Welding control method, welding system, welding control program, and multilayer build-up welding method

Abstract

In the automatic welding of multilayer build-up welding, the condition setting time can be reduced, and the appropriate welding condition can be automatically set. In a welding control method for performing multilayer build-up welding using a welding system provided with a welding robot, a robot control device, a camera, and a data processing device for processing image data obtained from the camera, the welding control method includes the steps of: before welding, using the welding robot; determining a welding condition at an arbitrary position on the welding start side of the workpiece to be welded; starting welding of a first layer in which the data processing device inputs the image data obtained from the camera and outputs a feature quantity of the image data to the robot control device based on the welding conditions, the robot control device controlling at least one of the welding conditions based on the feature quantity; and after the welding of the first layer, determining teaching point information after the second layer before the welding of the second layer on the basis of the construction information obtained by the control in the welding of the first layer.

Description

Technical Field

[0001] This invention relates to welding control methods, welding systems, welding control procedures, and multi-layer surfacing welding methods. Background Technology

[0002] In recent years, various manufacturing fields such as shipbuilding, bridge construction, and building construction have seen the emergence of automated welding technologies that use vision sensors (hereinafter also referred to as "cameras") to control welding based on welding images in order to achieve high welding quality. For example, Patent Document 1 discloses a welding control technology that aims to accurately recognize images even when there are deviations in image brightness, changes in the electric arc or molten pool, or changes in the presence or location of sputtering. The technology uses images obtained from a camera as input to a machine learning model, obtains state correlation information output from the machine learning model, and controls the welding mechanism that performs arc welding based on the obtained state correlation information, thereby enabling accurate image recognition even when there are deviations in image brightness.

[0003] Existing technical documents

[0004] Patent documents

[0005] Patent Document 1: Japanese Patent Application Publication No. 2018-192524

[0006] This allows for automated and precise welding using a camera. However, for automated welding to achieve high precision, the welding conditions, which serve as a reference, need to be set automatically at appropriate values. In particular, multi-layer welding, which involves forming multiple layers in multiple welding passes, requires automatically setting the welding conditions based on the number of welding passes. Furthermore, while automatic condition setting conventionally involves training or sensing the welding position, the larger the workpiece (hereinafter referred to as the "workpiece"), the longer the weld length, and the more areas require training or sensing, thus increasing the time spent on training or sensing. In other words, in automated welding of multi-layer welds, a function is sought that can reduce the time spent on condition setting and automatically set appropriate welding conditions. This function is arguably more important as the workpiece becomes larger. Patent Document 1 discloses a welding control method after setting conditions, but it does not consider the aforementioned function. Summary of the Invention

[0007] The problem that the invention aims to solve

[0008] The present invention was made in view of the above-mentioned problems, and its object is to provide a welding control method, welding system, welding control program, and multi-layer welding method that can reduce the time for setting conditions and automatically set appropriate welding conditions in automatic welding of multi-layer weld overlay.

[0009] Solution for solving the problem

[0010] The present invention is composed of the following structure.

[0011] (1) A welding control method, which is a welding control method for multi-layer overlay welding using a welding system,

[0012] Its features are,

[0013] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0014] The welding control method includes the following steps:

[0015] Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece.

[0016] Welding of the first layer is initiated based on the welding conditions. During welding of the first layer, the data processing device inputs image data obtained from the camera and outputs feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values.

[0017] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

[0018] (2) A welding control method, which is a welding control method for multi-layer surfacing welding using a welding system,

[0019] Its features are,

[0020] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera;

[0021] The welding control method includes the following steps:

[0022] Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece.

[0023] Welding of the first layer is initiated based on the welding conditions. During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values.

[0024] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for the second layer is determined before welding the second layer; and

[0025] Regarding the second layer and beyond, based on the construction information obtained from the control during the welding of the nth layer, before welding the (n+1)th layer, at least one of the following is determined: the teaching point information of the (n+1)th layer, the welding conditions of the teaching point information, and the stacking mode of the teaching point information, where n is a natural number greater than 2.

[0026] (3) A welding system for performing multi-layer overlay welding, wherein,

[0027] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0028] Before welding, the welding robot determines the welding conditions at any position on the welding start side of the workpiece.

[0029] The welding robot initiates welding of the first layer based on the welding conditions.

[0030] During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device.

[0031] The robot control device controls at least one of the welding conditions based on the feature quantity. After welding the first layer, the robot control device determines the teaching point information for subsequent layers before welding the second layer based on the construction information obtained from the control during the welding of the first layer.

[0032] (4) A welding control program, which is a welding control program for multi-layer overlay welding, wherein,

[0033] The welding control program is used by the robot control device to control the welding robot to achieve the following functions:

[0034] The welding robot performs the first layer of welding based on welding conditions determined at any position on the welding start side of the workpiece before welding.

[0035] In the welding of the first layer, at least one of the welding conditions is controlled based on feature quantities of image data, wherein the feature quantities of the image data are obtained from image data acquired from a camera that includes the molten pool, welding wire, and arc within the image; and

[0036] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

[0037] (5) A multi-layer welding method, which is performed using a welding system,

[0038] Its features are,

[0039] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0040] The multi-layer welding method includes the following steps:

[0041] Before welding, determine the welding conditions at any position on the welding start side of the workpiece.

[0042] Welding of the first layer is initiated based on the aforementioned welding conditions. During welding of the first layer, at least one of the welding conditions is controlled based on feature values ​​from image data, wherein the feature values ​​are obtained from image data acquired from the camera that includes the molten pool, welding wire, and arc within the image.

[0043] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

[0044] Invention Effects

[0045] According to the present invention, in the automatic welding of multi-layer overlay welding, the time for setting conditions can be reduced, and appropriate welding conditions can be automatically set. Attached Figure Description

[0046] Figure 1 This is a schematic diagram showing a structural example of the welding system of this embodiment.

[0047] Figure 2 This is a flowchart illustrating an example of welding control processing according to this embodiment.

[0048] Figure 3 This is a perspective view illustrating the configuration of the imaging device in this embodiment.

[0049] Figure 4 This is an example of a welding image captured by a camera device in the case of welding in a lateral orientation.

[0050] Figure 5 This is a block diagram illustrating a structural example of the data processing apparatus of this embodiment.

[0051] Figure 6 This is an explanatory diagram showing an example of a screen used for teaching exercises.

[0052] Figure 7 This is an explanatory diagram showing a specific example of a welding image obtained from welding in a lateral orientation, and an example of welding information within that welding image.

[0053] Figure 8 This is a conceptual diagram illustrating the offset amount from the welding line in this embodiment.

[0054] Figure 9 This is a conceptual diagram illustrating the method of adding and updating teaching points in this embodiment.

[0055] Explanation of reference numerals in the attached figures:

[0056] 50 Welding System

[0057] 100 mobile welding robots

[0058] 200 welding torch

[0059] 210 nozzle

[0060] 211 Welding Wire

[0061] 300 feed device

[0062] 400 welding power supply

[0063] 410 Power cables

[0064] 420 catheter

[0065] 430 Power Cable

[0066] 500 Protective Gas Supply Source

[0067] 510 Gas Tube

[0068] 600 Robot Control Device

[0069] 601 Data Retention Department

[0070] 602 Bevel Shape Information Calculation Section

[0071] 603 Welding Condition Acquisition Section

[0072] 604 Control Department

[0073] 610 Robot control cables

[0074] 620 power supply control cable

[0075] 700 camera device

[0076] 800 Data Processing Unit

[0077] 810 Main Body

[0078] 815 Non-volatile storage device

[0079] 816 Input / Output Interface

[0080] 817 Communication Interface

[0081] 818 Image Output Interface

[0082] 819 Calculation Department

[0083] 820 Input Section

[0084] 830 Display Section. Detailed Implementation

[0085] Hereinafter, a welding system according to an embodiment of the present invention will be described with reference to the accompanying drawings. It should be noted that in the drawings, the same reference numerals are used to indicate the correspondence of the same constituent elements. The welding system of the present invention is suitable for mobile welding robots, but is not limited to the structure of this embodiment. For example, it could also be a 6-axis welding robot or an automatic welding device with a drive unit such as a trolley. Furthermore, in this embodiment, as an example, single-sided welding in a lateral posture is selected, but it is not limited to this, and the construction method and welding posture are not particularly limited.

[0086] Furthermore, in this embodiment, as described later, the image is captured at least in a manner that includes the molten pool during welding in the image data, and feature points of the molten pool, welding wire, and arc are extracted. There are no particular limitations on the method for extracting these feature points, but in this embodiment, the data processing apparatus described later is used to extract the feature points.

[0087] Figure 1 This is a schematic diagram showing a structural example of the welding system according to this embodiment. The welding system 50 is configured to include a mobile welding robot 100, a feed device 300, a welding power source 400, a shielding gas supply source 500, a robot control device 600, an imaging device 700, and a data processing device 800. It should be noted that when the features of this embodiment are applied to a 6-axis welding robot, an automatic welding device with a drive unit such as a trolley, etc., as described above, further structures may be included accordingly. Furthermore, the various components constituting the welding system 50 can be communicatively connected using various wired or wireless communication methods. The communication method is not limited to one; multiple communication methods can be combined and connected.

[0088] (Robot control device)

[0089] The robot control device 600 is connected to the mobile welding robot 100 via a robot control cable 610 and to the welding power supply 400 via a power control cable 620. The robot control device 600 has a data storage unit 601 that stores training data, including predetermined motion modes, welding start positions, welding end positions, construction conditions, welding conditions, and a parameter table (hereinafter also referred to as "condition DB") of the mobile welding robot 100. Based on this training data, the robot control device 600 sends instruction information as commands to the mobile welding robot 100 and the welding power supply 400, controlling the motion of the mobile welding robot 100 and the welding conditions.

[0090] Alternatively, the robot control device 600 may also include a bevel shape information calculation unit 602 that calculates bevel shape information based on detection data obtained through sensing such as touch sensing before welding, and a welding condition acquisition unit 603 that performs a layered design based on at least one of the bevel shape information and condition DB, and sets or modifies the welding conditions of the aforementioned training data for each pass to obtain the welding conditions. Furthermore, the bevel shape information calculation unit 602 and the welding condition acquisition unit 603 constitute a control unit 604. The control unit 604 may be configured using, for example, a CPU (Central Processing Unit), an MPU (Micro Processing Unit), a DSP (Digital Signal Processor), or a FPGA (Field Programmable Gate Array). It should be noted that in this embodiment, welding condition setting utilizing the most preferred sensing method is applied from an efficiency point of view, but it is not limited to this; for example, direct training and welding condition setting may also be performed. Furthermore, in this invention, the sensing position or training position (hereinafter referred to as the "teaching position" or "teaching point") of the initial layer (hereinafter also referred to as the "first layer") is any position on the welding start side, as described later. It should be noted that "any position on the welding start side" means the welding start position or its vicinity. In this embodiment, the welding start position is set from the viewpoint of the accuracy of the bevel shape information, but the welding conditions for the welding start position can also be set based on the bevel shape information near the welding start position. Furthermore, the stacking design and welding condition settings for subsequent passes in this invention will also be described later.

[0091] Layer design refers to determining the appropriate number of layers and passes for a given bevel. A layer refers to the number of times weld passes overlap along the thickness direction of the plate. In multi-layer weld overlay design, maintaining a constant layer height for each layer is generally a primary consideration. Therefore, it is preferable to design the layer based on bevel shape information and to set or modify welding conditions based on the calculated layer design information. Here, layer design information can include the number of layers, the number of passes, and the layer height (≈ the height of one pass), but welding conditions are determined based at least on the calculated layer height. For example, if the gap between the start and end positions of welding is large, as welding progresses, the gap width increases, thus requiring an increase in deposition rate to maintain a constant layer height. In this case, welding conditions for deposition rate can include wire feed rate, welding speed, etc., and these welding conditions can be set or modified based on the bevel shape information and the layer design information.

[0092] Figure 2 This is a flowchart illustrating an example of welding control processing according to this embodiment. In this embodiment, the robot control device 600 begins welding under welding conditions based on training data that has been set or corrected according to sensing. The imaging device 700 captures images in such a way that the molten pool during welding is included in the image data. The data processing device 800 that acquires the captured image data extracts feature points. Then, the robot control device 600 receives correction signals for various processing from the information of each feature point as instruction information (hereinafter also referred to as "instructions") (S1). The robot control device 600 sequentially executes controls such as gap processing (S2), bar manipulation processing (S3), imitation, amplitude processing (S4), and speed processing (S5), which will be described later. The robot control device 600 processes each control according to the correction signals received during welding, and then outputs information updating the status of various control conditions to the data processing device 800 (S6). The above processing (S1 to S6) is repeated until the welding as intended is completed (S7).

[0093] The welding power source 400 supplies power to the welding wire 211 (which serves as a consumable electrode) and the workpiece Wo according to instructions from the robot control device 600, thereby generating an electric arc between the welding wire 211 and the workpiece Wo. Power from the welding power source 400 is delivered to the feed device 300 via the power cable 410, and from the feed device 300 to the welding torch (hereinafter referred to as "welding gun") 200 via the conduit 420. Furthermore, power from the welding power source 400 is supplied to the welding wire 211 via the conductive tip at the front end of the welding gun 200. It should be noted that the current during welding operations can be direct current (DC) or alternating current (AC), and its waveform is not particularly limited. Therefore, the current can also be a rectangular wave, a triangular wave, or other pulses.

[0094] Alternatively, the welding power source 400 may connect power cable 410 as the positive electrode to the welding torch 200 side and power cable 430 as the negative electrode to the workpiece Wo. It should be noted that this is the case of welding with reverse polarity. In the case of welding with positive polarity, the power cable of the positive electrode is connected to the workpiece Wo side, and the power cable of the negative electrode is connected to the welding torch 200 side.

[0095] (Protective gas supply source)

[0096] The protective gas supply source 500 consists of a container filled with protective gas, valves, and other auxiliary components. The protective gas is supplied from the protective gas supply source 500 to the feed device 300 via a gas pipe 510. The protective gas supplied to the feed device 300 is then supplied to the welding torch 200 via a conduit 420. The protective gas supplied to the welding torch 200 flows within the torch 200 and is guided towards the nozzle 210, exiting from the front end of the torch 200. For example, argon (Ar), carbon dioxide (CO2), or a mixture thereof can be used as the protective gas in this embodiment.

[0097] In this embodiment, the conduit 420 has a conductive path formed on its outer sheath for functioning as a power cable, a protective tube for the welding wire 211 is disposed inside the tube, and a flow path for the protective gas is formed. However, the conduit 420 is not limited to this; for example, a conduit formed by bundling a power supply cable and a protective gas supply hose around the protective tube for feeding the welding wire 211 to the welding torch 200. Alternatively, for example, the tube for conveying the welding wire 211 and the protective gas can be provided separately from the power cable.

[0098] (Feeding device)

[0099] The feed device 300 extracts the welding wire 211 and delivers it to the welding torch 200. The welding wire 211 fed by the feed device 300 is not particularly limited and is selected based on the nature of the workpiece Wo, the welding method, etc., for example, solid welding wire or flux-cored welding wire can be used. Furthermore, the wire diameter is not particularly limited, but in this embodiment, the preferred upper limit is 1.6 mm and the lower limit is 0.9 mm.

[0100] In addition, in this embodiment, a voltage is applied between the workpiece Wo and the welding wire 211, and the voltage drop phenomenon generated when the welding wire 211 contacts the workpiece Wo is used to sense the voltage. Figure 3 The surface of the bevel 10 shown is a contact sensor, etc., used as the detection mechanism. The detection mechanism is not limited to the contact sensor of this embodiment; an image sensor, a laser sensor, or a combination of these detection mechanisms may also be used. However, the contact sensor of this embodiment is preferred for the sake of simplicity of the device structure.

[0101] (Filming device)

[0102] The imaging device 700 (hereinafter also referred to as a "vision sensor" or "camera") is, for example, a camera equipped with a CMOS (Complementary Metal-Oxide-Semiconductor) as a vision sensor. There are no particular restrictions on the placement of the imaging device 700; it can be directly mounted on the mobile welding robot 100, or it can be fixed in a specific location in the vicinity as a monitoring camera. When the imaging device 700 is directly mounted on the mobile welding robot 100, the imaging device 700 moves in tandem with the movement of the mobile welding robot 100, capturing the periphery of the front end of the welding torch 200. Multiple cameras can constitute the imaging device 700. For example, multiple cameras with different functions and placement positions can be used to constitute the imaging device 700.

[0103] Furthermore, there are no particular limitations on the direction in which the imaging device 700 captures images. For example, if the direction of welding progress is forward, it can be configured to capture images from the front side, or from the side or rear side. Therefore, the shooting range captured by the imaging device 700 can be appropriately determined. It should be noted that in order to suppress interference from the welding torch 200, it is preferable to capture images from the front side; in this embodiment, it is configured to capture images from the front side. The captured image information is sent to the data processing device 800 and utilized on the data processing device 800 side. At this time, the data processing device 800 can, for example, extract any image from the captured image information at predetermined intervals and utilize it in the processing described later. The extraction method and extraction settings here can be switched, for example, depending on the structure and function of the imaging device 700, the performance of the data processing device 800, etc.

[0104] In this embodiment, a camera 700, directly mounted and fixed to the mobile welding robot 100, is used to capture welding images as moving images, with the target (object) included in the image data encompassing at least the workpiece Wo, the welding wire 211, and the arc within the capture range. It should be noted that various shooting settings for the welding images can be preset or switched according to the operating conditions of the welding system 50. Examples of shooting settings include frame rate, image pixel count, resolution, and shutter speed.

[0105] Figure 3 This is a perspective view illustrating the configuration of the imaging device 700 in this embodiment. It should be noted that the orientation of the welding torch 200 and the bevel varies depending on the welding posture; therefore... Figure 3The orientation shown is an example. In this embodiment, the workpiece Wo is a butt joint. The workpiece Wo consists of two metal plates joined together with a bevel. In this embodiment, a lateral orientation is taken as an example. In this case, the workpiece on the upper plate side is designated as W1 (hereinafter referred to as "upper plate W1"), and the workpiece on the lower plate side is designated as W2 (hereinafter referred to as "lower plate W2"). ​​In addition, a ceramic backing 14 is installed on the back side of the two joined workpieces W1 and W2. It should be noted that the backing 14 can be a metal backing or a structure without a backing. Therefore, the material of the backing is not particularly limited and can vary depending on the material of the workpiece Wo, etc. In the butt joint, arc welding is performed in one direction along the bevel. Hereinafter, the direction of welding progress will be referred to as the "welding line direction". Figure 3 In the image, arrows indicate the direction of welding progress. Therefore, the welding torch 200 is located behind the imaging device 700.

[0106] Figure 4 This is an example of a welding image captured by the imaging device 700 in the case of welding in a lateral orientation. Figure 4 The image data has coordinates in a coordinate plane formed by the X-axis and Y-axis. In this embodiment with a lateral orientation, the X-axis represents the welding line direction, and the Y-axis represents a direction orthogonal to the X-axis. The imaging device 700 captures an image containing the area of ​​the welding position of the workpiece Wo in arc welding. The captured image includes the molten pool, the welding wire 211, and the arc.

[0107] The vision sensor of the imaging device 700 in this embodiment is capable of continuously capturing still images of 1024×768 pixels. In other words, the imaging device 700 can capture welding images as moving images. The resolution of the still images that can be captured by the imaging device 700 is not particularly limited. For example, if the imaging device 700 is composed of multiple cameras, each camera can acquire welding images at different resolutions. Furthermore, preprocessing such as cutting arbitrary feature regions from the captured welding images can be performed before inputting the learned model (described later) to shorten processing time. The arbitrary feature regions can also be a fixed-size range arranged with a predetermined area centered. Additionally, the size of the arbitrary feature regions can be varied depending on the welding condition.

[0108] (Data processing device)

[0109] Figure 5This is a block diagram illustrating a structural example of the data processing apparatus 800 according to this embodiment. The data processing apparatus 800 is, for example, a computer. The computer is configured to include a main body 810, an input unit 820, and a display unit 830. The main body 810 is configured to include a CPU 811, a GPU (Graphical Processing Unit) 812, a ROM 813, a RAM 814, a non-volatile storage device 815, an input / output interface 816, a communication interface 817, an image output interface 818, and a calculation unit 819. The CPU 811, GPU 812, ROM 813, RAM 814, non-volatile storage device 815, input / output interface 816, communication interface 817, image output interface 818, and calculation unit 819 are communicatively connected to each other via a bus or signal lines.

[0110] The non-volatile storage device 815 stores a learning program 815A that performs deep learning using prescribed learning data, a learned model 815B generated by executing the learning program 815A, an information generation program 815C that generates welding information related to welding using the learned model 815B, and image data 815D. In addition to these, the non-volatile storage device 815 also stores an operating system and application programs.

[0111] The data processing device 800 performs various functions through the execution of programs by the CPU 811 and GPU 812. In this embodiment, the data processing device 800 performs the functions of generating a learned model using machine learning and various processing functions during actual welding using the learned model. Details of these functions will be described later. It should be noted that the data processing device 800 can also be separated into the function of generating the learned model and the function of performing control processing based on information output from the learned model during actual welding. From a general viewpoint, it is more preferable to separate the data processing device 800 according to their respective functions. The GPU 812 is used as a computing device for executing the learning program 815A and the information generation program 815C. The ROM 813 stores the BIOS (Basic Input Output System) and the like, which are executed by the CPU 811. The RAM 814 is used as a working area for programs read from the non-volatile storage device 815.

[0112] The input / output interface 816 is connected to the input unit 820, which consists of a keyboard, mouse, etc. An imaging device 700, serving as a vision sensor, is also connected to the input / output interface 816. Image data output from the imaging device 700 is transmitted to the CPU 811 via the input / output interface 816. The communication interface 817 is a communication module for wired or wireless communication. The image output interface 818 is connected to, for example, a display unit 830 consisting of a liquid crystal display or an organic EL (Electro-Luminescence) display, and outputs an image signal corresponding to the image data transmitted from the CPU 811 to the display unit 830. The calculation unit 819 cooperates with the CPU 811 and GPU 812 to perform various processes, such as calculating the geometric quantities used in the welding speed control of this embodiment. These geometric quantities will be described later.

[0113] (Model generation after learning)

[0114] The following description, in this embodiment, explains the feature points extracted from image data and the learned model for extracting these feature points. The learned model 815B in this embodiment is composed of a convolutional neural network and includes multiple convolutional layers and multiple pooling layers. It should be noted that the structure of the convolutional neural network is not limited to the structure described above; the number of layers and the structure can also be other structures.

[0115] After learning, the model 815B takes the welding image output from the imaging device 700 as input data and outputs feature information for calculating the gap width and welding progress direction of the molten pool. In this embodiment, the captured image input to the learned model 815B includes at least the molten pool, welding wire, and electric arc as targets. By inputting the captured image to the learned model 815B, feature points of each of the aforementioned targets or obtained from multiple targets are extracted. Based on the extracted feature points, the data processing device 800 obtains information such as the gap width and molten pool information (described later) as geometric quantities in real time. It should be noted that, as examples of geometric quantities of molten pool information mentioned here, the distance between the welding wire and the tip of the molten pool in the welding line direction, the width of the molten pool, or the area of ​​the molten pool can be given. For example, the molten pool information includes at least one of the distance between a specified position in the welding image and the tip of the molten pool, and the area of ​​the molten pool. The specified position can also be the position of the welding wire tip or the center point of the electric arc.

[0116] In this embodiment, the feature points used as welding information are the tip of the welding wire 211 (hereinafter also referred to as the "welding wire tip"), the center point of the arc (hereinafter also referred to as the "arc center"), the position of the tip of the weld pool to the left or right (or up or down in the case of a lateral orientation) relative to the welding progress direction, and the position of the ends of the weld pool to the left or right (or up or down in the case of a lateral orientation) relative to the welding progress direction. The input of feature points used as teaching data is performed by the operator specifying a specific position on the welding image according to the instructions on the operation screen supporting the teaching operation. Therefore, the teaching data is constructed by pairing the welding image with feature points specified by the operator as coordinate information for welding. In the learning process, parameters are adjusted by comparing the welding information output from the learning model with the welding information contained in the teaching data and feeding back the error. By repeatedly performing this process, machine learning progresses, generating a learned model 815B.

[0117] Figure 6 This is an explanatory diagram showing an example of a screen used in a teaching exercise. In Figure 6 The welding image shown includes a molten pool 15, welding wire 211, and an electric arc 16. Figure 6 In the image, the molten pool 15 is shown using a shaded line. Figure 7 This is an explanatory diagram showing a specific example of a weld image obtained from a lateral welding posture, and an example of welding information within that image. Here, for ease of explanation, the weld image is depicted with positions corresponding to coordinates represented by feature points. It should be noted that, as described above, the image has coordinates and forms a coordinate plane composed of the X and Y axes. It should be noted that, in relation to... Figure 6 as well as Figure 7 In the relevant explanations, "upper end" and "lower end" simply indicate the top and bottom in the image, and can be replaced with "right end" and "left end" depending on the welding posture and the orientation of the image.

[0118] In this embodiment, the imaging device 700 is arranged so that the welding line direction is parallel to the X-axis direction. Therefore, in this embodiment, the X-axis direction can also be referred to as the welding line direction. Furthermore, the Y-axis direction is perpendicular to the X-axis; in other words, it is the bevel width direction perpendicular to the welding line. Therefore, the Y-axis direction can also be referred to as the bevel width direction. It should be noted that the X-axis of the movable robot's movement in this embodiment is the welding line direction. Therefore, it can be said that the direction of the X-axis of the movable robot's movement is the same as the direction of the X-axis in the welding image, and the Y-axis direction is also the same as the direction of the Y-axis of the movable robot's movement and the direction of the Y-axis in the welding image.

[0119] In this embodiment, such as Figure 6As shown, the coordinates of the arc center PA (ArcX, ArcY), the wire tip PW (WireX, WireY), the lower molten pool tip PLD (Pool_Lead_Dx, Pool_Lead_Dy), the upper molten pool tip PLU (Pool_Lead_Ux, Pool_Lead_Uy), the lower molten pool tip PD (Pool_Dy), and the upper molten pool tip PU (Pool_Uy) are taught by the operator as feature points. The input of feature points is performed by the operator indicating specific positions on the screen. One example is assigning the coordinates of the boundary between the welding wire 211 and the arc as the position coordinates of the wire tip. Additionally, the lower molten pool tip, the upper molten pool tip, the lower molten pool tip, and the upper molten pool tip are examples of feature points related to the characteristics of the molten pool 15. For example, when the feature points at the lower and upper molten pool tips are known, the width of the molten pool 15 can be calculated.

[0120] (Geometric quantity data)

[0121] In this embodiment, the robot control device 600 or the data processing device 800 calculates at least two pre-set values: the difference in the X direction between the position of the weld pool tip and the position of the welding wire tip (hereinafter referred to as "LeadX"), and the difference between the position above and below the weld pool tip (hereinafter referred to as "LeadW"). In this embodiment, such numerical data calculated based on feature points is also called "geometric quantity data." It should be noted that the data processing device 800 calculates this geometric quantity data and sends it to the robot control device 600 at pre-set time intervals, allowing the robot control device 600 to perform welding control. It should also be noted that, regardless of the difference between the position above and below the weld pool tip, the position of the weld pool tip can be either above or below the weld pool tip, depending on the application. Furthermore, the robot control device 600 and the data processing device 800 can be integrated. Within the scope of not causing contradictions, the data processing device 800 may instead perform the processing that is performed by the robot control device 600, or the robot control device 600 may instead perform the processing that is performed by the data processing device 800.

[0122] <Welding Control Methods for Multi-Layer Welding>

[0123] For bevels on workpieces, the offset and width of the weld line typically vary depending on the location. Therefore, previously, sensing was required at multiple locations before welding. The larger the workpiece and the longer the weld length, the longer the sensing time became. In this invention, the pre-welding sensing is set only at the welding start point. A layered design is performed based on the bevel shape at the start point, and welding conditions are set based on the layered design. Welding begins with the set welding conditions, and during welding, the first layer of welding is completed simultaneously by using the aforementioned imaging device 700 and data processing device 800 to control welding conditions, mimic weld lines, and mimic width. Thus, even with larger workpieces, the sensing time can be shortened, improving welding efficiency. Furthermore, in the second layer, a layered design is performed at various locations along the weld line based on the construction information obtained in the first layer of welding, and welding conditions are set. Since the conditions at each location along the weld line can be set using the construction information obtained in the first layer of welding, time is not consumed. It should be noted that the locations where welding conditions are set on the weld line are designated as teaching locations in this invention. Furthermore, the optimal conditions for the second layer can be set based on the workpiece affected by the welding of the first layer. In other words, the conditions for the second layer can be set without time, taking into account the deformation caused by the thermal strain of the welding of the first layer.

[0124] It should be noted that welding can be performed on subsequent layers using welding conditions based on the layering design performed in the second layer, as in this embodiment. This is because, since layering design is performed using construction information such as gap width and imitation obtained in the welding of the first layer, it is possible to produce welded joints with stable quality. Therefore, layering design can be performed in the second layer, and the set conditions can be directly applied to the third layer and beyond. Alternatively, the second layer can also be processed using the imaging device 700 and the data processing device 800, and the teaching position can be determined again in the third layer to set welding conditions. Workpiece information can be obtained in each layer or pass, and layering design can be re-performed in the next layer or pass, and welding conditions can be reset. Hereinafter, the welding control method for multi-layer overlay welding will be explained in detail step by step.

[0125] [Before welding]

[0126] (STEP1: Sensing)

[0127] The sensing step before welding can be performed using the aforementioned contact sensor to determine the bevel shape, plate thickness, and start point by touch sensing. It should be noted that in this embodiment, the bevel shape and plate thickness are determined by touch sensing only at the welding start position. Based on the detected data of the bevel cross-sectional shape obtained from the sensing at the welding start position, a bevel shape information calculation step is performed to calculate the bevel shape information. Examples of bevel shape information include the bevel angle, plate thickness, bevel depth, estimated weld metal height, and gap. After this bevel shape information calculation step, the calculated data is input to the data holding unit 601 as a set value. In this embodiment, in the bevel shape information calculation step, at least one of the plate thickness, bevel depth, and estimated weld metal height, along with the gap, is calculated as bevel shape information and input to the data holding unit 601 as a set value. It should be noted that in the bevel shape information calculation step, if the gap is smaller than the wire diameter, the gap amount can be calculated as 0.

[0128] (STEP2: Setting welding conditions at the starting point)

[0129] Next, based on the bevel shape information data and conditions DB input to the data holding unit 601 as setting values, the setting conditions on the training program data are set in the stack design and each layer and pass corresponding to the stack design. In other words, a program is temporarily set to weld the entire length of the weld using welding conditions at the starting point position. In this embodiment, for example, welding current, arc voltage, welding speed, and oscillation conditions are set as welding conditions. It should be noted that in this embodiment, these are set, but they can also be modified. That is, setting is performed when the setting value has not yet been set, and the setting value is modified when it has been set.

[0130] [Welding of the first layer]

[0131] (STEP3: Welding control of the first layer)

[0132] Welding of the first layer begins using the welding conditions set at the starting point. In this embodiment, the first layer is formed in one pass, hence "first layer = one pass". Welding begins, and as described above, the imaging device 700 is activated to capture the welding image, which is then input to the data processing device 800. Geometric data is calculated from the welding image data and output to the robot control device 600. In this embodiment, the geometric data output to the robot control device 600 is referred to as LeadX and LeadW. The robot control device 600, having received the geometric data, is based on... Figure 2The welding process is controlled accordingly. It should be noted that in this embodiment, the first layer is an example of one pass, but there is no limit to the number of passes for each layer. Even if the welding of the first layer consists of two or more passes, the invention of this application can still be applied.

[0133] like Figure 2 As shown, the robot control device 600 receives information containing the aforementioned geometric data from the data processing device 800 at preset time intervals. This reception is also referred to as command reception. When receiving commands, the robot control device 600, in gap processing S2, continuously changes the welding conditions set at the starting point position based on the received geometric data of the LeadW. In this way, by changing the welding conditions based on the LeadW, the first layer of welding can be performed only at the starting point. It should be noted that in this embodiment, the geometric data of the LeadW is processed into the gap value in gap processing S2. It should be noted that in this embodiment, the unit of the gap value is set to mm, and the geometric data of the initial layer of the LeadW is called the "root gap". There are no particular limitations on the method of changing the welding conditions corresponding to the geometric data of the LeadW, but in this embodiment, a database (hereinafter also referred to as "machining DB") with pre-set welding conditions corresponding to the gap values ​​is prepared according to the welding mode, and the welding conditions can be changed based on the machining DB and the received geometric data of the LeadW. It should be noted that the welding mode mentioned here refers to fixed conditions such as the type of welding wire, the type of shielding gas, the tilt angle, and the pitch. These fixed conditions are also called "standard items." It should be noted that the welding data (DB) can be stored, for example, in a storage area such as the data storage unit 601 of the robot control device 600.

[0134] exist Figure 2In the skewing process S3, the robot control device 600 determines the skewing method based on information received from the instruction or information from preprocessing such as the gap processing S2. Skewing is also referred to as oscillation. In this embodiment, the robot control device 600 determines the oscillation method using conditions such as the offset amount, oscillation amplitude, end stop time (lower end side), and end stop time (upper end side) determined in the gap processing S2. If oscillation is not taken as a condition, the skewing process S3 is skipped, but in this embodiment, it is preferable to use oscillation in the welding of the first layer. This is because by using oscillation in the first layer, the welding quality of the initial layer is improved, and a flat weld bead shape can be obtained, thus making the welding of the second layer and subsequent layers more stable. It has a particularly significant effect in welding difficult postures such as lateral postures. Furthermore, it is even more preferable to change the welding conditions at each position during oscillation. Specifically, the welding conditions during oscillation can be changed when (a) moving from one end to the other end, (b) being at one end, and (c) being at the other end. When taking the lateral posture as an example, the current-voltage, upper end stopping time, lower end stopping time, and lower end moving from the upper end to the lower end can be changed during the following conditions: upper end stopping time, lower end stopping time, etc.

[0135] Regarding imitation and amplitude processing, in this embodiment, Figure 2 The imitation and amplitude processing S4 is performed as shown. In the imitation and amplitude processing S4, the robot control device 600 performs welding line imitation and width imitation based on the information received from the command or the information from preprocessing. The imitation method is not limited, but in this embodiment, the center position of LeadW is taken as the welding center position, and the welding line imitation control is performed based on the offset direction and offset amount (hereinafter also referred to as "offset amount from welding line") calculated based on the welding center position and the coordinate position PW of the welding wire tip. Taking the lateral posture in this embodiment as an example, it can be determined whether the offset is in the downward direction or the upward direction. In addition, the width imitation control is performed by calculating the amplitude correction amount based on the value of LeadW. Figure 8 This is a conceptual diagram illustrating the offset amount from the welding line in this embodiment.

[0136] In this invention, the conditions set based on the gap are further modified. This modification further improves control precision, maintaining stable weld quality even under various combinations of welding conditions, oscillation methods, welding postures, root gaps, backing gaps, and misalignments. The modified conditions can be any one of welding speed, welding current, arc voltage, or protrusion length, but from the viewpoint of ease of control and precision, it is preferable to modify at least the welding speed. In this embodiment, to modify the welding speed, in Figure 2Speed ​​processing is performed in S5. Figure 2 The speed processing S5 shown calculates a correction amount for the welding speed using the gap width value received from the command and LeadX, which is one of the weld pool information points for the welding progress direction. It should be noted that in the speed processing S5 of this embodiment, LeadW is used as the gap width value. Furthermore, while LeadX is used in this embodiment as the weld pool information for the welding progress direction, other geometrical data such as the weld pool area could also be used.

[0137] Based on the modified welding conditions obtained through the above processing, the robot control device 600 sends instruction values ​​to the mobile welding robot 100 and the welding power source 400. This process is repeated as described above. Figure 2 The process continues from steps S1 to S6 until the welding of the first layer is completed.

[0138] [After welding the first layer ~ Before welding the second layer]

[0139] (STEP4: Determining the Teaching Point Information for the Second Layer)

[0140] After the first layer of welding, the teaching point information at the welding location is determined based on the construction information obtained during the welding of the first layer. It should be noted that the teaching point information includes, for example, the teaching location (which can also be simply referred to as the "teaching point") and the bevel cross-sectional shape at the teaching location. Determining the teaching point information means adding or updating the teaching location and the bevel cross-sectional shape information at the teaching location.

[0141] Here, the construction information obtained during the welding of the first layer in this embodiment can be exemplified by the value of the gap width (LeadW) along the entire welded length or the offset from the weld line. The teaching point information is determined based on the gap width value or the offset from the weld line, or both.

[0142] For example, the teaching position, which serves as teaching point information, and the bevel cross-sectional shape at the teaching position are determined by adding or updating information at the welding start position and welding end position, the gap width, or the position where the offset from the welding line changes by a predetermined length. For example, in the case of welding with varying gap width, the teaching position and the bevel cross-sectional shape at the teaching position are added or updated at least at the position where the gap width changes by a predetermined length from the start position. Figure 9 This is a conceptual diagram illustrating a method for adding or updating the teaching position in this embodiment. Figure 9For example, when the gap width changes by 2mm, the teaching position and the bevel cross-sectional shape at the teaching position are added or updated. Positions of 4mm, 6mm, 8mm, and 10mm are added or updated as teaching positions, and the bevel cross-sectional shape at each teaching position is added or updated.

[0143] Regarding the determination of teach point information, a preferred method is to determine the teach point information based on both the value of the gap width and the offset from the welding line, either per half-cycle of swing or per robot control cycle.

[0144] (STEP5: Setting conditions for the second layer)

[0145] Based on the bevel shape information of each teaching position in the second layer obtained in STEP4, the stack-up design is performed with reference to condition DB. Alternatively, the stack-up design of each teaching point can be performed based on the gap width of each teaching position obtained during welding in the first layer and the bevel shape information other than the gap width obtained during sensing at the start point, and conditions can be set based on this stack-up design.

[0146] In this embodiment, the stack design information (hereinafter referred to as "reference stack design information") that serves as the reference gap width is pre-stored in the condition DB. Based on the reference stack design information, the robot control device 600 easily calculates the cross-sectional area equivalent to one pass at each taught position and determines the welding conditions according to this cross-sectional area. For example, if a fluctuation of 4 to 10 mm in the gap width can be anticipated based on the nature of the workpiece to be welded, a stack design with a central gap width of 7 mm is prepared in advance. The cross-sectional area equivalent to one pass at each taught position is calculated from the stack design with a gap width of 7 mm (number of layers - number of passes), and the welding conditions are set. At this time, for the cross-sectional area equivalent to one pass at the taught position, if the gap width is 10 mm, the cross-sectional area of ​​each pass relative to the reference stack design information is multiplied by (10 mm / 7 mm); if the gap width is 4 mm, it is multiplied by (4 mm / 7 mm), and so on, to easily calculate the cross-sectional area equivalent to one pass. Therefore, welding conditions for multiple teaching positions can be easily set using a pre-defined reference layer design information, thus significantly reducing the computational workload.

[0147] (STEP6: Welding after the second layer)

[0148] In this embodiment, welding passes after the second layer are welded based on the welding conditions of the stacked design generated at each teaching point. It should be noted that, alternatively, passes after the second layer may also be processed using the imaging device 700 and the data processing device 800, determining the teaching position again and setting conditions for the next layer, obtaining workpiece information in each layer or pass, re-designing the stacked structure in the next layer or pass, and resetting the welding conditions. Specifically, regarding passes after the second layer, it is also possible that, based on the construction information obtained from the control during welding of the nth layer (n being a natural number of 2 or more), before welding the (n+1)th layer, at least one of the teaching point information of the (n+1)th layer, the welding conditions of the teaching point information, and the stacked structure mode of the teaching point information is determined.

[0149] By using the above steps to control multi-layer welding, the welding operation time can be significantly shortened, and welding can be carried out under optimal conditions even after the second layer, thereby obtaining good welding quality.

[0150] (Selection of welding materials)

[0151] In this invention, the same welding material can be used for all layers, or the welding material can be changed layer by layer. The welding material used is selected from at least slag-cored wire, metal-cored wire, and solid wire. When using the same welding material for all layers, from the viewpoint of weld workability, it is preferable to use any one of slag-cored wire, metal-cored wire, and solid wire. Alternatively, when changing the welding material layer by layer, the same welding material can be used up to the first layer or the nth layer (n is a natural number chosen from 2 or higher), and a different welding material can be used from the (n+1)th layer onwards. Specifically, using slag-cored wire or metal-cored wire up to the difficult-to-weld first or second layer can yield a good weld bead shape. On the other hand, switching to solid wire from the third layer onwards can achieve good penetration.

[0152] This invention is not limited to the embodiments described above. Combining the various structures of the embodiments with each other, making changes based on the description in the specification and well-known technologies by those skilled in the art, and applying them are also intended by this invention and are included within the scope of the claims.

[0153] As stated above, the following matters are disclosed in this specification.

[0154] (1) A welding control method, which is a welding control method for multi-layer overlay welding using a welding system,

[0155] Its features are,

[0156] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0157] The welding control method includes the following steps:

[0158] Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece.

[0159] Welding of the first layer is initiated based on the welding conditions. During welding of the first layer, the data processing device inputs image data obtained from the camera and outputs feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values.

[0160] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

[0161] (2) The welding control method according to (1) is characterized in that,

[0162] Oscillation is performed during the welding of the first layer.

[0163] The welding control method includes the step of varying the welding conditions during oscillation at least when moving from one end to the other, when located at one end, and when located at the other end.

[0164] (3) The welding control method according to (2) is characterized in that,

[0165] The welding conditions during the oscillation are determined based on geometric data or construction information obtained during welding, according to a pre-set database.

[0166] (4) The welding control method according to (1) is characterized in that,

[0167] The construction information is at least either the gap or the offset from the weld line.

[0168] (5) The welding control method according to any one of (1) to (4), characterized in that,

[0169] The welding material should be selected from at least one of solid welding wire, slag-cored welding wire, and metal-cored welding wire.

[0170] When n is set to a natural number chosen from 2 or higher

[0171] The same welding material is used for welding the initial layer or up to the nth layer.

[0172] From the (n+1)th layer onwards, different welding materials are used compared to those used up to the nth layer.

[0173] (6) The welding control method according to any one of (1) to (4), characterized in that,

[0174] The welding material should be selected from at least one of solid welding wire, slag-cored welding wire, and metal-cored welding wire.

[0175] All layers are welded using the same welding material.

[0176] (7) A welding control method, which is a welding control method for multi-layer surfacing welding using a welding system,

[0177] Its features are,

[0178] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera;

[0179] The welding control method includes the following steps:

[0180] Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece.

[0181] Welding of the first layer is initiated based on the welding conditions. During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values.

[0182] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for the second layer is determined before welding the second layer; and

[0183] Regarding the second layer and beyond, based on the construction information obtained from the control during the welding of the nth layer, before welding the (n+1)th layer, at least one of the following is determined: the teaching point information of the (n+1)th layer, the welding conditions of the teaching point information, and the stacking mode of the teaching point information, where n is a natural number greater than 2.

[0184] (8) A welding system for performing multi-layer overlay welding, wherein,

[0185] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0186] Before welding, the welding robot determines the welding conditions at any position on the welding start side of the workpiece.

[0187] Welding of the first layer is initiated based on the aforementioned welding conditions.

[0188] During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device.

[0189] The robot control device controls at least one of the welding conditions based on the feature quantity. After welding the first layer, the robot control device determines the teaching point information for subsequent layers before welding the second layer based on the construction information obtained from the control during the welding of the first layer.

[0190] (9) A welding control program, which is a welding control program for multi-layer overlay welding, wherein,

[0191] The welding control program is used by the robot control device to control the welding robot to achieve the following functions:

[0192] The welding robot performs the first layer of welding based on welding conditions determined at any position on the welding start side of the workpiece before welding.

[0193] In the welding of the first layer, at least one of the welding conditions is controlled based on feature quantities of image data, wherein the feature quantities of the image data are obtained from image data acquired from a camera that includes the molten pool, welding wire, and arc within the image; and

[0194] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

[0195] (10) A multi-layer overlay welding method, which is performed using a welding system,

[0196] Its features are,

[0197] The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera.

[0198] The multi-layer welding method includes the following steps:

[0199] Before welding, determine the welding conditions at any position on the welding start side of the workpiece.

[0200] Welding of the first layer is initiated based on the aforementioned welding conditions. During welding of the first layer, at least one of the welding conditions is controlled based on feature values ​​from image data, wherein the feature values ​​are obtained from image data acquired from the camera that includes the molten pool, welding wire, and arc within the image.

[0201] After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

Claims

1. A welding control method, which is a welding control method for multi-layer surfacing welding using a welding system, characterized in that, The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera. The welding control method includes the following steps: Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece. Welding of the first layer is initiated based on the welding conditions. During welding of the first layer, the data processing device inputs image data obtained from the camera and outputs feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values. After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

2. The welding control method according to claim 1, characterized in that, Oscillation is performed during the welding of the first layer. The welding control method includes the step of varying the welding conditions during oscillation at least when moving from one end to the other, when at one end, and when at the other end.

3. The welding control method according to claim 2, characterized in that, The welding conditions during the oscillation are determined based on geometric data or construction information obtained during welding, according to a pre-set database.

4. The welding control method according to claim 1, characterized in that, The construction information is at least either the gap or the offset from the weld line.

5. The welding control method according to any one of claims 1 to 4, characterized in that, The welding material should be selected from at least one of solid welding wire, slag-cored welding wire, and metal-cored welding wire. When n is set to a natural number chosen from 2 or higher The same welding material is used for welding the initial layer or up to the nth layer. From the (n+1)th layer onwards, different welding materials are used compared to those used up to the nth layer.

6. The welding control method according to any one of claims 1 to 4, characterized in that, The welding material should be selected from at least one of solid welding wire, slag-cored welding wire, and metal-cored welding wire. All layers are welded using the same welding material.

7. A welding control method, which is a welding control method for multi-layer surfacing welding using a welding system, characterized in that, The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera. The welding control method includes the following steps: Before welding, the welding robot is used to determine the welding conditions at any position on the welding start side of the workpiece. Welding of the first layer is initiated based on the welding conditions. During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature values. After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information of the second layer is determined before the welding of the second layer. as well as Regarding the second layer and beyond, based on the construction information obtained from the control during the welding of the nth layer, before welding the (n+1)th layer, at least one of the following is determined: the teaching point information of the (n+1)th layer, the welding conditions of the teaching point information, and the stacking mode of the teaching point information, where n is a natural number greater than 2.

8. A welding system for performing multi-layer overlay welding, wherein, The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera. Before welding, the welding robot determines the welding conditions at any position on the welding start side of the workpiece. Welding of the first layer is initiated based on the aforementioned welding conditions. During the welding of the first layer, the data processing device inputs image data obtained from the camera and outputs the feature values ​​of the image data to the robot control device. The robot control device controls at least one of the welding conditions based on the feature quantity. After welding the first layer, the robot control device determines the teaching point information for subsequent layers before welding the second layer based on the construction information obtained from the control during the welding of the first layer.

9. A welding control program, which is a welding control program for multi-layer overlay welding, wherein, The welding control program is used by the robot control device to control the welding robot to achieve the following functions: The welding robot performs the first layer of welding based on welding conditions determined at any position on the welding start side of the workpiece before welding. In the welding of the first layer, at least one of the welding conditions is controlled based on feature quantities of image data, wherein the feature quantities of the image data are obtained from image data acquired from a camera that includes the molten pool, welding wire, and arc within the image; and After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

10. A multi-layer overlay welding method, performed using a welding system, characterized in that, The welding system includes at least a welding robot, a robot control device for controlling the welding robot, a camera, and a data processing device for processing image data obtained from the camera. The multi-layer welding method includes the following steps: Before welding, determine the welding conditions at any position on the welding start side of the workpiece. Welding of the first layer is initiated based on the aforementioned welding conditions. During welding of the first layer, at least one of the welding conditions is controlled based on feature values ​​from image data, wherein the feature values ​​are obtained from image data acquired from the camera that includes the molten pool, welding wire, and arc within the image. After the welding of the first layer, based on the construction information obtained from the control during the welding of the first layer, the teaching point information for subsequent layers is determined before the welding of the second layer.

Images