A welding robot gun cleaning control system, method and storage medium

Abstract

The embodiment of the specification provides a welding robot cleaning gun control system, method and storage medium, which comprises a cleaning gun device and a processor. The cleaning gun device is configured to perform a cleaning gun operation on at least one welding gun. The processor is configured to: determine at least one feature sequence of a target welding gun based on equipment parameters, working parameters and welding characteristics of a target workpiece, wherein the at least one feature sequence comprises a welding quality sequence; in response to the at least one feature sequence satisfying a first preset condition, determine a target cleaning time point of the target welding gun based on the at least one feature sequence; and control the cleaning gun device to perform the cleaning gun operation on the target welding gun at the target cleaning time point.

Description

Technical Field

[0001] This specification relates to the field of welding technology, and in particular to a welding robot torch cleaning control system, method, and storage medium. Background Technology

[0002] During welding, spatter accumulates on the welding torch nozzle and contact tip, affecting weld quality and the lifespan of welding equipment (such as the welding torch). Different workpiece materials, welding wire types, and welding parameters result in varying degrees of spatter adhesion to the welding torch's contact tip and / or nozzle, requiring timely cleaning to ensure both efficient welding and high-quality results. Furthermore, different spatter conditions place higher demands on cleaning tools (such as reamers); improper cleaning methods can cause physical damage to these tools, affecting cleaning effectiveness and tool lifespan.

[0003] Therefore, a welding robot torch cleaning control system, method, and storage medium are provided to improve the efficiency of welding torch cleaning. Summary of the Invention

[0004] One embodiment of this specification provides a welding robot torch cleaning control system, including a torch cleaning device and a processor. The torch cleaning device is configured to perform torch cleaning operations on at least one welding torch. The processor is configured to: determine at least one feature sequence of the target welding torch based on equipment parameters, operating parameters, and welding characteristics of the target workpiece, the at least one feature sequence including a welding quality sequence; in response to the at least one feature sequence satisfying a first preset condition, determine a target cleaning time point for the target welding torch based on the at least one feature sequence; and control the torch cleaning device to perform torch cleaning operations on the target welding torch at the target cleaning time point.

[0005] One embodiment of this specification provides a welding robot torch cleaning control method, which is executed by a processor of a welding robot torch cleaning control system, including: determining at least one feature sequence of a target welding torch based on equipment parameters, working parameters, and welding characteristics of the target workpiece, wherein the at least one feature sequence includes a welding quality sequence; determining a target cleaning time point of the target welding torch based on the at least one feature sequence in response to the at least one feature sequence satisfying a first preset condition; and controlling the torch cleaning device to perform a torch cleaning operation on the target welding torch at the target cleaning time point.

[0006] One embodiment of this specification provides a computer-readable storage medium that stores computer instructions. When a computer reads the computer instructions from the storage medium, the computer executes the aforementioned welding robot torch cleaning control method. Attached Figure Description

[0007] This specification will be further described by way of exemplary embodiments, which will be described in detail with reference to the accompanying drawings. These embodiments are not limiting; in these embodiments, the same reference numerals denote the same structures, wherein:

[0008] Figure 1 This is a schematic diagram of a welding robot torch cleaning control system according to some embodiments of this specification;

[0009] Figure 2 This is an exemplary flowchart of a welding robot torch cleaning control method according to some embodiments of this specification;

[0010] Figure 3 This is an exemplary flowchart of another welding robot torch cleaning control method according to some embodiments of this specification;

[0011] Figure 4 This is a schematic diagram illustrating the determination of a model based on sequences shown in some embodiments of this specification;

[0012] Figure 5 This is a schematic diagram of the model determined based on the parameters shown in some embodiments of this specification. Detailed Implementation

[0013] To more clearly illustrate the technical solutions of the embodiments in this specification, the accompanying drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are merely some examples or embodiments of this specification. For those skilled in the art, these drawings can be applied to other similar scenarios without creative effort. Unless obvious from the context or otherwise specified, the same reference numerals in the drawings represent the same structures or operations.

[0014] It should be understood that the terms “system,” “device,” “unit,” and / or “module” used herein are one way to distinguish different components, elements, parts, sections, or assemblies at different levels. However, if other terms can achieve the same purpose, they may be replaced by other expressions.

[0015] Unless the context clearly indicates an exception, words such as "a," "an," "a kind," and / or "the" do not specifically refer to the singular and may also include the plural. Generally speaking, the terms "comprising" and "including" only indicate the inclusion of explicitly identified steps and elements, which do not constitute an exclusive list, and the method or apparatus may also include other steps or elements.

[0016] Flowcharts are used in this specification to illustrate the operations performed by the system according to embodiments of this specification. It should be understood that the preceding or following operations are not necessarily performed in exact order. Instead, the steps can be processed in reverse order or simultaneously. Furthermore, other operations can be added to these processes, or one or more steps can be removed from them.

[0017] Figure 1 This is a schematic diagram of a welding robot torch cleaning control system according to some embodiments of this specification.

[0018] like Figure 1 As shown, the welding robot torch cleaning control system 100 includes a processor 110 and a torch cleaning device 120.

[0019] Processor 110 is used to analyze and / or process information and / or data in the welding robot torch cleaning control system 100. In some embodiments, processor 110 may be a computer processing device. For example, processor 110 may include a central processing unit (CPU), application-specific integrated circuit (ASIC), digital signal processor (DSP), microcontroller unit, or any combination thereof. In some embodiments, processor 110 may be a single server or a group of servers.

[0020] The processor 110 can be used to control one or more components (such as the cleaning device 120) in the welding robot torch cleaning control system 100. For example, the processor 110 can send control commands (such as program instructions or control signals) to the cleaning device 120 to instruct the cleaning device 120 to perform a cleaning operation on at least one welding torch. In some embodiments, the processor 110 can control the cleaning device 120 to perform a cleaning operation on a target welding torch at a target cleaning time. In some embodiments, the processor 110 can control one or more cleaning tools of the cleaning device 120 to perform a cleaning operation on the target welding torch at a target cleaning time based on cleaning parameters.

[0021] The cleaning device 120 is used to clean at least one welding torch (such as welding torch 140).

[0022] Welding torch 140 can refer to various welding torches that require cleaning during welding operations. Examples include gas welding torches, argon arc welding torches, metal inert gas (MIG) shielded welding torches (MIG torches), and metal active gas (MAG) shielded welding torches (MAG torches), etc. Figure 1 As shown, the welding torch 140 includes welding torch 140-1, welding torch 140-2, ..., welding torch 140-n.

[0023] In some embodiments, the welding torch 140 includes the welding torch of an automated welding system (not shown), such as a MIG or MAG welding torch of a welding robot. In some embodiments, one or more components (such as processor 110) of the welding robot torch cleaning control system 100 are communicatively connected to the automated welding system (such as the welding robot), and the processor 110 can send information and / or instructions to the welding robot to instruct the welding robot that its welding torch needs to be cleaned at a target cleaning time.

[0024] The torch cleaning operation refers to one or more operations within the torch cleaning process (or torch cleaning procedure) of a welding operation. It includes cleaning and maintaining one or more torch components of the target welding torch (such as welding torch 140-1), including but not limited to contact tips and nozzles. The torch cleaning device 120 performs the torch cleaning operation using various preset cleaning tools. For example, these tools include a reamer for removing spatter, a copper brush, a needle for removing blockages, and a spraying device for applying anti-spatter agent.

[0025] In some embodiments, the welding robot torch cleaning control system 100 further includes a sensing device 130. The sensing device 130 includes, but is not limited to, various acoustic and optical devices configured according to actual needs, used to acquire various monitoring data related to torch cleaning (such as acoustic data and optical data). Figure 1 As shown, the sensing device 130 includes an ultrasonic device 131 and an image acquisition device 132.

[0026] The ultrasonic device 131 includes an ultrasonic transmitting / receiving module integrated in the torch cleaning device 120 for acquiring ultrasonic monitoring data of one or more components of the target welding torch. In some embodiments, the ultrasonic monitoring data includes second spatter data (such as ultrasonic reflection data of spatter from the conductive tip and / or the inner wall of the nozzle) for analyzing and / or processing information related to the second spatter (such as thickness, positional distribution, etc.).

[0027] Image acquisition device 132 includes optical devices such as a camera, used to acquire image monitoring data of the target welding torch. In some embodiments, the image monitoring data includes first spatter data (such as spatter images of the outer surface of the conductive tip and / or nozzle), used to analyze and / or process information related to the first spatter (such as thickness, positional distribution, etc.). In some embodiments, the image monitoring data includes image data related to the weld in the welding process (such as weld surface images), used to determine information related to the weld and / or weld pool, in order to determine the cleaning time point of the target welding torch. For more information on first spatter data and second spatter data, see [link to relevant documentation]. Figure 2 And its description.

[0028] In some embodiments, the welding robot torch cleaning control system 100 further includes a storage device (not shown in the figures), which includes, but is not limited to, a mass storage device, a removable storage device, volatile read / write memory, or any combination thereof. The storage device is used to store data and / or instructions. For example, the storage device may store data and / or instructions used by the processor 110 to execute the exemplary methods described in this specification (such as those shown in the flowcharts). As another example, the storage device may be used to store various types of monitoring data related to torch cleaning (such as ultrasonic monitoring data, image monitoring data, etc.).

[0029] The above description is for illustrative purposes only, and actual application scenarios may vary.

[0030] It should be noted that the welding robot torch cleaning control system 100 is provided for illustrative purposes only and is not intended to limit the scope of this specification. Various modifications or variations can be made by those skilled in the art based on the description in this specification. However, such modifications and variations will not depart from the scope of this specification.

[0031] Figure 2 This is an exemplary flowchart of a welding robot torch cleaning control method according to some embodiments of this specification.

[0032] In some embodiments, process 200 may be executed by a processor (such as processor 110) of a welding robot torch cleaning control system (hereinafter referred to as torch cleaning control system). Figure 3 As shown, process 200 includes the following steps.

[0033] Step S210: Based on the equipment parameters and working parameters of the target welding torch and the welding characteristics of the target workpiece, determine at least one feature sequence of the target welding torch, wherein the at least one feature sequence includes a welding quality sequence.

[0034] A target welding torch refers to a welding torch that needs to be cleaned. It can be a welding torch used in various welding applications (such as automotive manufacturing and ship maintenance) to perform welding operations. For example, a target welding torch could be the welding torch of a welding robot performing the welding operation. For more information on welding torches, see [link to relevant documentation]. Figure 1 And its description.

[0035] Equipment parameters refer to the physical parameters related to the target welding torch, including but not limited to the nozzle parameters (such as structure and material parameters), contact tip parameters (such as structure and material parameters), and welding wire parameters (such as welding wire size and material parameters).

[0036] Working parameters refer to the process parameters of the target welding torch during welding operations, including but not limited to current, voltage, shielding gas ratio, and moving speed.

[0037] The equipment parameters and operating parameters of the target welding torch can be pre-stored in the torch cleaning control system (such as a storage device), from which the processor can retrieve them for subsequent processing.

[0038] The target workpiece refers to the raw material or component that needs to be welded by the target welding torch. For example, the target workpiece includes, but is not limited to, steel plates, aluminum plates, pipes, etc., that need to be welded.

[0039] The welding characteristics of the target workpiece refer to the weld-related features in the welding operation (such as the welding process). Weld-related features include, but are not limited to, the material of the target workpiece (such as carbon steel, stainless steel, aluminum alloy, etc.), weld characteristics (such as weld porosity, dimensions (such as width, length), appearance, and color), and weld pool characteristics (such as weld pool depth, weld pool width, and weld pool temperature). The processor can determine the welding characteristics using image monitoring data collected by various sensors (such as image acquisition device 132) during the welding operation of the target welding torch.

[0040] At least one characteristic sequence of the target welding torch refers to one or more data sequences of data related to the target welding torch changing over time during the welding operation. This data related to the target welding torch during the welding operation includes, but is not limited to, the target welding torch's operating parameters (such as current and voltage), welding progress, weld characteristics, and weld pool characteristics. For example, characteristic sequences include operating parameter sequences, welding progress sequences, and weld characteristic sequences.

[0041] In some embodiments, the feature sequence includes a welding quality sequence. Welding quality reflects the welding effect of the target welding torch on the target workpiece, and can be represented based on a quality score, such as a value in the range [0, 100]. The larger the value, the better the welding effect and the higher the welding quality.

[0042] The processor can determine the welding quality based on weld characteristics (such as porosity) at any given time point (e.g., the current moment). For example, based on the weld characteristics at the current moment, the processor can retrieve or match the corresponding welding quality from a preset reference table (e.g., a data mapping table).

[0043] The welding quality sequence is generated based on the welding quality at multiple time points. These multiple time points can be in the form of a time series with preset time intervals (such as 10 seconds).

[0044] In some embodiments, the welding quality sequence includes welding quality at multiple time points within a preset future time period (e.g., 10 minutes, 1 hour, etc.), reflecting the trend of welding quality change within the preset future time period. In some embodiments, the processor determines the welding quality sequence based on the target welding torch's equipment parameters, operating parameters, and welding characteristics using a sequence determination model. See more details... Figure 4 And its description.

[0045] Step S220: In response to at least one feature sequence satisfying a first preset condition, the target cleaning time point of the target welding torch is determined based on at least one feature sequence.

[0046] The first preset condition refers to a condition that at least one characteristic sequence must meet. It can be determined based on experience or the actual needs of the welding operation. For example, the first preset condition could be that the current welding progress sequentially reaches a preset welding progress (such as 30%, 50%).

[0047] In some embodiments, the first preset condition relates to weld characteristics and / or weld pool characteristics. The first preset condition includes one of the following: the surface uniformity of the weld is greater than a uniformity threshold, the weld pool width is within a preset weld width threshold range, and the weld pool depth is within a preset weld depth threshold range.

[0048] Surface uniformity reflects the smoothness of a weld. A processor can determine the surface uniformity of a weld based on weld features. For example, the processor uses image analysis algorithms to determine weld features (such as weld width, the number and distribution of bumps and depressions on the weld surface) based on surface image data, and then determines the surface uniformity of the weld based on these features. For instance, the more uniform the weld width and the fewer the bumps and depressions, the greater the surface uniformity. In some embodiments, the processor retrieves or matches the corresponding surface uniformity from a preset correspondence table between weld features and surface uniformity.

[0049] The preset weld width threshold range and preset weld depth threshold range are preset according to the actual situation of the welding operation (such as the type of welding torch, the material properties of the target workpiece, etc.). For example, for MIG welding of low carbon steel, the preset weld width threshold range can be set to 5-8mm and the preset weld depth threshold range can be set to 2-4mm.

[0050] In some embodiments of this specification, considering the linkage between the welding process and the torch cleaning process in actual welding operations, the execution of the welding process is evaluated by setting a first preset condition, so as to evaluate the appropriate entry point for the subsequent torch cleaning process and improve the efficiency of the evaluation.

[0051] In some embodiments, in response to a first preset condition being met, the processor determines a target feature sequence from at least one feature sequence. The target feature sequence refers to one or a portion of the at least one feature sequence.

[0052] In some embodiments, the target feature sequence includes a target quality sequence corresponding to the welding quality sequence. The processor can determine the target time point from the welding quality sequence where the welding quality first falls below a preset scoring threshold (e.g., 60), and use the welding quality sequence segment up to that target time point as the target quality sequence. The target time point can be referred to as the predicted cleanup time point.

[0053] The target quality sequence is used to determine the estimated rate of decline in weld quality. The estimated rate of decline is calculated based on the rate of decline of weld quality scores over a unit time interval (e.g., 10 seconds). As an example, in the target quality sequence, the weld quality score corresponding to the current time point t1 is s1, and the target time point t... n The corresponding welding quality score is s n t1 and t n If the number of time intervals (e.g., 10 seconds) is k1, then the predicted descent rate v is... p =(s1-s n ) / k1. Wherein, the target time point in the target quality sequence is the predicted cleanup time point.

[0054] The target cleaning time point refers to the actual time when the target welding torch needs to be cleaned. The target cleaning time point can be the current moment or a future moment (such as the predicted cleaning time point).

[0055] In some embodiments, the processor calibrates the predicted cleaning time point based on actual welding quality data to determine the target cleaning time point. For example, the processor acquires the current time point t1 and the time point t after t1 in real time. m The weld characteristics were determined, and the time point t was identified. m Welding quality score m , where t1 and t m The time interval between them is k2, where k2 is less than k1. Then, using a similar method to determine the estimated rate of decline, the actual rate of decline in the welding quality score is calculated. It should be noted that k2 is a preset value (e.g., 3, 4, etc.). The processor can start from the current time point t1 and continuously monitor the changes in weld characteristics over k2 time intervals to obtain the time point t. m Corresponding welding quality score s m This allows for the assessment of changes in the actual welding quality score.

[0056] If the actual descent rate is less than or equal to the estimated descent rate, the processor will use the predicted cleanup time point as the target cleanup time point.

[0057] If the actual rate of decline is greater than the predicted rate of decline, it indicates that the actual time point when the actual weld quality score drops below the preset score threshold (e.g., 60) will be earlier than the predicted cleaning time point. Therefore, the target cleaning time point needs to be advanced. The target cleaning time point can be calculated based on the actual rate of decline and the preset score threshold.

[0058] In some embodiments of this specification, considering the complexity of actual welding operations, a more accurate target cleaning time can be obtained by combining the predicted cleaning time with the real-time monitoring of the welding quality score change trend. This avoids frequent welding torch cleaning affecting the progress of the welding operation, and also avoids cleaning the welding torch too late, which would affect the quality of the welding operation.

[0059] Step S230: Control the cleaning device to perform a cleaning operation on the target welding torch at the target cleaning time point.

[0060] In some embodiments, the processor may generate instructions and / or alarm messages to instruct the target welding torch to be cleaned at the target cleaning time.

[0061] In some embodiments, at the target cleaning point, the processor determines cleaning parameters based on the spatter distribution characteristics of the target welding torch to control one or more cleaning tools to perform torch cleaning operations on the target welding torch. For more information on cleaning parameters, see [link to relevant documentation]. Figure 3 And its description.

[0062] In some embodiments of this specification, feature sequences are dynamically generated to accurately determine the target cleaning time, avoiding resource waste or cleaning delays caused by cleaning the target welding torch at fixed intervals, and improving the maintenance efficiency of the welding torch.

[0063] Figure 3 This is an exemplary flowchart of another welding robot torch cleaning control method according to some embodiments of this specification.

[0064] In some embodiments, process 300 may be executed by a processor (such as processor 110) of the welding robot torch cleaning control system. Figure 3 As shown, process 300 includes the following steps.

[0065] Step S310: Obtain spatter data from the target welding torch.

[0066] Spatter data refers to data related to spatter generated by the target welding torch during welding operations. Spatter includes, but is not limited to, metallic deposits distributed on the target welding torch, around the weld, and on the surface of the target workpiece.

[0067] In some embodiments, the processor obtains spatter data through a sensing device (such as sensing device 130), and the spatter data includes data in various forms such as images and videos. It should be noted that the processor can directly obtain spatter data from a third-party system (such as an automated welding system), which collects the spatter data when the automated welding system (such as a welding robot) performs welding operations on the target welding torch.

[0068] In some embodiments, the spatter data includes first spatter data of the visible area of ​​the target welding torch and second spatter data of the inner wall of the target welding torch.

[0069] The first spatter refers to the visible spatter from the target welding torch. In some embodiments, the first spatter data includes spatter data from the outer surface of the conductive tip and / or nozzle of the target welding torch, and the first spatter data is determined by an image sequence acquired by an image acquisition device during the welding operation of the target welding torch. The image sequence can be used to determine the first spatter features corresponding to the first spatter. For example, the processor determines the first spatter features (such as size (e.g., diameter, thickness, etc.), location, quantity, etc.) based on an image recognition algorithm using one or more frames from the image sequence.

[0070] The second spatter refers to the invisible spatter from the target welding torch. In some embodiments, the second spatter data includes spatter data from the inner wall of the conductive tip and / or nozzle, which is determined by ultrasonic reflection data acquired by an ultrasonic device at the target cleaning point. The ultrasonic reflection data can be used to determine the second spatter characteristics corresponding to the second spatter. For example, the processor determines the second spatter characteristics (such as size, location, quantity, etc.) based on the ultrasonic reflection data and a time-difference method.

[0071] Step S320: Based on the spatter data, determine the spatter distribution characteristics of the target welding torch.

[0072] Spatter distribution characteristics are used to reflect the distribution of spatter from the target welding torch, including the distribution characteristics of visible and invisible spatter.

[0073] In some embodiments, the processor determines splash distribution characteristics based on a first splash feature and a second splash feature. The splash distribution characteristics can be presented in the form of a splash heatmap.

[0074] In some embodiments, the processor performs mutual registration (e.g., position coordinates) based on the first spatter feature and the second spatter feature to obtain the positional distribution of visible and invisible spatter in the same space (e.g., spatial coordinate system). The processor can further perform segmentation and labeling of visible and invisible spatter regions based on device parameters (e.g., the structure and material of the conductive tip and nozzle), the positional distribution of visible and invisible spatter, to obtain a spatter thermal map of the target welding torch.

[0075] The spatter heatmap is used to comprehensively reflect the distribution characteristics of spatter in the visible external area and the invisible internal wall area of ​​the conductive nozzle and / or spray nozzle. For example, different areas (such as the visible external area and the invisible internal wall area) are marked with different colors (such as red, blue, etc.) and / or transparency; visible spatter of different sizes (such as thickness) and invisible spatter of different sizes (such as thickness) are marked with different color depths (such as dark red, light blue, etc.) to distinguish between sparse and thick areas. Different areas also include location information (such as center point coordinates, boundary coordinates (such as vertex coordinates, etc.)).

[0076] In some embodiments of this specification, the thermal distribution diagram of spatter can more clearly present or distinguish the characteristics and distribution of spatter in different areas of the entire target welding torch (such as the contact tip or nozzle), providing a basis for subsequent targeted cleaning operations.

[0077] Step S330: Based on the spatter distribution characteristics, control the cleaning device to perform a cleaning operation on the target welding torch.

[0078] In some embodiments, the processor determines multiple spatter zones based on spatter distribution characteristics. These multiple spatter zones include sparse spatter zones and thick spatter zones. Based on these multiple spatter zones, the processor determines the cleaning parameters of the cleaning device using a parameter determination model. The parameter determination model is a trained machine learning model. Based on the cleaning parameters, the processor controls one or more cleaning tools to perform a cleaning operation on the target welding torch at the target cleaning time point.

[0079] For information on parameter determination models, please refer to [link / reference]. Figure 5 And its description.

[0080] A sparse splash zone refers to an area where visible and / or invisible splashes are concentrated and their size (e.g., thickness) is less than a preset size threshold. A thick splash zone refers to an area where visible and / or invisible splashes are concentrated and their size (e.g., thickness) is greater than or equal to a preset size threshold.

[0081] Cleaning parameters refer to the process parameters used when cleaning the target welding torch. These include, but are not limited to, the type of cleaning tool, cleaning duration, frequency, cycle, and force. Different cleaning tools may require different cleaning parameters for cleaning spatter with different characteristics.

[0082] In some embodiments, the cleaning tool includes a reamer, and the cleaning parameters include the reamer's feed depth and its rotational speed.

[0083] The feed depth refers to the depth to which the reamer extends axially into the contact tip and / or nozzle. The processor can control the reamer to perform a cleaning operation on the contact tip and / or nozzle of the target welding torch based on the reamer's feed depth parameters and its corresponding rotational speed parameters.

[0084] In some embodiments, the processor controls a reamer to clean up splatter in sparse splatter areas.

[0085] In some embodiments, the cleaning parameters also include axial oscillation parameters of the reamer at different feed depths. The axial oscillation parameter refers to the amplitude and frequency of the reamer's periodic oscillation along the axial direction during rotary cleaning.

[0086] In some embodiments, for thick splash zones, the cleaning tool further includes a high-pressure pulse cleaning device, and the cleaning parameters include the high-pressure pulse air pressure parameters of the high-pressure pulse cleaning device, which are positively correlated with the thickness of the splashes in the thick splash zone.

[0087] High-pressure pulse cleaning equipment includes, but is not limited to, high-pressure pulse reverse-flushing dust collectors and high-pressure pulse electric field equipment. The processor can control the high-pressure pulse cleaning equipment to assist in cleaning thick splash zones based on high-pressure pulse air pressure parameters. The greater the thickness of the splash, the higher the high-pressure pulse air pressure.

[0088] In some embodiments of this specification, welding torch cleaning is performed using a high-pressure pulse cleaning device based on the thickness of the spatter. This enhances the physical impact on thick spatter layers, reduces the load and energy consumption of mechanical cleaning (such as reamer cleaning), and improves the cleaning effect.

[0089] In some embodiments, for each of the plurality of splash zones, the processor determines the spray density of each splash zone and controls the spraying device to spray anti-splatter agent in each splash zone according to the spray density.

[0090] Antispatter agents are used to reduce the amount of spatter adhering to the target welding torch during subsequent welding operations. Spray density characterizes the amount of antispatter agent sprayed per unit area.

[0091] The coating density can differ between sparse spatter zones and thick spatter zones. Thick spatter zones indicate a relatively high probability of spatter adhesion during welding, and their corresponding coating density can be set higher. Conversely, sparse spatter zones indicate a relatively low probability of spatter adhesion during welding, and their corresponding coating density can be set lower.

[0092] In some embodiments, the processor may also employ a density allocation scheme with a smooth transition between adjacent splash areas to avoid abrupt changes in coating amount affecting the anti-adhesion effect. The smooth transition density allocation scheme includes ensuring that the difference in coating density between sparse and thick splash areas is less than a preset density threshold.

[0093] In some embodiments of this specification, by differentially controlling the spray density of sparse splash areas and thick splash areas, the adhesion characteristics of splashes can be precisely matched, reducing the cost loss caused by excessive use of anti-splatter agent while ensuring the cleaning effect.

[0094] In some embodiments, the output of the parameter determination model includes the degree of damage to the reamer. When cleaning the target welding torch, the processor determines the vibration monitoring frequency based on the degree of damage to the reamer; based on the vibration monitoring frequency, it controls the vibration sensing unit to collect the vibration information of the reamer; and based on the vibration information, it determines whether there is an abnormality in the reamer.

[0095] The degree of damage to the reamer is used to indicate the extent of damage (such as wear) to the reamer estimated before the cleaning operation during the removal of splatter. It should be noted that after the cleaning operation is completed, the processor can determine the actual degree of damage to the reamer based on changes in the reamer's physical parameters (such as the length and diameter of the blade).

[0096] In some embodiments, the processor can estimate the degree of damage to the reamer based on different spatter zones and their spatter characteristics (such as thickness, hardness, etc.). The greater the thickness and hardness of the spatter, the greater the probability of reamer damage and the greater the degree of damage. In some embodiments, the output of the parameter determination model includes the degree of damage to the reamer, and the processor can determine the degree of damage to the reamer based on the parameter determination model.

[0097] Vibration monitoring frequency refers to the time interval for collecting vibration signals from the reamer during the cleaning operation. The vibration monitoring frequency may vary depending on the degree of damage to the reamer. The vibration monitoring frequency can be a preset value determined based on experience.

[0098] In some embodiments, the processor can set a reamer damage reference table based on historical welding operation data. The reamer damage reference table includes spatter information (such as its location on the contact tip or nozzle, material, size, etc.), the degree of damage to the reamer (reference value), and the vibration monitoring frequency (reference value). During cleaning operations, the processor determines the degree of damage to the reamer and its corresponding vibration monitoring frequency from the reamer damage reference table based on the actual spatter characteristics, thereby controlling the vibration sensing unit to collect the reamer's vibration information. The degree of damage is positively correlated with the vibration monitoring frequency; for example, the greater the damage to the reamer, the higher the vibration monitoring frequency.

[0099] Vibration information reflects the intensity or amplitude of the reamer's vibration during cleaning, and can be used to assess any abnormalities in the reamer. In some embodiments, the processor compares (e.g., performs difference analysis) the current vibration information of the reamer, which is collected in real time by the vibration sensing unit during the cleaning operation, with standard vibration information (such as historical vibration information under normal conditions) to determine if there is an abnormality in the reamer. In response to the presence of an abnormality in the reamer, cleaning is stopped and a warning message is issued.

[0100] In some embodiments of this specification, the safety of the cleaning operation can be improved by detecting abnormal conditions of the reamer.

[0101] In some embodiments of this specification, different cleaning tools and their cleaning parameters are used to take into account the characteristics of different spatter, making the cleaning operation of the target welding gun more targeted, improving cleaning efficiency, extending the service life of vulnerable cleaning tools (such as reamers), and reducing costs.

[0102] Figure 4 This is a schematic diagram of a model determined according to some embodiments shown in this specification.

[0103] In some embodiments, the processor determines at least one feature sequence based on the equipment parameters, operating parameters of the target welding torch, and welding characteristics of the target workpiece using a sequence determination model.

[0104] A sequence determination model is a model used to determine at least one feature sequence of a target welding torch. It can be a model generated by modeling based on various analytical algorithms (such as mathematical algorithms) (such as mathematical models). In some embodiments, the sequence determination model is a trained machine learning model, such as any one or a combination of a Long Short Term Memory (LSTM) neural network model or other custom model structures.

[0105] like Figure 4 As shown, the inputs of the sequence determination model 420 include equipment parameters 411, working parameters 412, and welding features 413, and the output includes welding quality sequence 430.

[0106] For more information on equipment parameters, operating parameters, welding characteristics, and quality characteristic sequences, please refer to [link / reference]. Figure 2 And its description.

[0107] In some embodiments, the input to the sequence determination model 420 also includes spatter distribution characteristics at the previous cleaning point (not shown in the figure). The spatter distribution characteristics at the previous cleaning point refer to the spatter distribution characteristics of the target welding torch during the previous cleaning. When the processor makes the current prediction to obtain the weld quality sequence 430, it obtains the spatter distribution characteristics at the previous cleaning point from the historical data of the previous cleaning operation on the target welding torch.

[0108] In some embodiments, the input to the sequence determination model 420 further includes the antispatter agent spray density (not shown) in different spatter zones at the previous cleaning point. The antispatter agent spray density in different spatter zones at the previous cleaning point refers to the antispatter agent spray density in different spatter zones of the target welding torch during the previous cleaning. The processor can obtain the antispatter agent spray density in different spatter zones at the previous cleaning point from historical data of the previous cleaning operation on the target welding torch.

[0109] In some embodiments, the inputs to the sequence determination model 420 also include environmental parameters (not shown in the figure). Environmental parameters include temperature, humidity, etc., during the welding operation.

[0110] For more information on splatter zones, splatter distribution characteristics, and spray density, please refer to [link to relevant information]. Figure 3 And its description.

[0111] The sequence determination model 420 can be obtained by training multiple sets of first training samples with a first label. The first training samples are obtained based on historical welding data from multiple welding operations within a historical period (such as the past six months, one month, etc.). Here, historical welding data refers to historical data of welding steps (or welding processes) in historical welding operations.

[0112] Historical welding data includes sample welding data from multiple sample welding torches. This sample welding data includes sample equipment parameters, sample operating parameters, and sample welding characteristics corresponding to the sample workpiece at multiple sample time points. Historical welding data also includes the sample welding quality for each sample time point. The sample welding quality represents the actual welding quality evaluated based on the sample welding data during historical welding operations.

[0113] The processor can use sample welding data from a specific sample time point (hereinafter referred to as the first sample time point) as a set of first training samples. As an example, for multiple sample time points in a historical welding operation, the processor can obtain the sample equipment parameters, sample operating parameters, and sample welding characteristics at one of the first sample time points to generate a set of first training samples. For multiple different historical welding operations, the processor can generate multiple sets of first training samples.

[0114] In some embodiments, the processor constructs a sample welding quality sequence based on the sample welding quality at multiple sample time points following the first sample time point (hereinafter referred to as the second sample time point), and uses the sample welding quality sequence as a first label corresponding to a set of first training samples. The first label can be labeled based on manual annotation or other feasible methods.

[0115] During training, the processor determines the value of the loss function based on the difference between the output of the initial sequence determination model and the first training label. The parameters of the initial sequence determination model can be iteratively updated based on the value of the loss function until the training termination condition is met (e.g., the loss function converges, a certain number of iterations are performed, etc.). The updated initial sequence determination model can then be used as the trained sequence determination model.

[0116] In some embodiments, historical welding data also includes sample environment parameters, and the first training sample used to train the sequence determination model also includes sample environment parameters.

[0117] The processor can obtain the sample environmental parameters (such as sample temperature and sample humidity) at the first sample time point from historical welding operation data, and construct the first training sample based on the sample equipment parameters, sample operating parameters, sample welding characteristics, and sample environmental parameters at the first sample time point, for use in training the sequence determination model. For details on training the sequence determination model, please refer to the previous description of sequence determination model training; it will not be repeated here.

[0118] In some embodiments, the first training samples used to train the sequence determination model also include the sample splash distribution characteristics corresponding to the previous historical cleanup point.

[0119] It should be noted that welding operations include the welding process and the torch cleaning process (or torch cleaning procedure). In a single welding operation, the welding and torch cleaning processes can be performed alternately. The previous historical cleaning time point can be the sample time point corresponding to the most recent cleaning operation before the first sample time point.

[0120] The processor can extract the distribution characteristics of splatter from the previous historical cleaning point in the sample cleaning data to construct the first training sample for training the sequence determination model. For details on training the sequence determination model, please refer to the previous description. For information on sample cleaning data, please refer to [link to relevant documentation]. Figure 5 And its description.

[0121] In some embodiments, the first training samples used to train the sequence determination model further include the sample spraying density of different sample splash areas corresponding to the previous historical cleaning time point. The processor can obtain the sample spraying density of different sample splash areas corresponding to the previous historical cleaning time point from the sample cleaning data to construct the first training samples for training the sequence determination model. For the training of the sequence determination model, please refer to the previous description of the sequence determination model training, which will not be repeated here.

[0122] Some embodiments in this specification, through sequence determination models, can learn the influence of welding data (such as welding torch equipment parameters, working parameters, welding characteristics of the workpiece, etc.) on welding quality during welding operations, thereby enabling a more accurate understanding of the trend of welding quality changes and providing a good foundation for subsequent determination of target cleaning points.

[0123] Figure 5 This is a schematic diagram of the model determined based on the parameters shown in some embodiments of this specification.

[0124] In some embodiments, the processor determines the cleaning parameters of the gun cleaning device based on the characteristics of the splatter distribution by using a parameter determination model.

[0125] A parameter determination model is a model used to determine the cleaning parameters when cleaning a target welding torch. It can be a model generated through modeling based on various analysis algorithms (such as mathematical algorithms). In some embodiments, the parameter determination model is a trained machine learning model, such as any one or a combination of Deep Neural Networks (DNN) models or other custom model structures.

[0126] like Figure 5 As shown, the input of the parameter determination model 520 includes the splash distribution characteristics 511, and the output includes the cleaning parameters 530.

[0127] The splash distribution characteristics 511 are determined based on the splash distribution characteristics of the splash zone 501.

[0128] The spatter zone 501 includes both visible and invisible areas of the target welding torch, such as the outer surface of the contact tip and / or nozzle of the target welding torch and the inner wall of the contact tip and / or nozzle of the target welding torch. The spatter zone 501 also includes spatter areas within the visible and / or invisible areas, such as sparse spatter areas and thick spatter areas.

[0129] like Figure 5 As shown, the splash area 501 includes splash area A1, splash area A2, ..., splash area A n .

[0130] In some embodiments, the input to the parameter determination model 520 is a spatter thermal map of the target welding torch (not shown in the figure), wherein the spatter thermal map can be determined based on spatter distribution characteristics 511.

[0131] Cleanup parameter 530 includes each splash zone (e.g., splash zone A) n The corresponding cleaning tools (not shown in the figure) and their cleaning parameters, such as... Figure 5 As shown, cleaning parameters 530 include cleaning parameters for splash zone A1, cleaning parameters for splash zone A2, ..., cleaning parameters for splash zone An Cleaning parameters. The cleaning tool can be represented by a preset cleaning tool type, which can be a numerical value (e.g., 1, 2, 3, etc.) or other forms. For example, 1 represents a reamer, 2 represents a high-pressure pulse cleaning device, and 3 represents a spraying device, etc. Different splash zones (e.g., splash zone A) n (etc.) can be represented based on location information (such as the coordinates of the center point of the region and / or the coordinates of the boundary points of the region) and color information (such as RGB values).

[0132] In some embodiments, cleaning parameters (e.g., splash zone A) n Cleaning parameters include the reamer feed depth and rotational speed. In some embodiments, cleaning parameters (e.g., splash zone A) are... n Cleaning parameters include reamer feed depth, rotational speed, and axial oscillation parameters.

[0133] In some embodiments, the output of the parameter determination model 520 may also include the degree of damage to the reamer 540.

[0134] In some embodiments, the output of the parameter determination model 520 may also include high-pressure pulse air pressure parameters of the high-pressure pulse cleaning device. For example, for a thick splash zone, the output of the parameter determination model 520 may include reamer cleaning parameters (such as feed depth and rotational speed, axial oscillation parameters) and high-pressure pulse air pressure parameters of the high-pressure pulse cleaning device.

[0135] In some embodiments, the output of the parameter determination model 520 may also include the spraying density of the spraying device corresponding to different splash areas, which is used to instruct the spraying device to spray anti-splatter agent onto different splash areas at the corresponding spraying density after the cleaning operation is completed.

[0136] More information about sparse splatter zones, thick splatter zones, splatter distribution characteristics, cleaning tools, cleaning parameters, and the degree of damage to the reamer can be found elsewhere in this manual (e.g., Figure 3 ).

[0137] The parameter determination model 520 can be obtained by training multiple sets of second training samples with second labels. These second training samples can be obtained based on historical welding data and historical torch cleaning data from multiple welding operations within a historical time period (such as the past six months, one month, etc.). Historical torch cleaning data refers to historical data on the torch cleaning step (or torch cleaning process) in historical welding operations. Specifically, historical welding data includes sample welding data from multiple sample welding torches, and historical torch cleaning data includes sample torch cleaning data from multiple sample welding torches. For more information on sample welding data, see [link to relevant documentation]. Figure 4 And its description.

[0138] In some embodiments, the sample welding data includes sample spatter data generated by the sample welding torch during the welding process. The processor can generate sample spatter features based on a set of sample spatter data as a set of second training samples. In some embodiments, the second training samples are in the form of sample spatter heatmaps, and the processor generates sample spatter heatmaps based on these sample spatter features as a set of second training samples.

[0139] The sample cleaning data includes sample cleaning parameters. These parameters include actual cleaning parameters used in historical cleaning processes (such as sample cleaning time point, sample feed depth, sample rotation speed, and sample axial oscillation parameters), which can be used to generate second labels for second training samples. The sample cleaning parameters can be those used in cleaning operations that meet the requirements for cleaning effectiveness (such as cleaning time being lower than a preset time threshold, and no wear on the conductive tip and nozzle of the sample welding torch).

[0140] The second label is determined based on the sample cleaning parameters corresponding to the second training sample. As an example only, for a set of second training samples (such as a sample splatter heatmap), the second label includes the reamer cleaning parameters corresponding to one or more sparse areas in the sample splatter heatmap, the reamer cleaning parameters corresponding to one or more thick areas in the sample splatter heatmap, the high-pressure pulse air pressure parameters of the high-pressure pulse cleaning equipment, and the spraying density of the spraying device in each splatter area. The second label can be based on manual annotation or other feasible methods.

[0141] During training, the difference between the model's output and the second label is determined based on the initial parameters, and the value of the loss function is determined accordingly. The parameters of the initial parameter determination model can be iteratively updated based on the value of the loss function until the training termination condition is met (e.g., the loss function converges, a certain number of iterations have been performed, etc.). The updated initial parameter determination model can then be used as the trained parameter determination model.

[0142] Some embodiments in this specification, through parameter determination models, can learn the relationship between the spatter distribution characteristics of the welding torch and the cleaning parameters. This enables the automatic determination of the cleaning parameters of the target welding torch during actual torch cleaning, reducing labor costs while improving the accuracy of the cleaning parameters.

[0143] It should be noted that the above description of the process is for illustrative purposes only and does not limit the scope of this specification. Those skilled in the art can make various modifications and changes to the process under the guidance of this specification. However, these modifications and changes remain within the scope of this specification.

[0144] The basic concepts have been described above. Obviously, for those skilled in the art, the detailed disclosure above is merely illustrative and does not constitute a limitation of this specification. Although not explicitly stated herein, those skilled in the art may make various modifications, improvements, and corrections to this specification. Such modifications, improvements, and corrections are suggested in this specification and therefore remain within the spirit and scope of the exemplary embodiments described herein.

[0145] Furthermore, this specification uses specific terms to describe embodiments thereof. For example, "an embodiment," "one embodiment," and / or "some embodiments" refer to a particular feature, structure, or characteristic associated with at least one embodiment of this specification. Therefore, it should be emphasized and noted that references to "an embodiment," "one embodiment," or "an alternative embodiment" in different locations throughout this specification do not necessarily refer to the same embodiment. Moreover, certain features, structures, or characteristics in one or more embodiments of this specification can be appropriately combined.

[0146] Finally, it should be understood that the embodiments described in this specification are merely illustrative of the principles of the embodiments described herein. Other variations may also fall within the scope of this specification. Therefore, alternative configurations of the embodiments described herein are intended to be consistent with the teachings of this specification, rather than as examples or limitations. Accordingly, the embodiments described herein are not limited to those explicitly introduced and described herein.

Claims

1. A welding robot torch cleaning control system, characterized in that, The invention includes a cleaning device and a processor, the cleaning device comprising one or more cleaning tools, and the cleaning device being configured to perform a cleaning operation on at least one welding torch. The processor is configured to: Based on equipment parameters, operating parameters, and welding characteristics of the target workpiece, at least one feature sequence of the target welding torch is determined, wherein the at least one feature sequence includes a welding quality sequence, which is generated based on the welding quality at multiple time points. In response to the at least one feature sequence satisfying a first preset condition, Based on the at least one feature sequence, the target cleaning time point of the target welding torch is determined; Acquire spatter data of the target welding torch, the spatter data including first spatter data of the visible area of ​​the target welding torch and second spatter data of the inner wall of the target welding torch; Based on the spatter data, the spatter distribution characteristics of the target welding torch are determined; Based on the distribution characteristics of the splashes, multiple splash zones are determined, including sparse splash zones and thick splash zones; Based on the multiple splash zones, the cleaning parameters of the gun cleaning device are determined by a parameter determination model, wherein the parameter determination model is a trained machine learning model; Based on the cleaning parameters, the one or more cleaning tools are controlled to perform a cleaning operation on the target welding torch at the target cleaning time point.

2. The system according to claim 1, characterized in that, The processor is further configured to: The sequence determination model is a trained machine learning model that determines at least one feature sequence based on the equipment parameters, the working parameters, and the welding features.

3. A method for controlling the cleaning of a welding torch in a welding robot, characterized in that, The method is executed by a processor of a welding robot torch cleaning control system, the welding robot torch cleaning control system including a torch cleaning device, the torch cleaning device including one or more cleaning tools, and the method includes; Based on equipment parameters, operating parameters, and welding characteristics of the target workpiece, at least one feature sequence of the target welding torch is determined, wherein the at least one feature sequence includes a welding quality sequence, which is generated based on the welding quality at multiple time points. In response to the at least one feature sequence satisfying a first preset condition, Based on the at least one feature sequence, the target cleaning time point of the target welding torch is determined; Acquire spatter data of the target welding torch, the spatter data including first spatter data of the visible area of ​​the target welding torch and second spatter data of the inner wall of the target welding torch; Based on the spatter data, the spatter distribution characteristics of the target welding torch are determined; Based on the distribution characteristics of the splashes, multiple splash zones are determined, including sparse splash zones and thick splash zones; Based on the multiple splash zones, the cleaning parameters of the gun cleaning device are determined by a parameter determination model, wherein the parameter determination model is a trained machine learning model; Based on the cleaning parameters, the one or more cleaning tools are controlled to perform a cleaning operation on the target welding torch at the target cleaning time point.

4. The method according to claim 3, characterized in that, The determination of at least one feature sequence of the target welding torch based on equipment parameters, operating parameters, and welding characteristics of the target workpiece includes: The sequence determination model is a trained machine learning model that determines at least one feature sequence based on the equipment parameters, the working parameters, and the welding features.

5. A computer-readable storage medium storing computer instructions, wherein when a computer reads the computer instructions in the storage medium, the computer executes the welding robot torch cleaning control method as described in any one of claims 3 to 4.

Images