3D operation path generation device, 3D operation path generation method, and 3D operation path generation program

Abstract

A 3D operation path generation device according to the present invention comprises: a memory that stores at least diagram data for 2D shapes for the melting of a metal welding wire by a welding robot and 3D shape data for a metal layered object that is to be produced by the welding robot; and a processor that generates an operation route for the welding robot for production of the metal layered object by the layering of a plurality of single metal layers that are each a layer of the welding wire as melted into a 2D shape by the welding robot on the basis of the diagram data for the 2D shapes in a prescribed layering direction.

Description

3D motion path generating device, 3D motion path generating method, and 3D motion path generating program

[0001] The present disclosure relates to a 3D motion path generation device, a 3D motion path generation method, and a 3D motion path generation program.

[0002] Patent Literature 1 discloses a robot operation simulation method for verifying whether the teaching points of a created program are desirable without actually operating the robot. In this simulation method, after creating a program for a robot to be used in workpiece machining, a robot simulator is used to verify the appropriateness of the teaching points. A positional deviation tolerance area is set around each teaching point defined in the program, and a simulation is performed to verify that the positional deviation tolerance area is included in the robot's range of motion and that there is no interference with other components within that area. If any of the conditions is not met, the teaching points are corrected and verified again. If the simulation determines that the teaching points are appropriate, a program containing the teaching points is given to the robot, and the teaching points are corrected according to the position of the workpiece, after which the robot performs machining.

[0003] Also, Non-Patent Document 1 discloses a method for easily generating a solid model of a 3D figure by selecting and specifying 2D figures viewed from several viewpoints, such as a side view, a top view, a front view, etc.

[0004] Japanese Patent Application Publication No. 8-328632

[0005] Armonicos, “2D / 3D Conversion”, [online], [Retrieved July 16, 2024], Internet <URL: https: / / www.armonicos.co.jp / laboratory / 31 / >

[0006] The present disclosure has been devised in consideration of the current circumstances, and aims to easily and efficiently generate motion paths when processing metal laminates using a welding robot, without the need for expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0007] The present disclosure provides a 3D motion path generation device including: a memory that stores at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate manufactured by the welding robot; and a processor that generates a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one-layer metal layers by the welding robot in a predetermined stacking direction.

[0008] The present disclosure also provides a 3D motion path generation method executed by a 3D motion path generation device, the 3D motion path generation method including: storing in a memory at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one metal layers in a predetermined stacking direction by the welding robot.

[0009] The present disclosure also provides a program for causing a 3D motion path generation device, which is a computer, to realize the following processes: a process of storing in a memory at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and a process of generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one-layer metal layers in a predetermined stacking direction by the welding robot.

[0010] According to the present disclosure, motion paths for processing metal laminates using a welding robot can be easily and efficiently generated without the need for expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0011] FIG. 1 is a block diagram showing an example of the hardware configuration of a 3D motion path generation device according to this embodiment; FIG. 2 is a diagram showing an example of the concept of operation for generating a 3D motion path using 2D-shaped graphic data; FIG. 3 is a diagram showing an example of the concept of operation for generating a 3D motion path using 2D-shaped graphic data; FIG. 4 is a diagram showing an example of welding wire lamination when the change between layers is small; FIG. 5 is a diagram showing an example of welding wire lamination when the change between layers is large; FIG. 6 is a diagram showing an example of the outline of operation for generating a 3D motion path for processing a metal laminate according to this embodiment; FIG. 7 is a diagram showing an example of the concept of bead lamination width, weld height, and penetration length; A diagram showing an example of an unacceptable case. A flowchart showing an example of the overall operation procedure of a 3D operation path generation device according to this embodiment in chronological order. A flowchart showing an example of the operation procedure of the intermediate path generation process of FIG. 8 in chronological order. A flowchart showing an example of the operation procedure of the overlap determination process of FIG. 8 in chronological order. A flowchart showing an example of the operation procedure of the intermediate path generation process according to embodiment 2 in chronological order. A flowchart showing an example of the operation procedure of the overlap determination process according to embodiment 2 in chronological order. A flowchart showing an example of the operation procedure of the correction process by a trained AI model in chronological order. A flowchart showing an example of the operation procedure of the learning process of the ... after the visual inspection in chronological order.

[0012] (Background to the present disclosure) A technology is already known for manufacturing metal laminates having any shape desired by a user, using a 3D printer capable of handling metal powders such as steel, aluminum, titanium, or copper. The metal powder is irradiated with a laser beam to melt and solidify the metal layers, and these layers are gradually stacked by repeatedly forming layers. Conventionally, this type of metal laminate has often been manufactured using a dedicated system (e.g., the 3D printer described above). However, the high cost of such dedicated systems has hindered their widespread adoption. On the other hand, assuming that a user owns a welding robot and uses it to manufacture metal laminates, the introduction cost is often a fraction of that of a very expensive 3D printer, and initial costs can be reduced. However, in order to generate the motion path of the welding robot (hereinafter sometimes referred to as the "motion path") when processing and manufacturing metal laminates using the welding robot, it was necessary to purchase a dedicated application such as 3D-Computer Aided Design (CAD) and generate the motion path, which is the motion path of the welding robot, in accordance with the metal laminate to be manufactured.

[0013] In reality, the motion path for lamination must be generated using a dedicated application such as 3D-CAD, and the motion path of the welding robot must then be input into an offline teaching system. This not only increases costs by purchasing the dedicated application, but also inevitably increases the user's workload by having to learn how to operate the offline teaching system. Furthermore, depending on the user's environment, if 3D-CAD is not available and only 2D shapes are available (see, for example, Non-Patent Document 1), it is possible to generate a 3D model of the metal laminate itself, but it is difficult to use the 3D model itself to generate a motion path that will serve as the path for moving the welding robot when manufacturing the metal laminate, which is the target product.

[0014] Therefore, in the following embodiments, examples of a 3D motion path generation device, a 3D motion path generation method, and a 3D motion path generation program are described that can easily and efficiently generate motion paths when processing metal laminates using a welding robot, without the need for expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0015] Hereinafter, with reference to the accompanying drawings as appropriate, detailed descriptions of embodiments specifically disclosing a 3D motion path generation device, a 3D motion path generation method, and a 3D motion path generation program according to the present disclosure will be described in detail. However, unnecessary detailed descriptions may be omitted. For example, detailed descriptions of well-known matters or redundant descriptions of substantially identical configurations may be omitted. This is to avoid unnecessary redundancy in the following description and to facilitate understanding by those skilled in the art. Note that the accompanying drawings and the following description are provided to enable those skilled in the art to fully understand the present disclosure, and are not intended to limit the subject matter recited in the claims.

[0016] (Embodiment 1) 1. Configuration of 3D Motion Path Generator First, the configuration of a 3D motion path generation device 1 according to this embodiment will be described with reference to Fig. 1. Fig. 1 is a block diagram showing an example of the hardware configuration of the 3D motion path generation device 1 according to this embodiment. The 3D motion path generation device 1 includes at least a processor 2, a memory 3, an input device 4, and an output device 5. The processor 2, memory 3, input device 4, and output device 5 are connected via an internal bus 10 so that data signals can be input and output to and from each other. The 3D motion path generation device 1 is configured using, for example, a personal computer or a server computer, but may also be configured using a tablet terminal.

[0017] The 3D motion path generation device 1 generates a motion path, which is a three-dimensional motion path of a laser head (not shown) attached to the tip of a welding robot (not shown) when the welding robot manufactures a metal laminate (see FIGS. 2 and 3 ), in response to a user's operation using the input device 4. The motion path for manufacturing the metal laminate generated by the 3D motion path generation device 1 is input to the welding robot or a robot controller (not shown) for controlling the movement of the welding robot, and is used when manufacturing the metal laminate.

[0018] In this embodiment, a detailed description of the method for manufacturing a metal laminate using a welding robot will be omitted, but the outline is as follows. Specifically, the welding robot is equipped with a welding torch (not shown) at its tip that constantly holds a fed welding wire, and melts the metal welding wire held by the welding torch by an arc or the like, according to 2D shape reference graphic data held by the 3D motion path generation device 1. The welding robot treats the molten and solidified 2D-shaped welding wire as a single metal layer. After generating the single metal layer, the welding robot places (in other words, stacks) another metal layer generated in a similar manner on top of the previously generated metal layer. The welding robot repeatedly generates and stacks one metal layer to manufacture the metal laminate, which is the manufacturing target.

[0019] Here, a metal laminate manufactured by a welding robot will be described with reference to FIGS. 2 and 3. FIGS. 2 and 3 are diagrams showing an example of the operation concept of generating a 3D motion path using 2D shape graphic data. In FIG. 2, direction DR1 is the lamination direction of the welding wire (one metal layer) generated by the welding robot melting and solidifying a metal welding wire along the reference 2D shape graphic data Pth1 in order for the welding robot to manufacture the metal laminates WP1 and WP2. The welding robot moves a welding torch (not shown) holding the metal welding wire along the reference 2D shape graphic data Pth1, and during this movement, the welding wire is melted by an arc or the like, and the welding wire formed by subsequent solidification is treated as one metal layer.

[0020] 2, when a single metal layer corresponding to each of the plurality of 2D shape graphic data Pth1 is stacked along the direction DR1 without applying any parameters (e.g., magnification, reduction, rotation angle) to any of the plurality of 2D shape graphic data Pth1, a cylindrical metal laminate WP1 is formed. In this case, the welding robot moves the welding torch (i.e., welding wire) along the 2D shape graphic data Pth1 to generate one metal layer, then moves the welding torch to the next upper layer position and similarly generates one metal layer corresponding to that upper layer position, and so on.

[0021] On the other hand, if a specific parameter (e.g., an enlargement or reduction rate) is applied to some of the multiple pieces of 2D-shaped graphic data Pth1 and multiple metal layers corresponding to the 2D-shaped graphic data Pth1 are stacked along the direction DR1, a substantially cylindrical metal laminate WP2 is formed with a slightly recessed center compared to the upper and lower ends along the direction DR1. In this case, the welding robot moves the welding torch along the 2D-shaped graphic data Pth1 to generate one metal layer, then moves the welding torch to the next upper layer position and moves the welding torch along 2D-shaped graphic data Pth2 having parameters different from the parameters of the 2D-shaped graphic data Pth1 (e.g., reduction rate) to generate one metal layer, and repeats the same process, moving the welding torch along the 2D-shaped graphic data PthZ corresponding to the top layer position to generate one metal layer.

[0022] 2 , 2D-shaped graphic data Pth3, in which multiple protrusions Pj1, Pj2, Pj3, and Pj4 are further arranged on 2D-shaped graphic data Pth1, is provided with specific parameters (e.g., a rotation angle) for the graphic data, and multiple metal layers corresponding to the 2D-shaped graphic data Pth3 are stacked along the direction DR1 to form a cylindrical screw-shaped metal laminate WP3. In this case, the welding robot moves a welding torch along the 2D-shaped graphic data Pth3 to generate one metal layer, then moves the welding torch to the next upper layer position and moves the welding torch along 2D-shaped graphic data Pth3 having parameters different from the parameters (e.g., rotation angle) of the 2D-shaped graphic data Pth3 to generate one metal layer, and repeats the same process to generate one metal layer by moving the welding torch along the 2D-shaped graphic data Pth3 corresponding to the top layer position.

[0023] The processor 2 is configured by at least one of, for example, a Central Processing Unit (CPU), a Digital Signal Processor (DSP), a Graphical Processing Unit (GPU), or a Field Programmable Gate Array (FPGA). The processor 2 functions as a controller that manages the overall operation of the 3D motion path generation device 1. The processor 2 performs control processing for overseeing the operation of each part of the 3D motion path generation device 1, data input / output processing between each part of the 3D motion path generation device 1, data calculation processing, and data storage processing. The processor 2 operates according to a program stored in the memory 3. The processor 2 uses the memory 3 during operation, and temporarily stores data generated or acquired by the processor 2 in the memory 3. In the 3D motion path generation device 1, the processor 2 executes various processes (for example, the processes of the flowcharts shown in Figs. 8 to 10) in cooperation with the memory 3. By cooperating with the memory 3, the processor 2 realizes the functions of the overall control unit 11, the intermediate path generation unit 12, and the overlap determination unit 13.

[0024] The overall control unit 11 controls the generation of a motion path (motion path) for a welding robot for manufacturing a metal laminate based on 2D shape graphic data (see FIG. 2) stored in the memory 3 and 3D shape data for the metal laminate, which is the object to be manufactured. For example, the overall control unit 11 generates a motion path for the welding robot for manufacturing a metal laminate by forming a 2D shape welding wire melted and solidified by the welding robot based on the 2D shape graphic data into one metal layer and stacking multiple single metal layers in a predetermined stacking direction (e.g., see direction DR1 in FIG. 2 or 3). Details of the operation procedure of the overall control unit 11 will be described later with reference to FIG. 8.

[0025] The intermediate pass generator 12 is called as needed during the processing of the operation procedure of the overall controller 11, and performs processing to generate one or more intermediate passes based on the allowable range value of the penetration length (see FIG. 6) of one metal layer corresponding to the welding conditions stored in the memory 3. An intermediate pass is an intermediate operation path (operation path) that is placed between adjacent metal layers when the adjacent metal layers do not overlap in the stacking direction (see, for example, direction DR1 in FIG. 2 or FIG. 3). Details of the operation procedure of the intermediate pass generator 12 will be described later with reference to FIG. 9.

[0026] Here, the penetration length of one metal layer will be described with reference to FIG. 6 . FIG. 6 is a diagram showing a conceptual example of the bead stack width, weld height, and penetration length. In the description of FIG. 6 , the mutually orthogonal X, Y, and Z directions are as shown. Specifically, the X direction indicates the length direction (longitudinal direction) of the bead BD1 formed on the lower plate BPL1 by welding with a welding wire arc or the like by a welding robot; in other words, the direction of the weld line of the welding performed by the welding robot. The Y direction is parallel to the lower plate BPL1 and orthogonal to the X direction, and further indicates the width direction of the bead BD1. The Z direction is orthogonal to the X and Y directions, and further indicates the height direction of the bead BD1.

[0027] In this embodiment, for ease of explanation, it is assumed that a substantially linear bead BD1, shown as a welding result WD1 by a welding robot in FIG. 6 , is stacked in the Z direction. It is assumed that this bead BD1 has an elliptical shape when cut in the YZ plane. The stacking width of the bead BD1 refers to the length equivalent to the major axis of the elliptical bead BD1 when the bead BD1 of the welding result WD1 is cut along a cut surface SEC1 parallel to the YZ plane. The welding height of the bead BD1 refers to the length (height) in the Z direction of the minor axis of the elliptical bead BD1 when the bead BD1 of the welding result WD1 is cut along a cut surface SEC1 parallel to the YZ plane, i.e., the height from the lower plate BPL1. The penetration length of the bead BD1 indicates the length (depth) of the short diameter of the elliptical bead BD1 in the -Z direction when the bead BD1 of the welding result WD1 is cut along a cutting plane SEC1 parallel to the YZ plane, i.e., the depth from the lower plate BPL1.

[0028] The overlap determination unit 13 is called as needed during the processing of the operating procedure of the overall control unit 11, and performs processing according to the overlap distance (see below) between adjacent metal layers in the stacking direction (e.g., see direction DR1 in FIG. 2 or FIG. 3 ) based on the allowable range value of the penetration length of one metal layer corresponding to the welding conditions stored in the memory 3. Details of the operating procedure of the overlap determination unit 13 will be described later with reference to FIG. 9 .

[0029] The memory 3 is configured using, for example, Random Access Memory (RAM) and Read Only Memory (ROM), and temporarily stores programs necessary for the operation of the 3D motion path generation device 1 and data acquired or generated during operation. The RAM is, for example, a work memory used during the operation of the 3D motion path generation device 1. The ROM, for example, stores and holds programs for controlling the 3D motion path generation device 1 in advance.

[0030] The memory 3 stores, for example, at least 2D shape graphic data (see Figure 2) that serves as a base showing the shape in which the metal powder will be placed by the welding robot, and 3D shape data of the metal laminate (see Figure 2 or Figure 3) manufactured by the welding robot.

[0031] The memory 3 also stores, for example, a combination of data on multiple welding conditions used by a welding robot to manufacture a metal laminate, as well as data on the allowable range of values ​​for the penetration length of one metal layer corresponding to the combination of welding condition data. Examples of combinations of welding condition data include welding current, welding voltage, welding wire feed rate, welding wire diameter, and welding wire extension length. The shape of the laminate (specifically, the lamination width, weld height, and penetration length of the bead BD1 (see FIG. 6)) changes depending on the combination of these welding condition data and the welding robot's operating speed. For example, the memory 3 stores multiple combinations of welding condition data categorized by lamination width (bead width) and weld height, and the welding robot's operating speed. When the welding robot performs lamination, the combination of welding condition data and the welding robot's operating speed corresponding to the lamination width (bead width) and weld height data expected for the metal laminate to be manufactured are selected and used. The welding robot's operating speed may always be constant.

[0032] The input device 4 is connected to the 3D motion path generation apparatus 1 so as to be able to input and output data therebetween, and is a device that accepts input operations from a user of the 3D motion path generation apparatus 1. A signal based on the input operation from the user is input from the input device 4 to the 3D motion path generation apparatus 1. The input device 4 may be, for example, a mouse, a keyboard, a touch panel, or the like, or a combination of these.

[0033] The output device 5 is connected to the 3D motion path generation device 1 so as to be able to input and output data therefrom, and displays the motion path and metal laminate (see FIGS. 2 and 3) executed by the 3D motion path generation device 1. The output device 5 is, for example, a display such as a Liquid Crystal Display (LCD) or an organic EL display. The input device 4 and the output device 5 may be configured as an integrated device, in which case the input device 4 and the output device 5 are configured as a touch panel display.

[0034] 2. Generation of intermediate paths Next, the generation of intermediate paths will be described with reference to Figures 4A, 4B, and 5. Figure 4A is a diagram schematically illustrating an example of metal lamination when the change between laminations is small. Figure 4B is a diagram schematically illustrating an example of metal lamination when the change between laminations is large. Figure 5 is a diagram schematically illustrating an example of an outline of a 3D motion path generation operation for processing a metal laminate according to this embodiment.

[0035] 4A shows how adjacent metal layers are all overlapped along the stacking direction of the metal layers (e.g., direction DR1). That is, a pair of 2D shape graphic data Pth1 and Pth2, a pair of 2D shape graphic data Pth2 and Pth3, and a pair of 2D shape graphic data Pth3 and Pth4, which correspond to each metal layer, are all overlapped along direction DR1. In this case, each metal layer can be stacked.

[0036] 4B shows that adjacent metal layers do not overlap along the stacking direction (e.g., direction DR1) of the metal layers. That is, the pair of 2D shape graphic data Pth1 and Pth2, the pair of 2D shape graphic data Pth2 and Pth3, and the pair of 2D shape graphic data Pth3 and Pth4, which correspond to each metal layer, do not overlap along direction DR1. As described above, the stacking width and the welding height are determined by the welding conditions (e.g., welding current, welding voltage, welding wire feed speed) and the operating speed of the welding robot.

[0037] Therefore, in the present embodiment, when adjacent metal layers do not overlap in the stacking direction (see direction DR1), as in the case of 2D-shaped graphic data Pth1, Pth2, Pth3, and Pth4 shown in FIG. 4B , processor 2 of 3D motion path generation device 1 generates intermediate paths between the adjacent metal layers. Specifically, processor 2 changes at least the welding conditions (e.g., welding current, welding voltage, welding wire feed speed) to generate intermediate paths by modifying the 2D-shaped graphic data corresponding to the metal layers, and generates motion paths for the welding robot so that all adjacent metal layers overlap in the stacking direction (see direction DR1). That is, intermediate path Pth5 is generated between the pair of 2D-shaped graphic data Pth1 and Pth2, intermediate path Pth6 is generated between the pair of 2D-shaped graphic data Pth2 and Pth3, and intermediate path Pth7 is generated between the pair of 2D-shaped graphic data Pth3 and Pth4. As a result, all adjacent metal layers overlap each other along the stacking direction of one metal layer (for example, direction DR1).

[0038] For example, as shown in FIG. 5 , an initial motion path Ps corresponding to one metal layer is placed at the start of lamination. The processor 2 of the 3D motion path generation device 1 then places the final motion path Pe in response to a user's input device 4. After placing the final motion path Pe, the processor 2 receives a specification of a change point (e.g., enlargement rate, reduction rate) in the lamination direction (see direction DR1) in response to the user's input device 4. Based on this specification, the processor 2 calculates the shape and number of intermediate paths Pms to Pme required to be placed between the initial motion path Ps and the final motion path Pe (see FIG. 5 ). The lamination height per layer (= welding height + penetration length) is determined by the welding conditions (e.g., welding current, welding voltage, welding wire feed speed). Therefore, the number of layers can be calculated by (height / lamination height per layer). When generating multiple intermediate paths, the processor 2 generates multiple intermediate paths Pms to Pme1 of the same shape from the lamination start position to the lamination end position using the same welding conditions, for example. Furthermore, processor 2 determines whether an intermediate pass based on the same welding conditions overlaps with final operation pass Pe in the stacking direction (see direction DR1) when the intermediate pass is placed, and if it determines that the intermediate pass does not overlap with final operation pass Pe, it changes the welding conditions for the intermediate pass immediately before the stacking end position and generates and places intermediate pass Pme corresponding to the changed welding conditions. Since intermediate pass Pme is generated according to different welding conditions from the other intermediate passes Pms to Pme1, its shape (e.g., size, diameter) is different; in the example of Figure 5, intermediate pass Pme has a smaller diameter than the other intermediate passes Pms to Pme1.

[0039] 3. Operational Procedure of the 3D Motion Path Generator Next, an example of the operational procedure of the 3D motion path generator 1 according to this embodiment will be described with reference to FIGS. 7A , 7B , 8 , 9 , and 10 . FIG. 7A is a diagram showing an example in which the penetration length between motion paths can be calculated when stacking in the stacking direction. FIG. 7B is a diagram showing an example in which the penetration length between motion paths cannot be calculated when stacking in the stacking direction. FIG. 8 is a flowchart showing an example of the overall operational procedure of the 3D motion path generator 1 according to this embodiment in chronological order. FIG. 9 is a flowchart showing an example of the operational procedure of the intermediate path generation process of FIG. 8 in chronological order. FIG. 10 is a flowchart showing an example of the operational procedure of the overlap determination process of FIG. 8 in chronological order. The series of processes shown in FIG. 8 is mainly executed by the overall control unit 11 of the processor 2, the series of processes shown in FIG. 9 is mainly executed by the intermediate path generation unit 12 of the processor 2, and the series of processes shown in FIG. 10 is mainly executed by the overlap determination unit 13 of the processor 2.

[0040] 8 , processor 2 reads from memory 3 2D shape graphic data characterizing the shape of a welding wire, which is a metal layer to be laminated when a welding robot manufactures a metal laminate, and 3D shape data of the metal laminate, which is the object to be manufactured (Step St1). Based on the 3D shape data of the metal laminate read in Step St1, processor 2 reads from memory 3 parameters of the 2D shape graphic data (e.g., the height, magnification, and reduction rate of each metal layer) when the 2D shape graphic data is laminated along a lamination direction (e.g., see direction DR1) (Step St2). In response to a user's operation using input device 4 (e.g., designation of final movement path Pe), processor 2 calculates the number of laminations of the movement path from initial movement path Ps (see FIG. 5 ) to final movement path Pe (see FIG. 5 ) based on the 3D shape data of the metal laminate, and repeatedly executes a series of processes from Step St3 to Step St10 the number of laminations indicated by the calculation result. The series of processes from step St3 to step St10 is a so-called loop process, which is executed for each operation pass of one layer (i.e., one metal layer).

[0041] Processor 2 starts the lamination process for each operation path (Step 3) and first generates an operation path corresponding to one metal layer (e.g., the initial operation path Ps in FIG. 5) (Step 4). After step St4, processor 2 determines whether an operation path for the previous layer has already been generated (Step 5). If processor 2 determines that an operation path for the previous layer has not already been generated (Step 5, NO), it ends the lamination process for each path (Step 10).

[0042] On the other hand, if processor 2 determines that the action path of the previous layer has already been generated (Step 5, YES), it determines whether the most recently generated action path and the action path of the previous layer overlap in the stacking direction (e.g., see direction DR1) (Step 6). Examples of overlapping and non-overlapping in the stacking direction (e.g., see direction DR1) will be described with reference to Figures 7A and 7B.

[0043] FIG. 7A shows an example in which a movement path Pth11 generated and arranged in the nth position (n: an integer equal to or greater than 1) and a movement path Pth12 generated and arranged in the (n+1)th position overlap along the stacking direction (see, for example, direction DR1). The stacking direction (see, for example, direction DR1) is parallel to the Z direction (see FIG. 6). In this embodiment, the movement paths Pth11 and Pth12 are assumed to be elliptical. When the movement paths Pth11 and Pth12 overlap along the stacking direction (see, for example, direction DR1), there is an overlapping portion between the movement paths Pth11 and Pth12. In this case, the length d1 of the straight line Ln1 connecting the elliptical center C11 of the operating path Pth11 and the elliptical center C12 of the operating path Pth12 within the overlapping area is the penetration length (see Figure 6) when one metal layer corresponding to the operating path Pth11 and one metal layer corresponding to the operating path Pth12 are stacked.

[0044] 7B shows an example in which the nth generated and arranged motion path Pth11 and the (n+1)th generated and arranged motion path Pth12 do not overlap in the stacking direction (see, for example, direction DR1). In this embodiment, the motion paths Pth11 and Pth12 are assumed to be elliptical. When the motion paths Pth11 and Pth12 do not overlap in the stacking direction (see, for example, direction DR1), the processor 2 generates and arranges one or more intermediate paths Pth13 and Pth14 between the motion paths Pth11 and Pth12. For ease of explanation, the shapes of the intermediate paths Pth13 and Pth14 are assumed to be the same as those of the motion paths Pth11 and Pth12.

[0045] As a result, the length d1 of a straight line connecting the center C11 of the ellipse of the operating path Pth11 and the center C13 of the ellipse of the intermediate path Pth13 within the range of the overlapping portion of the operating path Pth11 and the intermediate path Pth13 becomes the penetration length (see FIG. 6 ) when one metal layer corresponding to the operating path Pth11 and one metal layer corresponding to the intermediate path Pth13 are laminated together. Similarly, the length d3 of a straight line connecting the center C13 of the ellipse of the intermediate path Pth13 and the center C14 of the ellipse of the intermediate path Pth14 within the range of the overlapping portion of the intermediate path Pth13 and the intermediate path Pth14 becomes the penetration length (see FIG. 6 ) when one metal layer corresponding to the intermediate path Pth13 and one metal layer corresponding to the intermediate path Pth14 are laminated together. Furthermore, within the range where the intermediate path Pth14 and the operating path Pth12 overlap, the length d2 of the straight line connecting the elliptical center C14 of the intermediate path Pth14 and the elliptical center C12 of the operating path Pth12 becomes the penetration length (see Figure 6) when one metal layer corresponding to the intermediate path Pth14 and one metal layer corresponding to the operating path Pth12 are stacked.

[0046] 8, if processor 2 determines that the most recently generated action path and the action path of the previous layer do not overlap in the stacking direction (e.g., see direction DR1) (St6, NO), it performs intermediate path generation processing (see FIG. 9) to generate one or more intermediate paths between the most recently generated action path and the action path of the previous layer (St7). After processing step St7, processor 2 proceeds to step St9.

[0047] On the other hand, if processor 2 determines that the most recently generated motion path and the motion path of the previous layer overlap in the stacking direction (e.g., see direction DR1) (St6, YES), it performs overlap determination processing (see FIG. 10) according to the allowable range value of the penetration length of one metal layer corresponding to the currently set welding conditions (St8). After processing step St8, the processing of processor 2 proceeds to step St9.

[0048] The processor 2 determines whether or not the intermediate path generation process in step St7 or the overlap determination process in step St8 resulted in an error (i.e., it is difficult to generate a 3D motion path for the metal laminate) (St9). If the processor 2 determines that there is an error (i.e., it is difficult to generate a 3D motion path for the metal laminate) (St9, YES), the processor 2 ends the operation procedure shown in Fig. 8. On the other hand, if the processor 2 determines that there is no error (i.e., it is difficult to generate a 3D motion path for the metal laminate) (St9, NO), the processor 2 ends the stacking process for the current motion path (St10). After the stacking process for the motion paths for the number of laminations (a series of processes from step St3 to step St10) is completed, the processor 2 ends the operation procedure shown in Fig. 8.

[0049] In FIG. 9 , processor 2 reads from memory 3 and acquires the allowable range value for the penetration length of one metal layer corresponding to the currently set welding conditions (Step 11). Processor 2 generates intermediate paths that can be stacked within the allowable range value for the penetration length acquired in Step St11 based on the distance between the most recently generated motion path and the motion path of the previous layer (e.g., the linear distance from the center C11 of the ellipse to the center C12 of the ellipse in FIG. 7B ) (Step 12). Here, the allowable range value for the penetration length is a range specified by a coefficient or numerical value for the penetration length of one metal layer corresponding to the welding conditions. For example, if the penetration length is 5 mm and the allowable range value is ±20%, the allowable penetration length range is 4 to 6 mm. Note that the numerical value is ±1 mm.

[0050] Processor 2 determines whether the overlap between the last added intermediate path (i.e., the intermediate path generated in the most recent step St12) and the latest operation path generated immediately before the intermediate path generation process is within the allowable range (St13). If it is determined that the overlap between the last added intermediate path (i.e., the intermediate path generated in the most recent step St12) and the latest operation path generated immediately before the intermediate path generation process is within the allowable range (St13, YES), the intermediate path generation process by processor 2 ends.

[0051] On the other hand, if the processor 2 determines that the overlap between the last added intermediate path (i.e., the intermediate path generated in step St12) and the most recent operation path generated immediately before the intermediate path generation process is not within the allowable range (NO in step St13), the processor 2 determines whether there are other usable welding conditions different from the currently set welding conditions (St14). That is, in step St14, the processor 2 determines whether there are welding conditions different from the currently set welding conditions, and the overlap between the last added intermediate path and the most recent operation path generated immediately before the intermediate path generation process is within the allowable range. If the processor 2 determines that there are no other usable welding conditions different from the currently set welding conditions (NO in St14), the processor 2 performs error processing (i.e., there are no appropriate welding conditions for generating intermediate paths to fill each operation path) and outputs a notification to that effect to the output device 5 (St15). This allows the user to quickly recognize the need to add appropriate welding conditions for generating intermediate paths to fill each operation path.

[0052] On the other hand, if processor 2 determines that there are other usable welding conditions different from the currently set welding conditions (St14, YES), it changes the welding conditions for the intermediate pass to the other welding conditions (St16). For example, in step St16, if processor 2 determines that there are multiple applicable welding conditions, it may select the welding conditions that are registered first in memory 3, giving priority to the order in which the other welding conditions are registered. Alternatively, if processor 2 determines that there are multiple applicable welding conditions, it may select a welding condition that has the same or similar welding width (layer width) as the currently set welding conditions. Furthermore, processor 2 reads and obtains from memory 3 the allowable range value for the penetration length of one metal layer that corresponds to the changed welding conditions (St16). Processor 2 updates (changes) the shape of the last added intermediate pass (i.e., the intermediate pass generated in step St12) according to the welding conditions changed in step St16 (St17). Processor 2 determines whether the overlap between the intermediate path updated in step St17 and the latest operation path generated immediately before the intermediate path generation process is within the allowable range value of the penetration length of one metal layer corresponding to the changed welding conditions obtained in step St16 (St18).

[0053] If it is determined that the overlap between the intermediate path updated in step St17 and the latest operation path generated immediately before the intermediate path generation process is within the allowable range value of the penetration length of one metal layer corresponding to the changed welding conditions obtained in step St16 (St18, YES), the intermediate path generation process by processor 2 is terminated.

[0054] On the other hand, if it is determined that the overlap between the intermediate path updated in step St17 and the latest operation path generated immediately before the intermediate path generation process is not within the allowable range value of the penetration length of one metal layer corresponding to the changed welding conditions obtained in step St16 (St18, NO), the processing of processor 2 returns to step St12.

[0055] 10 , processor 2 reads and acquires from memory 3 the allowable range value of the penetration length of one metal layer corresponding to the currently set welding conditions (St21). Processor 2 determines whether the overlap between the most recently generated motion path and the motion path of the previous layer is within the allowable range value of the penetration length of one metal layer acquired in step St21 (St22). If it is determined that the overlap between the most recently generated motion path and the motion path of the previous layer is within the allowable range value of the penetration length of one metal layer acquired in step St21 (YES in St22), the overlap determination process by processor 2 ends.

[0056] On the other hand, if the processor 2 determines that the overlap between the most recently generated motion path and the motion path of the previous layer is not within the allowable range of the penetration length of the first metal layer acquired in step St21 (NO in St22), it determines whether the distance corresponding to the overlap between the most recently generated motion path and the motion path of the previous layer is less than the lower limit of the allowable range of the penetration length of the first metal layer acquired in step St21 (St23). If the processor 2 determines that the distance corresponding to the overlap between the most recently generated motion path and the motion path of the previous layer is less than the lower limit of the allowable range of the penetration length of the first metal layer acquired in step St21 (YES in St23), it generates an insufficient contact notification and outputs it to the output device 5 (St24). The insufficient contact notification is an example of a first notification, and is a notification that notifies the user that the overlap between the most recently generated motion path and the motion path of the previous layer is too short, which may result in peeling of the laminated portion.

[0057] On the other hand, if the processor 2 determines that the distance corresponding to the overlap between the most recently generated motion path and the motion path of the previous layer is not less than the lower limit of the allowable range value of the penetration length of one metal layer acquired in step St21 (St23, NO), it generates a notification of insufficient stacking spacing and outputs it to the output device 5 (St25). The notification of insufficient stacking spacing is an example of a second notification, and is a notification that notifies the user that, for example, the distance corresponding to the overlap between the most recently generated motion path and the motion path of the previous layer is too long, and therefore the penetration length is not sufficiently secured, which may cause the stack shape to collapse.

[0058] After step St24 or step St25, processor 2 determines that there is an error (i.e., it is difficult to generate the 3D motion path of the metal laminate) (St26), and ends the overlap determination process.

[0059] (Embodiment 2) In embodiment 2, a process for correcting the allowable range value of the penetration length of one metal layer corresponding to the welding conditions using a trained AI model, and a process for training the AI ​​model will be described. In the 3D motion path generation device according to embodiment 2, a configuration for performing correction processing using a trained AI model is added to the configuration of the 3D motion path generation device according to embodiment 1.

[0060] 4. Configuration of 3D motion path generation device and external devices First, the configuration of a 3D motion path generation device 1 according to embodiment 2 will be described with reference to Fig. 1. Note that the same components as those of the 3D motion path generation device 1 according to embodiment 1 will be assigned the same reference numerals, and explanations will be simplified or omitted, and only differences will be described. The 3D motion path generation device 1 according to embodiment 2 includes a processor 2, a memory 3, an input device 4, an output device 5, and a communication I / F 6.

[0061] The processor 2, in cooperation with the memory 3, further realizes the functions of an AI model correction unit 14 and an AI model learning unit 15 in addition to the overall control unit 11, intermediate path generation unit 12, and overlap determination unit 13. Note that the AI ​​model learning unit 15 may be omitted.

[0062] The AI ​​model correction unit 14 performs a process of correcting the allowable range value of the penetration length of one metal layer that corresponds to the current welding conditions, using the trained AI model 3a stored in the memory 3. Details of the correction by the AI ​​model correction unit 14 will be described later with reference to FIG.

[0063] The AI ​​model learning unit 15 receives as input parameters of 2D shape graphic data (e.g., height, magnification, and reduction rate of each metal layer) and welding conditions (e.g., welding current, welding voltage, welding wire feed rate, welding wire diameter, and welding wire extension length) when stacking metal layers along the stacking direction to manufacture a metal laminate, and performs machine learning to generate or update a trained AI model by outputting an allowable range value for the penetration length of one metal layer. Details of learning by the AI ​​model learning unit 15 will be described later with reference to FIGS. 14 and 15.

[0064] In addition to the various data or information stored with reference to embodiment 1, the memory 3 stores a trained AI model 3a that has been generated or updated through machine learning by the AI ​​model learning device 20 or the AI ​​model learning unit 15.

[0065] The communication I / F 6 is configured using a circuit capable of communicating between the 3D motion path generation device 1 and an external device connected so as to be able to communicate data with the 3D motion path generation device 1. The communication I / F 6 performs data communication with at least one of the AI ​​model learning device 20 and the appearance inspection device 30, for example.

[0066] The AI ​​model learning device 20 is an example of an external device connected to the 3D motion path generation device 1 so as to be able to communicate data, and is connected, for example, via a wired or wireless network (not shown). Particularly when connected via a wireless network, the AI ​​model learning device 20 may be configured as a cloud computer. The AI ​​model learning device 20 performs machine learning using various data or information acquired from the 3D motion path generation device 1 to generate or update a trained AI model 22. Specifically, the AI ​​model learning device 20 receives inputs of parameters (e.g., height, magnification, and reduction rate of each metal layer) of 2D shape graphic data when layers are stacked along the stacking direction to manufacture a metal laminate, and welding conditions (e.g., welding current, welding voltage, welding wire feed rate, welding wire diameter, and welding wire protrusion length), and generates or updates the trained AI model by performing machine learning to output an allowable range value for the penetration length of each metal layer.

[0067] The virtual storage device 21 is configured by a storage such as a hard disk drive (HDD) or a solid state drive (SSD), etc. The virtual storage device 21 stores various learning data acquired from the 3D motion path generation device 1 or the appearance inspection device 30 via the 3D motion path generation device 1 for use in machine learning executed by the AI ​​model learning device 20.

[0068] The trained AI model 22 is an AI model that is generated or updated through machine learning (see above) by the AI ​​model learning device 20, similar to the trained AI model 3a stored in the memory 3 of the 3D motion path generation device 1. The trained AI model 22 executes a process of correcting the allowable range value of the penetration length of one metal layer that corresponds to the current welding conditions, for example.

[0069] The appearance inspection device 30 performs an appearance inspection of a metal laminate (see, for example, FIG. 2 or 3 ) manufactured by a welding robot (not shown) based on the motion path generated by the 3D motion path generation device 1. The appearance inspection device 30 sends appearance inspection information (for example, success or failure in manufacturing the metal laminate) that is information indicating the result of the appearance inspection to the 3D motion path generation device 1. Note that this appearance inspection information is used as learning data required for machine learning by the 3D motion path generation device 1 or the AI ​​model learning device 20.

[0070] 5. Operational Procedure of 3D Motion Path Generation Device Next, an example of the operational procedure of the 3D motion path generation device 1 according to this embodiment will be described with reference to Figs. 11, 12, 13, 14, and 15. Fig. 11 is a flowchart showing an example of the operational procedure of intermediate path generation processing according to embodiment 2 in chronological order. Fig. 12 is a flowchart showing an example of the operational procedure of overlap determination processing according to embodiment 2 in chronological order. Fig. 13 is a flowchart showing an example of the operational procedure of correction processing using a trained AI model in chronological order. Fig. 14 is a flowchart showing an example of the operational procedure of learning processing of an AI model in chronological order. Fig. 15 is a flowchart showing an example of the operational procedure of learning processing of an AI model after visual inspection in chronological order.

[0071] 11, the same steps as those in the corresponding intermediate path generation process of FIG. 9 according to embodiment 1 are assigned the same step numbers, and explanations thereof will be simplified or omitted, and only differences will be described. The series of processes shown in FIG. 11 are mainly executed by either the intermediate path generation unit 12 or the AI ​​model correction unit 14 of the processor 2.

[0072] In FIG. 11 , processor 2 reads from memory 3 and acquires the allowable range value of the penetration length of one metal layer corresponding to the currently set welding conditions (St11). After step St11, processor 2 executablely references trained AI model 3a stored in memory 3 and, using trained AI model 3a, corrects the allowable range value of the penetration length of one metal layer acquired in step St11 according to the current welding conditions (St30). Details of the processing of step St30 will be described later with reference to FIG. 13. Processor 2 generates intermediate paths that can be layered within the allowable range value of the penetration length corrected in step St30 based on the distance between the most recently generated motion path and the motion path of the previous layer (e.g., the linear distance from the ellipse center C11 to the ellipse center C12 in FIG. 7B ) (St12).

[0073] After step St16, the processor 2 executablely references the trained AI model 3a stored in the memory 3 and uses the trained AI model 3a to correct the allowable range value of the penetration length of one metal layer acquired in step St16 in accordance with the welding conditions changed in step St16 (St30). Details of the processing of step St30 will also be described later with reference to FIG. 13. The processor 2 updates (changes) the shape of the last-added intermediate path (i.e., the intermediate path generated in the most recent step St12) in accordance with the welding conditions changed in step St16 and the allowable range value of the penetration length corrected in step St30 (St17).

[0074] After step St15, or if it is determined that the overlap between the final additional intermediate path updated in step St17 and the motion path generated immediately before the final additional intermediate path is within the permissible range of the penetration length corrected in step St30 (YES in St18), the processor 2 executes a learning process for the AI ​​model (St40). Details of the process of step St30 will be described later with reference to FIGS. 14 and 15.

[0075] The series of processes shown in Fig. 12 are mainly executed by the overlap determination unit 13 of the processor 2. In the description of Fig. 12, the same processes as those in the corresponding overlap determination process of Fig. 10 according to the first embodiment are given the same step numbers, and the description will be simplified or omitted, and only the different contents will be described.

[0076] 12, the processor 2 determines whether or not the allowable range of the penetration length corrected by the trained AI model 3a in the intermediate path generation process described with reference to FIG. 11 exists (Step 20). If the processor 2 determines that the allowable range of the corrected penetration length exists (Step 20, YES), the processor 2 determines whether the overlap between the most recently generated motion path and the motion path of the previous layer is within the allowable range of the penetration length of the first metal layer corrected by the trained AI model 3a (Step 22). On the other hand, if the processor 2 determines that the allowable range of the corrected penetration length does not exist (Step 20, NO), the processor 2 reads and acquires from the memory 3 the allowable range of the penetration length of the first metal layer corresponding to the currently set welding conditions (Step 21).

[0077] The series of processes shown in Fig. 13 is mainly executed by the AI ​​model correction unit 14 of the processor 2. In Fig. 13, the processor 2 determines whether the trained AI model 3a is usable (St31). The processor 2 can determine whether the trained AI model 3a is usable, for example, depending on whether setting information indicating that the trained AI model 3a is saved in the memory 3 and is usable is stored in the memory 3. If it is determined that the trained AI model 3a is not usable (St31, NO), the correction process of the processor 2 shown in Fig. 13 ends.

[0078] On the other hand, if the processor 2 determines that the trained AI model 3a can be used (Step 31, YES), it refers to the trained AI model 3a and performs a process of correcting (in other words, updating) the allowable range value of the penetration length of the first metal layer corresponding to the current welding conditions (Step 32). Furthermore, the processor 2 may not only obtain the corrected allowable range value of the penetration length of the first metal layer using the trained AI model 3a, but also obtain the welding conditions corresponding to the corrected allowable range value. This allows the 3D motion path generation device 1 to update the allowable range value of the penetration length of the first metal layer and the welding conditions corresponding to the current welding conditions to more appropriate allowable range values ​​using the trained AI model 3a, thereby contributing to the generation of more appropriate intermediate paths.

[0079] The series of processes shown in Fig. 14 is mainly executed by the AI ​​model learning unit 15 of the processor 2 and the AI ​​model learning device 20. In Fig. 14, the processor 2 determines whether the trained AI model 3a is usable (St41). The processor 2 can determine whether the trained AI model 3a is usable, for example, depending on whether setting information indicating that the trained AI model 3a is saved in the memory 3 and is usable is stored in the memory 3. If it is determined that the trained AI model 3a is not usable (St41, NO), the learning process of the processor 2 or the AI ​​model learning device 20 shown in Fig. 14 ends.

[0080] On the other hand, if the processor 2 determines that the trained AI model 3a can be used (St41, YES), it determines whether an error was processed in the intermediate path generation process, for example, by referring to the work memory in the memory 3 at the time of execution of the intermediate path generation process (see Figure 11) (St42).

[0081] If the processor 2 determines that no error processing was performed in the intermediate path generation process (Step 42, NO), it retrieves learning data from the memory 3 when no error occurred and stores it in the virtual storage device 21 (Step 43). The learning data when no error occurred includes, but is not limited to, (1) information indicating that no error occurred, (2) parameters of 2D shape graphic data when stacking along the stacking direction to manufacture a metal laminate (e.g., height, magnification, and reduction ratio of each metal layer), (3) welding conditions (e.g., welding current, welding voltage, welding wire feed rate, welding wire diameter, and welding wire protrusion length), and (4) the allowable range of the penetration length of each metal layer. This learning data is stored in the virtual storage device 21 of the AI ​​model learning device 20, and machine learning is performed by the AI ​​model learning device 20 using this learning data. As a result, the AI ​​model learning device 20 can generate or update a trained AI model 3a that obtains more appropriate output based on the results (performance) of the intermediate path generation process, based on machine learning using the learning data.

[0082] When the processor 2 determines that an error has been processed in the intermediate path generation process (Step 42, YES), it retrieves learning data from the memory 3 at the time of the error occurrence and stores it in the virtual storage device 21 (Step 44). The learning data at the time of the error occurrence includes, but is not limited to, (1) information indicating that an error has occurred, (2) parameters of 2D shape graphic data (e.g., the height, magnification, and reduction ratio of each metal layer when stacking along the stacking direction to manufacture a metal laminate), (3) welding conditions (e.g., the welding current, welding voltage, welding wire feed rate, welding wire diameter, and welding wire protrusion length), and (4) the allowable range of the penetration length of each metal layer. This learning data is stored in the virtual storage device 21 of the AI ​​model learning device 20, and the AI ​​model learning device 20 performs machine learning using this learning data. As a result, the AI ​​model learning device 20 can generate or update a trained AI model 3a that obtains more appropriate output based on the results (performance) of the intermediate path generation process, based on machine learning using the learning data.

[0083] 15 is executed mainly by the AI ​​model learning unit 15 and the AI ​​model learning device 20 of the processor 2. In Fig. 15, the processor 2 determines whether or not there is a visual inspection device 30 connected to the 3D motion path generation device 1 by a known communication process via the communication I / F 6 (e.g., the presence or absence of a ping response) (St51). When the processor 2 determines that there is a visual inspection device 30 connected to the 3D motion path generation device 1 (St51, YES), the processor 2 inputs and acquires information indicating the results of a visual inspection by the visual inspection device 30 of a metal laminate (see, for example, Fig. 2 or 3 ) manufactured by a welding robot (not shown) based on the motion path generated by the 3D motion path generation device 1 (St52).

[0084] On the other hand, if the processor 2 determines that there is no appearance inspection device 30 connected to the 3D motion path generation device 1 (Step 51, NO), it inputs and acquires (Step 53) information indicating the results of a visual appearance inspection by an operator of a metal laminate (see, for example, FIG. 2 or 3 ) manufactured by a welding robot (not shown) based on the motion path generated by the 3D motion path generation device 1. This input may be realized, for example, by the operator operating the input device 4 of the 3D motion path generation device 1.

[0085] After step St52 or step St53, the processor 2 executes a learning process for the AI ​​model using information indicating the result of the appearance inspection acquired in step St52 or step St53 and the learning data (see above) stored in the virtual storage device 21 of the AI ​​model learning device 20 (St54). Note that the process of step St54 may be executed by the AI ​​model learning device 20, rather than by the processor 2 of the 3D movement path generation device 1. This allows the 3D movement path generation device 1 or the AI ​​model learning device 20 to generate or update a trained AI model 3a that obtains a more appropriate output based on information about the result of the appearance inspection (e.g., success or failure) and the result (performance) of the intermediate path generation process, based on machine learning using the learning data.

[0086] <Technology of the Present Disclosure> As described above, the present disclosure discloses the following technical ideas.

[0087] (Item 1) A 3D motion path generation device comprising: a memory (3) that stores at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and a processor (2) that generates a motion path of the welding robot for manufacturing the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data and stacking a plurality of the single metal layers in a predetermined stacking direction (direction DR1) by the welding robot. As a result, the 3D motion path generation device generates a 3D metal laminate model by stacking one metal layer corresponding to the 2D shape graphic data that resembles the shape of existing melted welding wire by the welding robot in the stacking direction (e.g., vertically upward). Therefore, the motion path of the welding torch equipped to the welding robot becomes the same shape represented by the 2D shape graphic data, and the metal laminate can be easily manufactured by the welding robot. This makes it possible to easily and efficiently generate motion paths when processing metal laminates using a welding robot, without the need for expensive dedicated equipment such as a 3D printer to manufacture metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0088] (Item 2) The 3D motion path generation device according to Item 1, wherein the processor, in lamination in the lamination direction when generating the motion path, arranges the nth (n: integer of 1 or more) metal layer and the (n+1)th metal layer at a predetermined lamination interval. This makes it easier for the 3D motion path generation device to stabilize the lamination of each metal layer that constitutes the 3D-shaped metal laminate.

[0089] (Item 3) The 3D motion path generation device according to Item 2, wherein the processor generates a shape of the (n+1)th metal layer that differs from the shape of the nth metal layer by changing parameters corresponding to the (n+1)th metal layer compared to parameters corresponding to the nth metal layer (e.g., circle radius, enlargement rate, reduction rate). This makes it possible for the 3D motion path generation device to arbitrarily set the shape of a single metal layer, and to generate a 3D motion path that conforms to the shape of any metal laminate desired by a user.

[0090] (Item 4) The 3D motion path generation device according to Item 2 or 3, wherein the processor generates and arranges one or more intermediate paths equivalent to the one metal layer between the nth metal layer and the (n+1)th metal layer in the stacking direction when the nth metal layer and the (n+1)th metal layer do not overlap in the stacking direction. This allows the 3D motion path generation device to stably stack the metal laminate so that the shape of the metal laminate does not collapse.

[0091] (Item 5) The 3D motion path generation device according to Item 4, wherein the memory stores an allowable range value for a penetration length of the one metal layer corresponding to welding conditions for the welding robot, and the processor generates one or more of the intermediate paths based on the allowable range value and a distance between the n-th one metal layer and the (n+1)-th one metal layer. This makes it possible for the 3D motion path generation device to appropriately generate intermediate paths for stably laminating the metal laminate so as not to distort the shape of the metal laminate.

[0092] (Item 6) The processor generates a plurality of the intermediate paths, and, if a distance between the (m-1)th intermediate path and the mth intermediate path (m: number of generated intermediate paths, m>n) is not within the allowable range, changes the current welding conditions used by the welding robot to other welding conditions and updates the shape of the mth intermediate path based on the change to the other welding conditions. This makes it possible for the 3D motion path generation device to adaptively change the shape of intermediate paths in accordance with usable welding conditions and appropriately generate a 3D motion path for manufacturing a metal laminate, which is an object to be manufactured, by a welding robot.

[0093] (Item 7) The 3D motion path generation device according to Item 5, wherein the processor corrects the penetration length tolerance value acquired from the memory using a trained model, and generates one or more of the intermediate paths based on the corrected penetration length tolerance value and the distance. This enables the 3D motion path generation device to update the penetration length tolerance value of one metal layer corresponding to current welding conditions to a more appropriate tolerance value using the trained AI model, thereby contributing to the generation of more appropriate intermediate paths.

[0094] (Item 8) The 3D motion path generation device according to Item 7, wherein the trained model is generated by machine learning using parameters corresponding to the one metal layer and welding conditions for the welding robot as input, and outputting an allowable range value for the penetration length of the one metal layer. This allows the 3D motion path generation device to generate or update a trained AI model that obtains more appropriate output based on machine learning using training data and taking into account results (performance) of an intermediate path generation process.

[0095] (Item 9) The 3D motion path generation device according to Item 8, wherein the trained model is generated by machine learning such that the input is the result of an appearance inspection of the metal laminate, added to the parameters and the welding conditions for the welding robot. This enables the 3D motion path generation device to generate or update a trained AI model that obtains a more appropriate output based on information on the appearance inspection result (e.g., success or failure) and the result (track record) of an intermediate path generation process, based on machine learning using training data.

[0096] (Item 10) The 3D motion path generation device according to Item 1, wherein the memory stores an allowable range value for the penetration length of the first metal layer corresponding to the welding conditions of the welding robot, and the processor outputs a first notification indicating insufficient contact between the nth metal layer of the first layer and the (n+1)th metal layer of the first layer when the nth metal layer of the first layer and the (n+1)th metal layer of the first layer overlap and the distance between the nth metal layer of the first layer and the (n+1)th metal layer of the first layer is less than a lower limit of the allowable range value. This makes it possible for the 3D motion path generation device to provide a user with a notification that appropriately reflects the state of the currently designed 3D motion path (e.g., the distance is too short and there is insufficient contact) in accordance with the length (distance) between the motion paths of adjacent single metal layers along the stacking direction.

[0097] (Item 11) The 3D motion path generation device according to Item 1, wherein the memory stores an allowable range value for the penetration length of the first metal layer corresponding to the welding conditions of the welding robot, and the processor outputs a second notification indicating that the lamination spacing between the nth metal layer and the (n+1)th metal layer is insufficient when the nth metal layer and the (n+1)th metal layer overlap and the distance between the nth metal layer and the (n+1)th metal layer is not less than a lower limit of the allowable range value. This makes it possible for the 3D motion path generation device to provide a user with a notification that appropriately reflects the state of the currently designed 3D motion path (e.g., a state in which the distance is too long and the lamination spacing is insufficient) in accordance with the length (distance) between the motion paths of adjacent single metal layers along the lamination direction.

[0098] (Item 12) A 3D motion path generation method executed by a 3D motion path generation device, comprising: storing in a memory at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data and stacking a plurality of the one metal layer layers in a predetermined stacking direction by the welding robot. Thus, according to the 3D motion path generation method, the welding robot can easily manufacture the metal laminate by stacking one metal layer corresponding to the 2D shape graphic data, which is assumed to be the shape of melted existing welding wire, in the stacking direction (e.g., vertically upward). Therefore, according to the 3D motion path generation method, motion paths for processing metal laminates using a welding robot can be easily and efficiently generated without the need for expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0099] (Item 13) A program for causing a 3D motion path generation device that is a computer to realize the following processes: storing in a memory at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data and stacking a plurality of the one metal layers in a predetermined stacking direction by the welding robot. As a result, according to the program, the 3D motion path generation device stacks one metal layer corresponding to the 2D shape graphic data that resembles the shape of existing melted welding wire by the welding robot in the stacking direction (e.g., vertically upward), thereby making it possible to easily manufacture a metal laminate by a welding robot. Therefore, according to the program, the 3D motion path generation device can easily and efficiently generate motion paths when processing metal laminates using a welding robot, even if the device does not have expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0100] Although various embodiments have been described above with reference to the drawings, it goes without saying that the present disclosure is not limited to such examples. It is clear that a person skilled in the art can conceive of various modifications, alterations, substitutions, additions, deletions, and equivalents within the scope of the claims, and it is understood that these also naturally fall within the technical scope of the present disclosure. Furthermore, the components of the various embodiments described above may be combined in any manner without departing from the spirit of the invention.

[0101] This application is based on a Japanese patent application (Patent Application No. 2024-134641) filed on August 9, 2024, the contents of which are incorporated herein by reference.

[0102] The present disclosure is useful as a 3D motion path generation device, a 3D motion path generation method, and a 3D motion path generation program that easily and efficiently generate motion paths when processing metal laminates using a welding robot, without the need for expensive dedicated equipment such as a 3D printer for manufacturing metal laminates or a dedicated application for generating motion paths required for processing metal laminates.

[0103] REFERENCE SIGNS LIST 1 3D motion path generating device 2 Processor 3 Memory 4 Input device 5 Output device 10 Internal bus 11 Overall control unit 12 Intermediate path generating unit 13 Overlap determining unit

Claims

1. A 3D motion path generation device comprising: a memory that stores at least 2D shape graphic data for melting a metallic welding wire by a welding robot and 3D shape data for a metal laminate produced by the welding robot; and a processor that generates a motion path for the welding robot to produce the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one metal layers in a predetermined stacking direction by the welding robot.

2. The 3D movement path generation device according to claim 1, wherein the processor stacks and arranges the nth (n: integer greater than or equal to 1)th metal layer and the (n+1)th metal layer at a predetermined stacking interval in the stacking direction when generating the movement path.

3. The 3D motion path generation device of claim 2, wherein the processor generates a shape of the (n+1)th metal layer to be different from the shape of the nth metal layer by changing parameters corresponding to the (n+1)th metal layer compared to parameters corresponding to the nth metal layer.

4. The 3D motion path generation device of claim 2, wherein the processor generates and places one or more intermediate paths equivalent to the one metal layer between the nth metal layer and the (n+1)th metal layer in the stacking direction when the nth metal layer and the (n+1)th metal layer do not overlap in the stacking direction.

5. The 3D motion path generation device according to claim 4, wherein the memory stores a tolerance value for the penetration length of the one metal layer corresponding to the welding conditions for the welding robot, and the processor generates one or more of the intermediate paths based on the distance between the nth one metal layer and the (n+1)th one metal layer and the tolerance value.

6. The 3D motion path generation device of claim 5, wherein the processor generates a plurality of the intermediate paths, and when the distance between the (m-1)th intermediate path and the mth intermediate path (m: number of generated intermediate paths, m>n) is not within the allowable range value, changes the current welding conditions used by the welding robot to other welding conditions and updates the shape of the mth intermediate path based on the change to the other welding conditions.

7. The 3D motion path generation device according to claim 5, wherein the processor corrects the tolerance value of the penetration length obtained from the memory using a trained model, and generates one or more of the intermediate paths based on the corrected tolerance value of the penetration length and the distance.

8. The 3D motion path generation device described in claim 7, wherein the trained model is generated by machine learning using parameters corresponding to the first metal layer and welding conditions for the welding robot as inputs, and outputting an allowable range value for the penetration length of the first metal layer.

9. The 3D motion path generation device according to claim 8, wherein the trained model is generated by machine learning to add the results of visual inspection of the metal laminate to the parameters and welding conditions for the welding robot as the input.

10. The 3D motion path generation device of claim 1, wherein the memory stores an allowable range value for the penetration length of the first metal layer corresponding to the welding conditions of the welding robot, and the processor outputs a first notification that there is insufficient contact between the nth metal layer of the first layer and the (n+1)th metal layer of the first layer when the nth metal layer of the first layer and the (n+1)th metal layer of the first layer overlap and the distance between the nth metal layer of the first layer and the (n+1)th metal layer of the first layer is less than a lower limit of the allowable range value.

11. The 3D motion path generation device of claim 1, wherein the memory stores an allowable range value for the penetration length of the first metal layer corresponding to the welding conditions of the welding robot, and the processor outputs a second notification that the lamination spacing between the nth metal layer and the (n+1)th metal layer is insufficient when the nth metal layer and the (n+1)th metal layer overlap and the distance between the nth metal layer and the (n+1)th metal layer is not less than the lower limit of the allowable range value.

12. A 3D motion path generation method executed by a 3D motion path generation device, comprising: storing in a memory at least 2D shape graphic data for melting a metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one metal layers in a predetermined stacking direction by the welding robot.

13. A program for causing a 3D motion path generation device, which is a computer, to perform the following processes: storing in memory at least 2D shape graphic data for melting metal welding wire by a welding robot and 3D shape data of a metal laminate to be manufactured by the welding robot; and generating a motion path for the welding robot to manufacture the metal laminate by forming the 2D shape welding wire melted by the welding robot into one metal layer based on the 2D shape graphic data, and stacking a plurality of the one metal layers in a predetermined stacking direction by the welding robot.

Images