Machinability determination device, machinability learning device, machinability determination procedure, machinability learning procedure, machinability determination program and machinability learning program

Abstract

A machinability determination device (100) comprises a three-dimensional shape extraction unit (10) that extracts a three-dimensional shape from three-dimensional shape data representing a planned machining shape, a three-dimensional shape rotation unit (12) that rotates the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction, a depth map transformation unit (14) that generates a depth map in which information about the depth in the desired machining axis direction extracted from the three-dimensional shape data is mapped to a two-dimensional image obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction, and an inference model (16) built by machine learning using a learning machining axis direction instruction.a learning depth map extracted from three-dimensional shape learning data, and editability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and determining the suitability of machining the planned machining shape by inference using the desired machining axis and the depth map generated by the depth map transformation unit (14).

Description

AREA OF TECHNOLOGY

[0001] The present disclosure relates to a machinability determination device, a machinability learning device, a machinability determination method, a machinability learning method, a machinability determination program and a machinability learning program. STATE OF THE ART

[0002] In cutting operations, where a material (machining target) is cut using a cutting tool (hereinafter referred to as "tool"), such as a drill, milling cutter, or tool bit, is cut, there are cases where cutting with the tool is difficult depending on the shape of the structure. Furthermore, material expansion or tool deformation can occur due to heat generated during machining, and machining defects can also occur due to vibration or tool deformation.

[0003] Patent reference 1 discloses an electrode manufacturing process in which the suitability of electrode production by electrical discharge machining (EDM) is determined. Furthermore, non-patent reference 1 discloses a technology for determining the suitability of the cutting capability using a device that is trained by inputting voxel data representing a three-dimensional shape into a 3D convolutional neural network (CNN) model. REFERENCES ON THE STATE OF THE TECHNOLOGY PATENT REFERENCE

[0004] Patent reference 1: Publication of Japanese patent application no. 2010-105080. NON-PATENT REFERENCE

[0005] Non-Patent Reference 1: Computer Aided Geometric Design 62 (2018) 263-275. SUMMARY OF THE INVENTION TASK TO BE SOLVED BY THE INVENTION

[0006] However, in the technology described in patent reference 1, all rib lines forming the electrode shape are extracted from input three-dimensional electrode data and shape identification information. The suitability of the machining is then determined based on the electrode contour identified from the extracted rib lines. This results in an increased computational effort for extracting the electrode contour from the three-dimensional electrode data and similar tasks. Furthermore, the technology described in non-patent reference 1 requires a large 3D CNN model to process the voxel data as three-dimensional data, thus increasing both the computational load and the amount of data required for training.

[0007] The subject of the present disclosure is the provision of a machinability determination device, a machinability determination method and a machinability determination program for determining the suitability of cutting (whether cutting is possible or impossible) as well as a machinability learning device, a machinability learning method and a machinability learning program for building the machinability determination device by learning, in which the number of learning data and the computational load are limited. MEANS TO SOLVE THE PROBLEM

[0008] A machinability determination device in the present disclosure comprises a three-dimensional shape extraction unit for extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a three-dimensional shape rotation unit for rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction; a depth map transformation unit for generating a depth map in which information about the depth in the desired machining axis direction extracted from the three-dimensional shape data is mapped to a two-dimensional image obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction;and an inference unit built by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and determines the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map generated by the depth map transformation unit.

[0009] A machinability determination method to be performed by a computer, as disclosed herein, comprises a step of extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a step of rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction; a step of generating a depth map in which information about the depth in the desired machining axis direction extracted from the three-dimensional shape data is mapped to a two-dimensional image obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction;and a step that is built up by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and determines the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map.

[0010] A machinability determination program in the present disclosure comprises the machinability determination program causing a computer to perform: a step of extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a step of rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction; a step of generating a depth map in which information about the depth in the desired machining axis direction extracted from the three-dimensional shape data is mapped to a two-dimensional image obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction;and a step that is built up by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and determines the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map. IMPACT OF THE INVENTION

[0011] According to the present disclosure, it is possible to provide a machinability determination device, a machinability determination method, and a machinability determination program that determine the suitability of the cutting operation by means of learning, wherein the number of pieces of learning data and the computational load are limited by using a depth map from a machining axis direction, while also providing a machinability learning device, a machinability learning method, and a machinability learning program that perform the learning. BRIEF DESCRIPTION OF THE DRAWINGS Fig. Figure 1a is a functional configuration representation showing an editability determination device according to a first embodiment, and Fig. Figure 1b is a functional configuration representation showing an editability learning device according to the first embodiment. Fig. Figure 2 is a hardware configuration diagram showing the machinability determination device and the machinability learning device according to the first embodiment. Fig. 3a, Fig. 3b and Fig. 3c are explanatory illustrations that demonstrate the cutting process using a cutting machine such as a milling machine or a machining center. Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d and Fig. 4e are explanatory descriptions that show cases in which processing is not possible. Fig. Figure 5 is a flowchart showing an example of a determination process of the trained machinability determination device according to the first embodiment. Fig. Figure 6a is a schematic representation showing an example of a three-dimensional shape extracted from three-dimensional CAD data, and Fig. 6b shows an example of a depth map derived from the one in Fig. The three-dimensional shape shown in 6a is created. Fig. Figure 7a is a schematic representation showing an example of a case in which a tool accesses a material in a machining axis direction, and Fig. Figure 7b is an explanatory representation indicating that shape information on a processing surface can be represented by a depth map using orthographic projection. Fig. Figure 8 is a functional configuration representation showing an editability determination device according to a second embodiment. Fig. Figure 9a is an explanatory illustration showing an area where cutting in the present machining axis direction is difficult, and Fig. 9b is an explanatory illustration showing a case where processing becomes possible by changing the direction of the processing axis. Fig. Figure 10 is a flowchart showing an example of a determination process of the trained editability determination device according to the second embodiment. Fig. Figure 11a is a functional configuration representation showing an editability determining device according to a third embodiment, and Fig. Figure 11b is a functional configuration representation showing an editability learning device according to the third embodiment. Fig. Figure 12 is a functional configuration representation showing an editability learning device according to a fourth embodiment. Fig. Figure 13a is a schematic representation showing an example of a depth map, and Fig. Figure 13b is a schematic representation showing an example of a heatmap that displays an inadequate editing area as editability information according to the depth map in Fig. 13a shows. Fig. Figure 14 is a functional configuration representation showing a machinability determining device according to a fifth embodiment. Fig. 15a is an explanatory illustration showing a planned processing method for a material form, Fig. 15b is an explanatory representation of a case in which a machining surface machining axis direction recording unit records the machining axis directions in a machining surface list generated by a machining surface division unit, Fig. 15c is an explanatory presentation of a case in which the in Fig. The material shape shown in drawing 15a is rotated 90° to the left by a three-dimensional shape rotation unit, and Fig. Figure 15d is an explanatory representation of a case in which the machining surface machining axis direction recording unit records the machining axis directions in the machining surface list in the state in which the material shape is rotated by 90°. Fig. Figure 16 is a flowchart showing an example of a determination process of the trained editability determination device according to the fifth embodiment. MODE FOR EXECUTING THE INVENTION

[0012] A machinability determination device and a machinability learning device according to each embodiment are described below with reference to the drawings. The following embodiments are only examples, and it is possible to combine embodiments appropriately and to modify each embodiment appropriately. (First embodiment)

[0013] Fig. Figure 1a is a functional configuration representation showing a machinability determination device 100 according to a first embodiment. The machinability determination device 100 comprises a three-dimensional shape extraction unit 10, which extracts information about the shape of a material as a machining object after machining (planned machining shape) from input three-dimensional CAD data; a three-dimensional shape rotation unit 12, which aligns a surface to be machined (hereinafter referred to as the "machining surface") in the extracted shape information in a desired machining axis direction (e.g., Z-axis direction); and a depth map transformation unit 14, which generates a depth map with respect to the machining axis direction, as shown in Figure 1. Fig. Figure 6b shows that the shape information is transformed in which the machining area is oriented in the machining axis direction, and an inference model 16 is used to determine whether machining is possible with respect to the generated depth map and to output a result of the determination. As described later, the inference model 16 is built by training a mathematical model such as CNN through machine learning using learning machining axis direction instructions, learning depth maps extracted from the three-dimensional shape training data according to the learning machining axis direction instructions, and machinability information describing the suitability of actual machining previously performed according to the learning machining axis direction instruction and the three-dimensional shape training data that were previously prepared for training.

[0014] Fig. Figure 1b is a functional configuration representation showing an editability learning device 110 according to the first embodiment. The editability learning device 110 comprises a shape edit axis direction database (DB) 20, an editability information DB 24 in which the editability information is stored as training data, and a learning device 22 that updates the inference model 16 with the editability information, in addition to the three-dimensional shape extraction unit 10, the three-dimensional shape rotation unit 12, the depth map transformation unit 14, and the previously described inference model 16.

[0015] The form machining axis direction DB 20 includes a form information DB 20A, which stores three-dimensional CAD data as three-dimensional form learning data, and a machining axis direction instruction DB 20B, which stores instructions for machining axis directions as learning machining axis direction instructions.

[0016] The data stored in Shape Information DB 20A, Machining Axis Direction Instruction DB 20B, and Machinability Information DB 24 each represent previous machining case data. For example, the three-dimensional CAD data stored in Shape Information DB 20A is training data of three-dimensional shapes actually used in machining, and the machining axis direction instructions stored in Machining Axis Direction Instruction DB 20B are also training data of machining axis direction instructions actually used in machining. Furthermore, Machinability Information DB 24 stores the suitability of a machining result using the three-dimensional CAD data stored in Shape Information DB 20A and the machining axis direction instructions stored in Machining Axis Direction Instruction DB 20B.The machining information stored in the machining information DB 24 for learning purposes corresponds to the three-dimensional CAD data stored in the shape information DB 20A and the machining axis direction instructions stored in the machining axis direction instruction DB 20B. In the first embodiment, the learning device 22 evaluates the inference model 16 by comparing the result of the assessment of the learning depth map, obtained using the three-dimensional CAD data and the machining axis direction instructions as example data for previous machining cases, with the machining information corresponding to the three-dimensional CAD data and the machining axis direction instructions used by the inference model 16 for the assessment, and updating the inference model 16.

[0017] Fig. Figure 2 is a hardware configuration diagram showing the editability determination device 100 and the editability learning device 110 according to the first embodiment. The editability determination device 100 and the editability learning device 110 are each configured by a computer with a processor 210, a storage device 220, and an input / output interface 230. The editability determination device 100 can be configured by a plurality of computers.

[0018] The Processor 210 is an IC (Integrated Circuit) that performs arithmetic processing. Specifically, the Processor 210 could be a CPU (Central Processing Unit), a DSP (Digital Signal Processor), a GPU (Graphics Processing Unit), or similar device. The Processor 210 is capable of rendering three-dimensional CAD data, changing the orientation of the rendered three-dimensional CAD data in any desired machining axis direction, and functioning as the Three-Dimensional Shape Extraction Unit 10, the Three-Dimensional Shape Rotation Unit 12, and the Depth Map Transformation Unit 14, as described above, by executing a program that generates the depth map from the three-dimensional CAD data.Furthermore, the processor 210 functions as the aforementioned learning device 22 by executing a program that performs machine learning using training data and, as the result of the machine learning, constructs the inference model 16, which determines machinability. In the first embodiment, the processor 210, by running a machinability learning program, functions as the three-dimensional shape extraction unit 10, the three-dimensional shape rotation unit 12, the depth map transformation unit 14, and the learning device 22, which updates the inference model 16. In the first embodiment, after the machine learning, the processor 210 further functions as the three-dimensional shape extraction unit 10, the three-dimensional shape rotation unit 12, the depth map transformation unit 14, and the inference model 16 by running a machinability determination program.Furthermore, the editability determination program and the editability learning program are provided via a recording medium, for example, which has recorded these programs.

[0019] The memory device 220 is configured by a volatile memory device such as RAM (Random Access Memory) or a non-volatile memory device such as ROM (Read Only Memory), HDD (Hard Disk Drive) or flash memory.

[0020] The input / output interface 230 is a port to which an input device 300 and an output device 310 are connected. Specifically, the input / output interface 230 is, for example, a USB (Universal Serial Bus) port, an IEEE 1394 port, a Thunderbolt port, or similar, and also includes a communication interface for Ethernet or similar. The input device 300 is a touch panel, a keyboard, a mouse, or similar. The output device 310 is a display, a printer, or similar. At the time of learning, the form editing axis direction DB 20 and the editability information DB 24 are each connected to the input / output interface 230 described above. Both the form editing axis direction DB 20 and the editability information DB 24 can be located in the storage device 220.

[0021] Fig. 3a, Fig. 3b and Fig. 3c are explanatory illustrations that demonstrate the cutting process using a cutting machine such as a milling machine or a machining center. Fig. Figure 3a is an explanatory illustration showing planar machining of a planar cutting surface 42 of a material 40 with a tool 30, such as a simple milling cutter. In planar machining, the cutting surface 42 is cut on a plane of the material 40 that is orthogonal to the machining axis direction with the tool 30, which moves in a direction parallel to the plane, as indicated by the arrows.

[0022] Fig. Figure 3b is an explanatory illustration showing the machining of a side surface of material 40 with a tool 32, such as an angle cutter. During side surface cutting, a cutting surface 44 is cut with the tool 32, which moves in the direction of the arrow.

[0023] Fig. Figure 3C is an explanatory illustration showing a drilling and pocketing operation of the cutting of material 40 in the machining axis direction with a tool 34, e.g., an end mill. In the drilling and pocketing operation, a cutting surface 46 of the material 40 is cut with the tool 34, which is fed in the machining axis direction indicated by the arrow.

[0024] In the first embodiment, the suitability of machining processes other than those described above is also determined. For example, it is also possible to determine the feasibility of machining with regard to the formation of a three-dimensional shape using electrical discharge machining (EDM), in which the material is processed by an arc discharge between an electrode and the material.

[0025] Fig. 4a, Fig. 4b, Fig. 4c, Fig. 4d and Fig. 4e are explanatory descriptions that show cases in which processing is not possible. Fig. Figure 4a shows a case of tool engagement where a different area of ​​the tool 34 engages the material 40 as a cutting edge, and Fig. Figure 4b shows a case where tool 34 is not suitable for the machining operation of material 40. In the Fig. 4a and Fig. In the cases shown in 4b, the problems can be solved according to rules in the machining process, such as changing the tool 34 to a tool suitable for the situation, and thus it is possible to assess the machinability even without using the machinability determination device 100 according to the first embodiment.

[0026] Fig. Figure 4c shows a case in which precise cutting of the cutting surface 44 becomes impossible due to vibration of the tool 32 or similar, and Fig. Figure 4d shows a case in which the tool 34 bends and the material 40 cannot be cut correctly. Fig. 4e shows a case where the processing method of material 40 is complicated and it is difficult to make an assessment based on the processing rules. In the Fig. 4c, Fig. 4d and Fig. In the cases described in section 4e, it is not possible to solve the problems based on the rules of machining, such as changing the tool, so that a skilled person has usually assessed the suitability of the machining based on their own experience. In the first embodiment, the suitability of the machining in cases such as those described in the Fig. 4c, Fig. 4d and Fig. 4e are shown, determined by the trained machinability determination device 100, without referring to the specialist.

[0027] Fig. Figure 5 is a flowchart showing an example of a determination process for the trained machinability determination unit 100 according to the first embodiment. In step S101, the three-dimensional CAD data is entered as the shape information into the three-dimensional shape extraction unit 10.

[0028] In step S102, the three-dimensional shape extraction unit 10 extracts the three-dimensional shape by rendering the input three-dimensional CAD data. Fig. Figure 6a is a schematic representation showing an example of the three-dimensional shape extracted from the three-dimensional CAD data.

[0029] In step S103, the three-dimensional shape rotation unit 12 rotates the three-dimensional shape extracted in step S102 according to the machining axis direction instruction. In the first embodiment, the machining surface of the material 40 is, for example, oriented in the Z-axis direction. Specifically, the machining surface of the material 40 is positioned so that it points directly in the direction of the Z-axis. The machining axis direction instruction can either be set beforehand on the Z-axis or entered in step S101.

[0030] In step S104, the depth map transformation unit 14 generates a depth map of the three-dimensional shape, whose processing surface is aligned in the direction of the Z-axis. Fig. 6b shows an example of the depth map derived from the in Fig. The three-dimensional shape shown in Figure 6a is generated. The depth map consists of two-dimensional image data generated by orthographic projection of the three-dimensional shape, whose processing surface is aligned in the Z-axis direction onto a plane orthogonal to the Z-axis. In the depth map, information about the depth in the Z-axis direction is assigned to the data of a two-dimensional image representing the processing surface. The depth information in the Z-axis direction is extracted from the three-dimensional CAD data as the input shape information. Fig. 6b is the information about the depth of the hue or the brightness. In the depth map, a shallow area is represented by a light hue, e.g., a yellow color, and a deep area is represented by a dark hue, e.g., a dark blue color. If depth is represented by a single color such as gray, a shallow area will be represented with high brightness and a deep area with low brightness.

[0031] In a cutting machine, such as a milling machine, the tool 32 approaches the material 40 in the machining axis direction (the Z-axis direction) to perform the machining, as shown in Fig. Figure 7a shows that the shape information of the machining surface, viewed in the Z-axis direction, is important for determining machinability, while the shape information of the side or back surfaces of material 40, which are not the machining surface, is not required in this case. Since the shape information on the machining surface can be represented by the depth map using orthographic projection, as shown by the arrows in Figure 7a, the shape information on the machining surface can be represented by the depth map using orthographic projection. Fig. As stated in 7b, the editability can be determined using the depth map, which is two-dimensional data instead of data representing the three-dimensional shape.

[0032] In step S105, the inference model 16 derives the machinability. As mentioned earlier, in the first embodiment, the inference model 16 is trained using the three-dimensional CAD data, the machining axis direction instructions, and the machinability information corresponding to the three-dimensional CAD data and the machining axis direction instructions, respectively. The trained inference model 16 performs an inference based on the suitability of machining in the machinability information used for training and outputs a result of the inference.

[0033] In step S106, inference model 16 determines whether the result of the inference output in step S105 indicates that processing is possible or not. The process continues to step S107 if the result of the determination in step S106 indicates that processing is possible, or the process continues to step S108 if the result of the determination in step S106 indicates that processing is not possible.

[0034] In step S107, inference model 16 outputs an affirmative editability determination to output device 310 and terminates the process. In step S108, inference model 16 outputs a non-editability determination to output device 310 and terminates the process.

[0035] As described above, in the first embodiment, the suitability of machining is determined by using the depth map as two-dimensional data generated from the three-dimensional shape. Preparing the data as two-dimensional data reduces the data size compared to the three-dimensional shape, thereby limiting the number of training data sets and the computational load at the time of training the inference model 16. Furthermore, the computational load for machining determination on the trained machining determination device 100 can be limited.

[0036] When implementing inference model 16 for determining the machinability of cutting operations, a model into which voxel data is input, as in conventional technology, must handle data representing the three-dimensional shape and therefore requires significant computational effort and a large amount of training data. However, in the first embodiment, the computational effort and training data can be reduced by inputting the depth map, which is two-dimensional data, into the model. Furthermore, with the same amount of training data, a higher determination accuracy can be achieved than with the conventional technique because the number of parameters in the neural network is smaller. (Second embodiment)

[0037] The following describes a machinability determination device 120 according to a second embodiment. The machinability determination device 120 according to the one described in Fig. The second embodiment shown in Figure 8 differs from the first embodiment in that the machinability determination device 120 comprises a machining axis direction adjustment unit 26, which adjusts the machining axis direction depending on the machinability determination output by the inference model 16, and a three-dimensional shape rotation unit 28, which rotates the three-dimensional shape into the machining axis direction adjusted by the machining axis direction adjustment unit 26. The other components, however, are the same as in the first embodiment; therefore, these components, which are identical to those in the first embodiment, are provided with the same reference numerals as in the first embodiment, and a detailed description thereof is omitted.Furthermore, the hardware configuration in the second embodiment is the same as in the first embodiment, so a detailed description is omitted. However, in the second embodiment, the processor 210 functions as the three-dimensional shape extraction unit 10, the three-dimensional shape rotation unit 28, the depth map transformation unit 14, and the inference model 16, while also functioning as the machining axis direction adjustment unit 26.

[0038] In the first embodiment, the machinability is determined by placing the machining axis on the Z-axis. However, at the time of determining the machinability, there may be an area where cutting in the given machining axis direction is difficult, such as a [missing information - likely a specific area or feature]. Fig. 9a undercut shape 38. Also in the one shown in Fig. In the case shown in 9a, processing is possible depending on the direction of the processing axis, as shown in Fig. 9b. In the second embodiment, when design data is provided, verification with respect to each machining axis direction is performed by repeating the machinability determination while the machining axis direction (the direction of material 40) is automatically changed. As shown in Fig. As shown in Figure 8, the machining axis direction is changed, for example, by rotating the material 40 in a vertical direction 50 or a horizontal direction 52. Some cutting machines are capable of changing the machining axis while the material 40 is fixed; however, even in such machines, machining control can be achieved simply and quickly by changing the machining surface through rotation of the material 40 rather than by changing the machining axis. In the second embodiment and a fifth embodiment, which will be described later, the actual machining axis on which the tool is set is fixed, for example, in the direction of the Z-axis, and the three-dimensional shape is rotated so that the machining surface coincides with the machining surface as in the case where the machining axis is changed from the Z-axis.

[0039] Fig. Figure 10 is a flowchart showing an example of the determination process of the trained machinability determination device 120 according to the second embodiment. The Fig. The flowchart shown in section 10 differs from the one in Fig. The flowchart shown in Figure 5 in the first embodiment is modified by including step S203 instead of step S103, and by including step S204 for determining whether the change in the machining axis direction has been made with respect to all directions, and step S205 for changing the machining axis direction instruction. The other steps, however, are the same as in the first embodiment, so these identical steps are provided with the same reference numerals as in the first embodiment, and a detailed description of them is omitted.

[0040] In step S203, the three-dimensional shape rotation unit 28 rotates the three-dimensional shape extracted in step S102 according to the machining axis direction instruction. In the second embodiment, the machining axis direction instruction is set to the Z-axis beforehand in the first determination process. The machining axis direction instruction can be entered at the stage of step S101. As described later, in the second embodiment, the machining axis direction adjustment unit 26 changes the machining axis direction instruction if the result of the determination indicates that machining with the existing machining axis direction instruction is not possible, and in step S203, the three-dimensional shape is rotated according to the machining axis direction instruction after the change.

[0041] If the result of the determination in step S106 indicates that machining is not possible (if the result of the determination is non-machinability), the machining axis direction adjustment unit 26 determines in step S204 whether machinability has been checked with respect to all machining axis directions. If the material 40 is considered a rectangular solid, there are a total of six possible machining axis direction instructions. Among these six possible machining axis direction instructions, the machining axis direction adjustment unit 26 registers the Z-axis that was set in the first determination and the machining axis direction instruction that was changed in the later step S205 in the memory device 220.In step S204, the machining axis direction adjustment unit 26 determines, using the storage device 220, whether the machinability has been checked with respect to all machining axis directions or not.

[0042] If step S204 determines that machinability has been checked with respect to all machining axis directions, the process continues to step S108. In step S108, the non-machinability determination is output to output device 310, and the process is terminated analogously to the first embodiment.

[0043] If step S204 determines that machinability has not been checked with respect to all machining axis directions, the process continues to step S205. In step S205, the machining axis direction adjustment unit 26 changes the machining axis direction instruction and enters the modified machining axis direction instruction into the three-dimensional form rotation unit 28.

[0044] In step S203, the three-dimensional shape is rotated according to the modified machining axis direction instruction. Specifically, the three-dimensional shape is rotated such that, when the three-dimensional shape is machined in the machining axis direction specified by the machining axis direction instruction input from the machining axis direction adjustment unit 26, the new machining surface is directly opposite a desired machining axis direction (the Z-axis direction in the second embodiment). In the following steps, the depth map transformation unit 14 generates a new depth map by orthographically projecting the new machining surface onto a plane orthogonal to the Z-axis direction, and the inference model 16 determines the suitability of machining the planned shape using the new depth map.If the result of the determination in step S106 indicates that processing is possible, the affirmative machinability determination in step S107 is output to output device 310 and the process is terminated.

[0045] In the second embodiment, as described above, verification with respect to each machining axis direction is performed when design data is provided by repeating the machinability determination while automatically changing the machining axis direction. Even if the material has a shape where machinability changes depending on the machining axis direction, it is thus possible to determine whether a component shape represented by the design data is non-machinable at a given angle, since the machinability determination device is configured to make the determination with respect to all machining axis directions. Furthermore, in the second embodiment, it is also possible to determine the machinability for the shapes of the side and back surfaces, for which information is lost when the three-dimensional shape is converted into the depth map. (Third embodiment)

[0046] The following describes a machinability determination device 130 and a machinability learning device 140 according to a third embodiment. The machinability learning device 140 according to the embodiment described in Fig. The third embodiment shown in Figure 11b differs from the setup of the first embodiment in that the learning device 22 updates an inference model 58 by comparing the result of the assessment of the depth map, which is obtained using the three-dimensional CAD data and the machining axis direction instructions as example data for past machining cases, the machining axis direction specified by the machining axis direction instruction, and information about the properties of the material 40, tool information (tool type, tool material, tool diameter, tool length, etc.), and a cutting parameter (tool feed rate, tool rotation speed, etc.).), which are stored in an editing information DB 20C, which is contained in a form editing axis direction editing information DB 48, using the inference model 58 with the editing information stored in the editability information DB 24. Furthermore, the editability determination device 130 differs according to the one in . Fig. The third embodiment shown in Figure 11a differs from the setup of the first embodiment in that the trained inference model 58 determines the depth map, which is obtained using the three-dimensional CAD data and the machining axis direction instructions, the machining axis direction specified by the machining axis direction instruction, and the machining information, including information on the properties of the material 40, the tool information, and the cutting parameters. However, the other components are the same as in the first embodiment, so these components, which are identical to those in the first embodiment, are provided with the same reference numerals as in the first embodiment, and a detailed description thereof is omitted.Furthermore, the hardware configuration in the third embodiment is the same as the hardware configuration in the first embodiment, so a detailed description of it is omitted.

[0047] The properties of the material 40, the tool information, and the cutting parameters each have a significant influence on determining its machinability. For example, if a suitable tool is not used for a difficult-to-cut material such as stainless steel, or if the tool feed rate or rotational speed is unsuitable for the difficult-to-cut material, there is a risk that the cut surface 44 cannot be cut correctly due to vibrations of the tool 32 or similar factors, as shown in Fig. 4c shown, or there is a risk that the tool 34 will bend and the material 40 cannot be cut correctly, as shown in Fig. Shown in 4D.

[0048] Even if the design data are the same, as described above, the suitability of the cutting operation changes depending on the information about the properties of the material 40, the tool information, and the cutting parameters. In the machinability determination device 130 and the machinability learning device 140 according to the third embodiment, the machinability determination is carried out by adding at least some of the information from the information about the properties of the material 40, the tool information, and the cutting parameter to the input in the inference model 58.

[0049] As described above, according to the third embodiment, a correct determination of feasibility can be made even if the machinability changes not only depending on the design data, but also on the properties of the material 40, the tool and the cutting parameter. (Fourth embodiment)

[0050] The following describes a machinability learning device 150 according to a fourth embodiment. The machinability learning device 150 according to the one described in Fig. The fourth embodiment shown in Figure 12 differs from the setup of the first embodiment in that a heatmap representing an inadequate processing area is stored in an editability information database 56, and a learning device 54 performs a model update of the inference model 16 by using the heatmap as training data. However, the other components are the same as in the first embodiment; therefore, these components, which are identical to those in the first embodiment, are provided with the same reference numerals as in the first embodiment, and a detailed description of them is omitted.Furthermore, a machinability determination device, constructed by the machinability learning device 150, is set up in the same way as the machinability determination device 100 according to the first embodiment, except that the inference model 16 outputs a determination result that identifies the inadequate machining area by means of learning to update the inference model 16 by using the heatmap representing the inadequate machining area as the machinability information, and therefore a detailed description of it is omitted. Moreover, the hardware configuration in the fourth embodiment is the same as the hardware configuration in the first embodiment, so a detailed description of it is also omitted.

[0051] Fig. Figure 13a is a schematic representation showing an example of the depth map, and Fig. Figure 13b is a schematic representation showing an example of the heatmap, which represents the inadequate editing area as the editability information according to the depth map in Fig. 13a shows.

[0052] In the Fig. In the depth map shown in Figure 13a, a sharp edge 60, described as a so-called pin corner, is located at a point corresponding to a recess. During cutting, the recess is machined with a tool such as an end mill, and since the tool, like an end mill, cuts the recess under rotation, a corner formed by cutting the material 40 is inevitably rounded, and it is impossible to reproduce a pin corner like edge 60.

[0053] In the fourth embodiment, the heatmap, which corresponds to the depth map, is used as training data for training the inference model 16. The heatmap is a map in which an area corresponding to the depth map is represented as in Fig. 13a corresponds to a grid-like structure with a predetermined cell size, and each cell is assigned a numerical value representing the possibility of editing. Fig. In 13b, for example, each white cell is an editable area 62 assigned a numerical value close to 0, and the black cell is a non-editable area 66 assigned a numerical value close to 1. Fig. In section 13b, each gray cell represents an intermediate range (64) with slight editing difficulties, to which a numerical value between 0 and 1 has been assigned. The numerical value assigned to each cell in the heatmap, which represents editability information, is determined based on the result of actual manual verification of whether editing is possible or not.

[0054] In the fourth embodiment, the learning device 54 evaluates the inference model 16 by comparing the result of the assessment of the depth map, which is obtained using the three-dimensional CAD data and the machining axis direction instructions as example data for previous machining cases, with the heatmap as the machinability information corresponding to the three-dimensional CAD data and the machining axis direction instructions used by the inference model 16 for the assessment, and updating the inference model 16.

[0055] The machinability determination device, produced by the learning process described above, outputs a heatmap as the result of the machinability determination. This heatmap digitizes the level of inadequate machining in relation to each cell as a numerical value. Consequently, according to the fourth embodiment, it is possible to precisely identify which area in the three-dimensional shape or depth map represents inadequate machining. (Fifth embodiment)

[0056] The following describes a machinability determination device 160 according to a fifth embodiment. The machinability determination device 160 according to the fifth embodiment, which is described in Fig. The device shown in Figure 14 differs from the device in the first embodiment in that the machinability determination device 160 comprises a machining surface division unit 70, which generates a machining surface list by listing all machining surfaces of the material 40 from the three-dimensional shape extracted by the three-dimensional shape extraction unit 10; a machining surface machining axis direction recording unit 74, which receives an input of the machining surface list generated by the machining surface division unit 70 and records the machining axis direction of each machining surface that is judged to be machinable in the machinability determination output by the inference model 16 in the machining surface list; and a termination determination unit 76, which performs a termination determination.if the affirmative machinability determination has been made for all machining surfaces described in the machining surface list, or if the machinability has already been verified in all of the machining axis directions, and the machinability determination is output with respect to the three-dimensional shape and the machining axis direction of each machining surface, and a machining axis direction adjustment unit 78 is outputting a different instruction of machining axis directions than the machining axis directions for which the machinability has already been verified, if the termination determination unit 76 determines that not all of the machining axis directions have been verified, and a three-dimensional shape rotation unit 72 is outputting a three-dimensional shape rotation unit that rotates the three-dimensional shape into the machining axis direction set by the machining axis direction adjustment unit 78,The other components are, however, the same as in the first embodiment, so these components, which are the same as in the first embodiment, are provided with the same reference numerals as in the first embodiment, and a detailed description of them is omitted. Furthermore, the hardware configuration in the fifth embodiment is the same as the hardware configuration in the first embodiment, so a detailed description of it is omitted. In the fifth embodiment, however, the processor 210 operates as the three-dimensional shape extraction unit 10, the three-dimensional shape rotation unit 72, the depth map transformation unit 14, and the inference model 16, while it also operates as the machining surface division unit 70 and the machining surface machining axis direction recording unit 74.The termination determination unit 76 and the processing axis direction adjustment unit 78 are operating.

[0057] Fig. Figure 15a is an explanatory illustration showing a planned machining shape 82 for a material shape 80. The machining surfaces (1), (2), (3), (4), (5), (6) and (7) were set with respect to the planned machining shape 82; however, in the Fig. In the state shown in 15a, the workable surfaces 84 and the workable surfaces (5), (6) and (7) are not workable.

[0058] Fig. Figure 15b is an explanatory representation of a case in which the machining surface machining axis direction recording unit 74 has recorded the machining axis directions in the machining surface list generated by the machining surface division unit 70. In the Fig. In the case shown in 15a, the machining of the machining surfaces (1) to (4) is possible, so that in the machining surface list, for example, 0° is specified for the Z-axis direction as their machining axis directions. In the case shown in Fig. In the case shown in 15a, however, it is not possible to process the processing areas (5) to (7), so that their processing axis directions are empty in the processing area list.

[0059] Fig. 15c is an explanatory presentation of a case in which the in Fig. Figure 15a shows that the material shape 80 is rotated 90° to the left by the three-dimensional shape rotation unit 72. This 90° rotation of the material shape 80 makes it possible to machine the machining surfaces (5) to (7), which was impossible before the rotation.

[0060] Fig. Figure 15d is an explanatory illustration for a case in which the machining surface machining axis direction recording unit 74 has recorded the machining axis directions in the machining surface list in the state after the 90° rotation of the material form 80. As in Fig. As shown in 15c, in the state after the 90° rotation of the material form 80, the machining of the machining surfaces (5) to (7) is possible, and thus +90° is described as the rotation angle in relation to the Z-axis as its machining axis directions in the machining surface list.

[0061] Fig. Figure 16 is a flowchart showing an example of the determination process of the trained editability determination device 160 according to the fifth embodiment. The Fig. The flowchart shown in section 16 differs from the one in Fig. The flowchart shown in Figure 5 in the first embodiment is modified by including step S302 instead of step S103 in the first embodiment and by including step S301 of generating a machining surface list by extracting the machining surfaces from the three-dimensional shape extracted by the three-dimensional shape extraction unit 10, step S304 in which the machining surface machining axis direction recording unit 74 records the machining surfaces determined to be machinable and the machining axis directions in the machining surface list, step S305 in which the termination determination unit 76 assesses whether all machining surfaces have been determined to be machinable or not, step S306 of outputting the affirmative machinability determination and the machining axis direction of each machining surface, and step S307 of assessing,whether the change in the machining axis direction has been made for all directions or not, and step S308 of changing the machining axis direction instruction. However, the other steps are the same as in the first embodiment, so these steps, which are the same as in the first embodiment, are provided with the same reference numerals as in the first embodiment and a detailed description of them is omitted.

[0062] In step S301, the machining area division unit 70 generates a machining area list, which is displayed in Fig. 15b or Fig. 15d is shown by extracting the machining surfaces from the three-dimensional shape extracted by the Three-Dimensional Shape Extraction Unit 10 in step S102.

[0063] In step S302, the three-dimensional shape rotation unit 72 rotates the three-dimensional shape extracted in step S102 according to the machining axis direction instruction. In the fifth embodiment, the machining axis direction instruction is set to the Z-axis beforehand in the first determination process. The machining axis direction instruction can be entered at the stage of step S101. As described later, in the fifth embodiment, the machining axis direction adjustment unit 78 modifies the machining axis direction instruction if the result of the determination is not affirmative machinability with the existing machining axis direction instruction, and in step S302, the three-dimensional shape is rotated according to the machining axis direction instruction after the modification.

[0064] In step S304, the machining surface-machining axis direction recording unit 74 records the machining surfaces and machining axis directions determined to be machinable in the machining surface list, based on the machinability determination issued by the inference model 16 in step S105.

[0065] In step S305, the termination determination unit 76 assesses whether all of the machining areas have been designated as machinable. Termination determination unit 76 assesses that all of the machining areas have been designated as machinable if all fields in the Machining Axis Direction column of the machining area list are filled with information about a significant angle, such as 0° or +90°, as shown, for example, in Fig. As shown in 15d. In step S305, the process is forwarded to step S306 if it is determined that all processing surfaces have been identified as machinable, or the process is forwarded to step S307 if it is determined that not all processing surfaces have been identified as machinable.

[0066] In step S306, the final affirmative machinability determination regarding the three-dimensional shape and the machining axis direction of each machining surface is output to output device 310, and the process is terminated. The information output in step S306 regarding the affirmative machinability determination and the machining axis direction of each machining surface is, for example, the information shown in Fig. 15d shown list of machining areas, in which the machining axis direction corresponding to each machining area is described.

[0067] If, in step S305, it is determined that not all machining surfaces have been identified as machinable, the termination determination unit 76 assesses in step S307 whether the machining axis direction change has been implemented for all directions. If the material 40 is considered a rectangular solid, there are a total of six possible machining axis direction instructions. Similar to the second embodiment, among these six possible machining axis direction instructions, the machining axis direction adjustment unit 78 registers the Z-axis set in the first determination and the machining axis direction instruction changed in the later step S308 in the memory device 220.In step S307, the termination determination unit 76 assesses, with reference to the storage device 220, whether the machinability has been checked with respect to all machining axis directions. Alternatively, it is also possible to provide the machining area list specifically with a column for the machining axis change history and to have the termination determination unit 76 assess, based on this column, whether the machinability has been checked with respect to all machining axis directions.

[0068] If, in step S307, it is determined that the change in the machining axis direction has been made for all directions, the process is forwarded to step S108. In step S108, the non-machinability determination is output to the output device 310, and the process is terminated analogously to the first embodiment.

[0069] If, in step S307, it is determined that the change in the machining axis direction has not been applied to all directions, the process proceeds to step S308. In step S308, the machining axis direction adjustment unit 78 modifies the machining axis direction instruction and inputs the modified machining axis direction instruction into the three-dimensional shape rotation unit 72.

[0070] In step S302, the three-dimensional shape rotation unit 72 rotates the three-dimensional shape according to the modified machining axis direction instruction. In the following steps, the depth map transformation and the machinability inference are performed, and if all machining surfaces are determined to be machinable in step S305, the machinability determination and the machining axis direction of each machining surface are output to the output device 310 in step S306, and the process is terminated.

[0071] Even if, as described above, according to the fifth embodiment, similar to the second embodiment, the material has a shape with machinability that changes depending on the machining axis direction, it can be determined whether the component shape represented by the design data is not machinable at which angle, since the machinability determination device is configured to make the determination with respect to all machining axis directions. By clarifying not only the machinability but also from which direction each machining surface can be machined, a setup process at the time of actual machining can be facilitated.

[0072] Furthermore, an inference unit in the claims corresponds to the inference model 16 or 58 described in the detailed description of the invention. REFERENCE MARK LIST

[0073] 10: Three-dimensional shape extraction unit, 12: Three-dimensional shape rotation unit, 14: Depth map transformation unit, 16: Inference model, 20A: Shape information DB, 20B: Machining axis direction instruction DB, 20C: Machining information DB, 22: Learning device, 24: Machinability information DB, 26: Machining axis direction adjustment unit, 30, 32, 34: Tool, 40: Material, 54: Learning device, 56: Machinability information DB, 58: Inference model, 70: Machining surface division unit, 72: Three-dimensional shape rotation unit, 74: Machining surface machining axis direction recording unit, 76: Completion determination unit, 78: Machining axis direction adjustment unit, 100: Editability determination device, 110: Editability learning device, 120: Editability determination device, 130: Editability determination device, 140: Editability learning device, 150: Editability learning device,160: Machinability determination device. QUOTES INCLUDED IN THE DESCRIPTION

[0000] This list of documents cited by the applicant was automatically generated and is included solely for the reader's convenience. The list is not part of the German patent or utility model application. The DPMA accepts no liability for any errors or omissions. Cited patent literature

[0000] JP 2010-105080

[0004] Cited non-patent literature

[0000] Computer Aided Geometric Design 62 (2018) 263-275

[0005]

Claims

[1] Machinability determination device, comprising: a three-dimensional shape extraction unit for extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a three-dimensional shape rotation unit for rotating the three-dimensional shape so that a machining surface of the three-dimensional shape is opposite a desired machining axis direction; a depth map transformation unit for generating a depth map in which information about the depth extracted from the three-dimensional shape data in the desired machining axis direction is assigned to a two-dimensional image obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction; and an inference unit constructed by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and determining the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map generated by the depth map transformation unit. [2] Machinability determination device according to claim 1, further comprising a machining axis direction adjustment unit for issuing an instruction for a machining axis direction that differs from the desired machining axis direction when the inference unit issues a non-machinability determination in the inference with respect to the machining surface that is opposite the desired machining axis direction, wherein The three-dimensional shape rotation unit rotates the three-dimensional shape so that a new machining surface is opposite the desired machining axis direction when the three-dimensional shape is machined in the machining axis direction specified by the machining axis direction instruction entered by the machining axis direction adjustment unit. The depth map transformation unit creates a new depth map by orthographically projecting the new editing surface onto a plane that is orthogonal to the desired editing axis direction, and The inference unit determines the suitability of processing the planned processing method using the new depth map. [3] Machinability determination device according to claim 1, wherein the inference unit is constructed by machine learning using the learning machining axis direction instruction, the learning depth map extracted from the three-dimensional shape learning data according to the learning machining axis direction instruction, learning machining information consisting of tool information including a tool type, tool material, tool diameter and tool length, information about a material of a machining object and a cutting parameter including the feed rate of a tool and a rotational speed of the tool, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction, the three-dimensional shape learning data and the learning machining information.and the suitability of machining the planned machining shape by inference using the desired machining axis direction, the depth map generated by the depth map transformation unit, and newly entered machining information consisting of tool information including tool type, tool material, tool diameter and tool length, information about the material of the machining object, and cutting parameters including tool feed rate and tool rotation speed. [4] Machinability determination device according to claim 1, wherein The editability information is a heatmap that shows an area on the editing surface that is unsuitable for editing, and The inference unit is built using machine learning with the learning machining axis direction instruction, the learning depth map extracted from the three-dimensional shape learning data according to the learning machining axis direction instruction, and the heatmap describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data. The result of this determination identifies the unsuitable area of ​​machining in the planned machining shape by inference using the desired machining axis direction and the depth map generated by the depth map transformation unit. [5] Machinability determination device according to claim 1, further comprising: a machining surface division unit for generating a machining surface list by listing all machining surfaces of a machining object from the three-dimensional shape extracted by the three-dimensional shape extraction unit; a machining surface machining axis direction recording unit for recording the machining axis direction of each machining surface that is assessed as machinable in a machinability determination issued by the inference unit in the machining surface list; a termination determination unit for making a termination determination when an affirmative machinability determination has been made for all machining surfaces described in the machining surface list or when machinability in all machining axis directions has already been verified, and for outputting the machinability determination with regard to the three-dimensional shape and the machining axis direction of each machining surface; and a machining axis direction adjustment unit for issuing an instruction for machining axis directions other than those for which machinability has already been verified, when the termination determination unit assesses that machinability has not been verified in all machining axis directions, wherein The three-dimensional shape rotation unit rotates the three-dimensional shape so that a new machining surface is opposite the desired machining axis direction when the three-dimensional shape is machined in the machining axis direction specified by the machining axis direction instruction entered by the machining axis direction adjustment unit. The depth map transformation unit creates a new depth map by orthographically projecting the new editing surface onto a plane that is orthogonal to the desired editing axis direction, and The inference unit determines the suitability of processing the planned processing method using the new depth map. [6] Editability learning facility, comprehensive: a three-dimensional shape extraction unit for extracting a three-dimensional learning shape from input three-dimensional shape learning data; a three-dimensional shape rotation unit for rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a learning machining axis direction specified by an input learning machining axis direction instruction; a depth map transformation unit for generating a learning depth map, in which information about the depth in the learning processing axis direction extracted from the three-dimensional shape learning data is assigned to a two-dimensional image obtained by orthographically projecting the processing surface onto a plane orthogonal to the learning processing axis direction; and an inference unit that is updated by a learning facility that performs machine learning using the learning machining axis direction instruction, the learning depth map, and the machinability information that describes the suitability of the actual machining that has already been performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, wherein the machinability learning device builds the machinability determination device according to claim 1 by machine learning. [7] Editability learning facility, comprehensive: a three-dimensional shape extraction unit for extracting a three-dimensional learning shape from input three-dimensional shape learning data; a three-dimensional shape rotation unit for rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a learning machining axis direction specified by an input learning machining axis direction instruction; a depth map transformation unit for generating a learning depth map, in which information about the depth in the learning processing axis direction extracted from the three-dimensional shape learning data is assigned to a two-dimensional image obtained by orthographically projecting the processing surface onto a plane orthogonal to the learning processing axis direction; and an inference unit that is updated by a learning device that performs machine learning using the learning machining axis direction instruction, the learning depth map, and learning machining information consisting of tool information, including tool type, tool material, tool diameter, and tool length; information about the material of a machined object; and cutting parameters, including a tool feed rate and tool rotation speed; and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction, the three-dimensional shape learning data, and the learning machining information. wherein the machinability learning device builds the machinability determination device according to claim 3 by machine learning. [8] Editability learning facility, comprehensive: a three-dimensional shape extraction unit for extracting a three-dimensional learning shape from input three-dimensional shape learning data; a three-dimensional shape rotation unit for rotating the three-dimensional shape so that a processing surface of the three-dimensional shape is opposite a learning processing axis direction specified by an input learning processing axis direction instruction; a depth map transformation unit for generating a learning depth map, in which information about the depth in the learning processing axis direction extracted from the three-dimensional shape learning data is assigned to a two-dimensional image obtained by orthographically projecting the processing surface onto a plane orthogonal to the learning processing axis direction; and an inference unit that is updated by a learning facility performing machine learning using the learning machining axis direction instruction, the learning depth map, and a heatmap that describes an unsuitable area of ​​actual machining that has already been performed according to the learning machining axis direction instruction and the learned three-dimensional shape training data, wherein the machinability learning device builds the machinability determination device according to claim 4 by machine learning. [9] A processability determination procedure to be performed by a computer, comprising: a step of extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a step of rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction; a step in generating a depth map, in which information about the depth extracted from the three-dimensional shape data in the desired machining axis direction is assigned to a two-dimensional image, which is obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction; and a step that is built by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and which determines the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map. [10] Editability determination program that causes a computer to execute: a step of extracting a three-dimensional shape from three-dimensional shape data representing a planned machining shape; a step of rotating the three-dimensional shape so that a machining surface in the three-dimensional shape is opposite a desired machining axis direction; a step in generating a depth map, in which information about the depth extracted from the three-dimensional shape data in the desired machining axis direction is assigned to a two-dimensional image, which is obtained by orthographically projecting the machining surface onto a plane orthogonal to the desired machining axis direction; and a step that is built by machine learning using a learning machining axis direction instruction, a learning depth map extracted from three-dimensional shape learning data according to the learning machining axis direction instruction, and machinability information describing the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, and which determines the suitability of machining the planned machining shape by inference using the desired machining axis direction and the depth map. [11] Editability learning method to be performed by a computer, comprising: a step of extracting a three-dimensional learning shape from input three-dimensional shape learning data; a step of rotating the three-dimensional shape so that a processing surface in the three-dimensional shape is opposite a learning processing axis direction specified by an entered learning processing axis direction instruction; a step in generating a learning depth map, in which information about the depth in the learning processing axis direction extracted from the three-dimensional shape learning data is assigned to a two-dimensional image obtained by orthographically projecting the processing surface onto a plane orthogonal to the learning processing axis direction; and a step of updating an inference unit that determines the suitability of machining a planned machining shape, with a learning facility that performs machine learning using the learning machining axis direction instruction, the learning depth map, and the machinability information that describes the suitability of the actual machining already performed according to the learning machining axis direction instruction and the three-dimensional shape learning data, wherein the machinability learning method constructs the machinability determination device according to claim 1 by machine learning. [12] Editability learning program that builds the editability determination device according to claim 1 by causing a computer to execute: a step of extracting a three-dimensional learning shape from input three-dimensional shape learning data; a step of rotating the three-dimensional shape so that a processing surface in the three-dimensional shape is opposite a learning processing axis direction specified by an entered learning processing axis direction instruction; a step in generating a learning depth map, in which information about the depth in the learning processing axis direction extracted from the three-dimensional shape learning data is assigned to a two-dimensional image obtained by orthographically projecting the processing surface onto a plane orthogonal to the learning processing axis direction; and a step of updating an inference unit that determines the suitability of processing a planned processing shape, with a learning facility that performs machine learning using the learning processing axis direction instruction, the learning depth map, and the machinability information that describes the suitability of the actual processing already performed according to the learning processing axis direction instruction and the three-dimensional shape learning data.

Images