acquired data being acquired by the sensing device at the at least one location associated with the formed set of layers.

[0167] Although described in detail in the embodiments of the present invention, many obvious modifications are possible without departing from the specific spirit or scope. It should be understood that the present invention described by the appended claims is not limited by the specific details described in the above explanation.

Claims

1. A method for forming a three-dimensional structure, the method comprising: determining one or more locations for positioning a sensing device based on coordinate information of a structure regarding the layout of the three-dimensional structure to be formed; forming a portion of the three-dimensional structure based on the coordinate information of the structure; positioning the sensing device at one of the one or more locations; determining a process state based on acquisition data obtained by the sensing device at one of the one or more determined locations regarding at least a portion of the formed portion of the three-dimensional structure. A method for forming a three-dimensional structure, including these steps.

2. The method according to claim 1, wherein the coordinate information of the structure includes information regarding the internal structure of the three-dimensional structure.

3. The method according to claim 1 or 2, wherein the coordinate information of the structure includes a tool path command for controlling a forming device for forming the three-dimensional structure.

4. The method according to any one of claims 1 to 3, wherein the step of determining the process state includes determining a parameter value of a structural feature of the formed portion of the three-dimensional structure based on the acquisition data.

5. The method according to claim 4, wherein the parameter value is at least one of length, width or diameter, height, roughness, color, homogeneity, thickness and tilt angle of a structural feature of the formed portion of the three-dimensional structure.

6. The method according to claim 4, wherein the structural feature includes a strand segment of the internal structure of the three-dimensional structure.

7. The method according to claim 6, wherein the parameter value of the feature is determined by detecting an edge of the strand segment.

8. The method according to claim 7, wherein the edge of the strand segment is detected by performing an artificial neural network process or a tilt-based detection process.

9. The step of determining the process state includes: determining an ideal axis of a strand segment of the portion of the three-dimensional structure based on the coordinate information of the structure; determining an actual axis of a strand segment of the formed portion based on the acquisition data of the formed portion; determining the process state based on a comparison between the ideal axis and the actual axis. The method according to any one of claims 1 to 8, including these steps.

10. The method according to claim 9, comprising the step of determining a defective process state when the difference between the ideal axis and the actual axis exceeds a threshold value.

11. The step of determining the process state comprises: comparing the ideal parameter values of the strand segments of the part of the three-dimensional structure with the determined parameter values of the strand segments of the formed part, wherein the determined parameter values of the strand segments are determined based on the acquired data of at least a part of the formed part of the three-dimensional structure, The method according to any one of claims 1 to 10, wherein the ideal parameter values of the strand segments are determined based on at least one of the coordinate information and the input value of the structure.

12. The method according to claim 11, comprising the step of determining a defective process state when the difference between the ideal parameter value and the determined parameter value exceeds a threshold value.

13. The step of determining the process state comprises: determining a plurality of parameter values associated with a plurality of strand segments of the formed part; determining statistical parameters of the plurality of parameter values; The method according to any one of claims 1 to 12, comprising the step of determining a defective process state when the difference between the statistical parameter and the comparison parameter exceeds a threshold value.

14. The method according to any one of claims 1 to 13, further comprising the step of adapting process parameters for forming the three-dimensional structure based on the determined process state.

15. The method according to any one of claims 1 to 14, wherein the part of the three-dimensional structure comprises one layer set consisting of a plurality of layers of the three-dimensional structure.

16. determining a plurality of locations for positioning the sensing device for each one layer set consisting of the plurality of layers; forming one layer set of the plurality of layer sets; sequentially positioning the sensing device at the locations of the plurality of locations consisting of locations from the formation of the layer set until the formation of the next layer set, and acquiring data of the formed layer set at each location. The method according to claim 15, further comprising:

17. The layer set includes at least a first sub-layer and at least a second sub-layer, The method according to claim 16, wherein the location among the plurality of locations is determined based on an intersection between the first sub-layer and the second sub-layer.

18. The method according to claim 17, wherein the location among the plurality of locations is between two intersections of the first sub-layer and the second sub-layer.

19. The method according to any one of claims 1 to 18, wherein the one or more locations include several locations, and the several locations are random, continuous, array-based, or determinable by a user.

20. The method includes Repeatedly and alternately Forming one layer set among the plurality of layer sets; Sequentially positioning the sensing device at the locations of the location group of the layer set from forming the layer set until forming a further layer set. The method according to claim 15.

21. A configuration for forming a three-dimensional structure, the configuration including A forming device for forming a three-dimensional structure based on structure coordinate information regarding the three-dimensional structure; A movable sensing device; A processor, The processor Determines one or more locations for positioning the sensing device based on structure coordinate information regarding the three-dimensional structure to be formed; Controls the positioning of the movable sensing device to the one or more locations; A configuration for forming a three-dimensional structure, configured to determine a process state based on acquisition data obtained by the sensing device at one of the one or more locations regarding at least a part of the portion of the three-dimensional structure to be formed.

22. The processor is further configured to generate a numerical code command adapted for forming the three-dimensional structure, and the adapted numerical code command includes structure coordinate information regarding the three-dimensional structure to be formed and information regarding positioning the sensing device at the one or more locations. The configuration according to claim 21.

23. A configuration for forming a three-dimensional structure, the configuration including A forming device for forming a three-dimensional structure including a plurality of layers; A movable sensing device; A processor, For one layer set consisting of the plurality of layers, determine at least one location for positioning the sensing device, control the formation of the layer set by the device based on the coordinate information of the structure regarding the three-dimensional structure, control the positioning of the movable sensing device to at least one location associated with the layer set from when the layer set is formed until a further layer set is formed, a processor configured to determine a process state based on acquired data of the formed layer set, the acquired data acquired by the sensing device at at least one location associated with the formed layer set. A configuration including. A computer-readable storage medium including instructions that, when executed by the processor of the configuration according to any one of claims 21 to 23, cause the configuration to execute the method according to any one of claims 1 to 20. ​

Images