Industrial robot control system and robot control method

Abstract

The present invention discloses an industrial robot control system and a robot control method, which relate to the technical field of industrial robot control, and include the following steps: when the industrial welding robot controls the welding head, the control information of the welding process is obtained in real time through sensors and controllers, and the obtained information is transmitted to the central control system; the obtained information is sorted, and the features of the sorted information are extracted. The present invention obtains welding control information in real time, analyzes characteristic parameters in combination with a machine learning model, and timely identifies potential anomalies in the welding process. If anomalies are detected, the welding traces are obtained and compared through an image sensor, a similarity difference index is generated, and compared with a reference threshold, and the influence of the anomaly on the welding quality is judged. If the similarity difference index exceeds the threshold, it is marked as low-quality welding, and the operation is stopped and maintained to ensure high-quality welding and production stability.

Description

Technical Field

[0001] The present invention relates to the technical field of industrial robot control, and in particular to an industrial robot control system and a robot control method. Background Art

[0002] Industrial robot control refers to the precise control and management of the movement, operation and behavior of industrial robots through various technical means and methods. Industrial robots are usually composed of multiple joints, each of which is driven by a motor or hydraulic system to achieve complex three-dimensional spatial movement. The control system mainly includes motion control, path planning, posture control, speed and acceleration control. Through the coordinated work of sensors, controllers and actuators, industrial robots are able to perform precise tasks such as welding, spraying, handling, assembly, etc. The control system also needs to handle environmental changes, obstacle avoidance and collaborative work with other equipment to ensure the efficient and safe operation of the robot.

[0003] The core of industrial robot control systems is control algorithms and software, which are usually based on advanced control theory and computer technology. These control algorithms can be divided into two categories: open-loop control and closed-loop control. Open-loop control relies on predetermined instructions and models, while closed-loop control uses sensor feedback to adjust the robot's behavior in real time, thereby achieving higher precision and adaptability. In addition, with the advancement of artificial intelligence and machine learning technology, modern industrial robot control systems have also introduced intelligent control methods, enabling robots to learn and adapt to new environments and tasks. Through this advanced control technology, industrial robots have not only improved production efficiency and product quality, but also significantly reduced labor costs and labor intensity, promoting the development of industrial automation and intelligent manufacturing.

[0004] When controlling the welding of industrial welding robots, the welding head needs to be precisely controlled. The control of the welding head refers to the precise control of the position, posture, motion trajectory, and welding parameters (such as current, voltage, welding speed, etc.) of the welding head when the industrial robot performs welding operations. Its function is to ensure that the welding head can accurately reach the specified position during the welding process and maintain the correct welding angle and distance, thereby achieving high-quality and stable welding results. This control can not only improve the accuracy and consistency of welding and reduce welding defects, but also adapt to different welding materials and complex welding environments, significantly improving production efficiency and product quality.

[0005] The prior art has the following deficiencies:

[0006] When the prior art controls the welding head of an industrial welding robot, abnormalities in the welding head control are usually not discovered in time. Since batch welding and batch inspection are usually adopted in industrial production, this means that during the welding process, there may be multiple products that have completed the welding operation, but the welding quality has not yet been tested. If the abnormality in the welding head control is not detected and corrected in time, the welding quality of these products is likely to not meet the standards, resulting in the entire batch of products being unqualified. Such a situation will not only cause waste of materials and time, but may also bring serious economic losses and production delays.

[0007] The above information disclosed in this Background section is only for enhancement of understanding of the background of the present disclosure and therefore it may contain information that does not constitute the prior art that is already known to one of ordinary skill in the art. Summary of the invention

[0008] The purpose of the present invention is to provide an industrial robot control system and a robot control method, which can obtain welding control information in real time, analyze characteristic parameters in combination with a machine learning model, and promptly identify potential anomalies in the welding process. If an anomaly is detected, the welding traces are acquired and compared through an image sensor, and a similarity difference index is generated and compared with a reference threshold to determine the impact of the anomaly on the welding quality. If the similarity difference index exceeds the threshold, it is marked as low-quality welding, and the operation is stopped and maintained to ensure high-quality welding and production stability, so as to solve the problems in the above-mentioned background technology.

[0009] In order to achieve the above object, the present invention provides the following technical solution: an industrial robot control method, comprising the following steps:

[0010] When the industrial welding robot controls the welding head, it obtains the control information of the welding process in real time through sensors and controllers, and transmits the acquired information to the central control system;

[0011] The acquired information is sorted, and features are extracted from the sorted information. The extracted features are then analyzed and processed for abnormalities. The extracted features are then input into a pre-trained machine learning model for analysis to determine whether there are potential abnormalities in the current welding process.

[0012] If the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor;

[0013] By comparing the difference between the last welding mark and the current welding mark through image processing technology, the quality of the current welding mark is evaluated;

[0014] Based on the results of the comparison analysis, determine whether the potential abnormalities in the welding process have an impact on the welding quality. If the comparison results show that there are obvious differences in the welding marks and affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process. The welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues.

[0015] Preferably, feature extraction is performed on the sorted information, wherein the extracted features include heat input, wire feeding speed and welding speed during welding. After extraction, the heat input during welding is subjected to abnormal analysis to generate a heat input unevenness index, and the wire feeding speed and welding speed are subjected to abnormal analysis to generate a wire feeding speed consistency index.

[0016] Preferably, the heat input unevenness index and the wire feeding speed consistency index after abnormal analysis are input into a pre-trained machine learning model, and a welding coefficient is generated by the model to determine whether there is a potential abnormality in the current welding process.

[0017] Preferably, the specific steps of analyzing and processing the heat input during welding to generate the heat input unevenness index are as follows:

[0018] In a welding process, obtain the welding current change function over time and the welding voltage as a function of time , and calculate the instantaneous heat input power , represents the heat input at time t, and the calculation expression is: ;

[0019] Calculate the cumulative heat input energy during a welding process , represents the cumulative heat input energy at time t, and the calculation expression is: , where and They represent the start time and end time of a welding process respectively. arrive The time period between is the total duration of a welding process;

[0020] Calculate heat input density , represents the heat input density at time t, and the calculation expression is: , where Represents the cumulative welding length at time t, and the calculation expression is: ,in, Represents the function of welding speed changing with time;

[0021] Calculate the local heat input non-uniformity, the calculation expression is: , where represents the rate of change of heat input density, , Indicated in The local heat input non-uniformity at the moment, Indicated in The rate of change of heat input density at time instant, n represents the total number of time points;

[0022] Calculate the heat input unevenness index, the calculation expression is: , where Indicates the heat input unevenness index.

[0023] Preferably, the specific steps of performing abnormal analysis on the wire feeding speed and the welding speed to generate the wire feeding speed consistency index are as follows:

[0024] During a welding process, the wire feeding speed data is collected in real time, and the wire feeding speed is calculated as a function of the time. To express;

[0025] Calculate the difference between wire feeding speed and welding speed at each moment , the calculation expression is: ;

[0026] The absolute value of the speed difference is integrated to calculate the total difference D in one welding process. The calculation expression is: , where T is the total length of one welding time;

[0027] Calculate the wire feeding speed consistency index, the calculation expression is: , where represents the total difference after normalization, , , , , Indicates the wire feeding speed consistency index.

[0028] Preferably, the welding coefficient generated by the industrial welding robot during one welding operation is compared and analyzed with a preset welding coefficient reference threshold value to determine whether there is a potential abnormality in the current welding process. The determination process is as follows:

[0029] If the welding coefficient is greater than or equal to the welding coefficient reference threshold, the welding process of the industrial welding robot is marked as a potential abnormality;

[0030] If the welding coefficient is less than the welding coefficient reference threshold, the process of the industrial welding robot welding is marked as efficient welding.

[0031] Preferably, when there is a potential abnormality in the welding process, a similarity analysis is performed on the welding traces of the previous welding and the welding traces of the current welding to generate a similarity difference index, and the quality of the current welding trace is evaluated by the similarity difference index.

[0032] Preferably, the steps of generating the similarity difference index are as follows:

[0033] Before and after the welding process begins, the image sensor is used to obtain the welding mark image of the previous welding. And the welding mark image of the current welding ;

[0034] Preprocess the acquired images, including grayscale, denoising and edge detection;

[0035] Use shape descriptors to extract weld shape features from preprocessed images;

[0036] The cosine similarity between the previous welding mark image and the current welding mark image is calculated using cosine similarity. The calculation expression is: ,in, and are the feature vector components of the previous welding mark image and the current welding mark image, respectively, N is the dimension of the feature vector, Represents the cosine similarity between the previous welding mark image and the current welding mark image;

[0037] According to the cosine similarity calculation results, the similarity difference index is generated, and the generated expression is: , where Represents the similarity difference index, Represents a weight factor used to adjust the sensitivity of the similarity difference index.

[0038] An industrial robot control system includes a data acquisition module, a data processing and anomaly detection module, an image acquisition module, an image analysis and quality assessment module, and a quality judgment and maintenance module;

[0039] The data acquisition module acquires the control information of the welding process in real time through sensors and controllers when the industrial welding robot controls the welding head, and transmits the acquired information to the central control system;

[0040] The data processing and anomaly detection module organizes the acquired information, extracts features from the organized information, performs anomaly analysis on the extracted features, and then inputs the extracted features into a pre-trained machine learning model for analysis to determine whether there are potential anomalies in the current welding process;

[0041] Image acquisition module, if the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor;

[0042] Image analysis and quality assessment module, which uses image processing technology to compare the difference between the previous welding mark and the current welding mark, and assess the quality of the current welding mark;

[0043] The quality judgment and maintenance module determines whether the potential abnormalities in the welding process have an impact on the welding quality based on the results of the comparison analysis. If the comparison results show that there are obvious differences in the welding marks and they affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process, the welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues.

[0044] In the above technical solution, the technical effects and advantages provided by the present invention are:

[0045] The present invention can timely identify potential anomalies in the welding process by acquiring control information of the welding process in real time when the industrial welding robot controls the welding head, and analyzing the characteristic parameters in combination with a pre-trained machine learning model. If potential anomalies are found, the welding traces are further acquired through an image sensor for detailed comparison and evaluation. This process ensures that potential abnormalities can be discovered and marked in a timely manner during the welding process, preventing undetected anomalies from affecting the welding quality in subsequent batch production, thereby improving the reliability and consistency of the overall welding process.

[0046] When the present invention identifies the existence of potential abnormalities in the welding process, the last welding mark and the current welding mark are obtained by the image sensor, and the image processing technology is used to perform a detailed similarity analysis to generate a similarity difference index. The similarity difference index is compared with a preset reference threshold value, so as to accurately judge whether the potential abnormalities in the welding process have an impact on the welding quality. If the similarity difference index exceeds the threshold value, the current welding mark is marked as low-quality welding, the welding process is stopped in time and maintenance management is carried out to prevent unqualified products from entering subsequent production links. If the similarity difference index is within the threshold range, it indicates that the welding quality is not affected by the abnormality and the welding operation can continue, thereby providing an accurate quality assessment means by comparing the welding marks, ensuring the high quality of the welding products and the stability of the production process. BRIEF DESCRIPTION OF THE DRAWINGS

[0047] In order to more clearly illustrate the embodiments of the present application or the technical solutions in the prior art, the drawings required for use in the embodiments will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can also be obtained based on these drawings.

[0048] Figure 1 The present invention is a method flow chart of an industrial robot control method.

[0049] Figure 2 The figure is a module schematic diagram of an industrial robot control system of the present invention. DETAILED DESCRIPTION

[0050] Example embodiments will now be described more fully with reference to the accompanying drawings. However, example embodiments can be implemented in a variety of forms and should not be construed as limited to the examples set forth herein; rather, these example embodiments are provided so that the description of the present disclosure will be more comprehensive and complete, and the concept of the example embodiments will be fully conveyed to those skilled in the art.

[0051] The present invention provides Figure 1 An industrial robot control method shown includes the following steps:

[0052] When the industrial welding robot controls the welding head, it obtains the control information of the welding process in real time through sensors and controllers, and transmits the acquired information to the central control system;

[0053] When an industrial welding robot performs welding tasks, various sensors and controllers installed on the welding robot collect various data and parameters in real time during the welding process, and transmit this information to the central control system for further processing and analysis. Specifically, the sensor can detect key parameters in the welding process, such as current, voltage, temperature, welding speed, position and angle of the welding head, etc. The controller is responsible for receiving these sensor data and dynamically adjusting and controlling the welding process according to the preset welding program and algorithm to ensure welding quality and efficiency.

[0054] The acquired information is sorted, and features are extracted from the sorted information. The extracted features are then analyzed and processed for abnormalities. The extracted features are then input into a pre-trained machine learning model for analysis to determine whether there are potential abnormalities in the current welding process.

[0055] Data cleaning includes steps such as noise removal, data cleaning and formatting to ensure data quality and consistency. This step is crucial for subsequent feature extraction and analysis.

[0056] Feature extraction is performed on the sorted information, wherein the extracted features include heat input, wire feeding speed and welding speed during welding. After extraction, the heat input during welding is analyzed abnormally to generate a heat input unevenness index, and the wire feeding speed and welding speed are analyzed abnormally to generate a wire feeding speed consistency index.

[0057] The heat input unevenness index and wire feeding speed consistency index after abnormal analysis are input into the pre-trained machine learning model, and the welding coefficient is generated by the model. The welding coefficient is used to determine whether there is potential abnormality in the current welding process.

[0058] When industrial welding robots are welding, uneven heat input during the welding process will have a serious impact on the welding quality. This is because the heat input during the welding process directly determines the weld depth, weld width and weld formation. If the heat input is uneven, the temperature distribution of each part of the weld will be inconsistent, which will affect the melting and solidification process of the weld metal. Specifically, areas with excessive heat input may cause overburning, metal overheating, and coarse grains, which will reduce the mechanical properties of the weld, and cause problems such as excessive heat-affected zone and increased weld brittleness. Areas with insufficient heat input may cause welding defects such as insufficient penetration, lack of fusion, inclusions and pores, which will seriously weaken the strength and durability of the weld and increase the risk of fatigue cracks and fractures in the weld. In addition, uneven heat input will also cause welding stress and deformation, resulting in substandard shape and dimensional accuracy of the weldment, requiring additional subsequent processing and repair, and increasing production costs and time. Therefore, maintaining the uniformity of heat input during welding is crucial to ensuring welding quality and improving production efficiency.

[0059] The specific steps for analyzing the heat input during welding to generate the heat input unevenness index are as follows:

[0060] In a welding process, obtain the welding current change function over time and the welding voltage as a function of time , and calculate the instantaneous heat input power , represents the heat input at time t, and the calculation expression is: ;

[0061] Calculate the cumulative heat input energy during a welding process , represents the cumulative heat input energy at time t, and the calculation expression is: , where and They represent the start time and end time of a welding process respectively. arrive The time period between is the total duration of a welding process;

[0062] Calculate heat input density , represents the heat input density at time t, and the calculation expression is: , where Represents the cumulative welding length at time t, and the calculation expression is: ,in, Represents the function of welding speed changing with time;

[0063] Calculate the local heat input non-uniformity, the calculation expression is: , where represents the rate of change of heat input density, , Indicated in The local heat input non-uniformity at the moment, Indicated in The rate of change of heat input density at time instant, n represents the total number of time points;

[0064] Calculate the heat input unevenness index, the calculation expression is: , where Indicates the heat input unevenness index.

[0065] It can be seen from the heat input unevenness index that the larger the performance value of the heat input unevenness index generated by the industrial welding robot during a welding operation, the higher the unevenness of the heat input during the welding process, and the greater the impact on the welding quality. Specifically, a higher heat input unevenness index means that there are significant heat input fluctuations during the welding process, which will lead to uneven welds, insufficient penetration, cracks, pores and other welding defects, which in turn affect the strength and durability of the weld. On the contrary, a lower heat input unevenness index indicates that the heat input is more uniform, the welding quality is more stable, the weld is well formed, and there are fewer defects.

[0066] When an industrial welding robot performs welding, the mismatch between the wire feeding speed and the welding speed will have a serious impact on the welding quality. The wire feeding speed refers to the rate at which the wire is fed into the molten pool during welding, while the welding speed refers to the speed at which the welding head moves on the surface of the workpiece. The matching of these two speeds is crucial to the quality of the weld. If the wire feeding speed is too fast and the welding speed is relatively slow, the wire will accumulate too much in the molten pool, resulting in an excessively large molten pool and a thick weld. Spatter and irregular accumulation may appear on the weld surface, affecting the appearance and mechanical properties of the weld. In addition, too much molten metal may cause the weld heat affected zone to become larger, resulting in greater thermal deformation and residual stress, and increasing the risk of welding cracks. Conversely, if the wire feeding speed is too slow and the welding speed is relatively fast, the welding arc will not be able to maintain a stable molten pool, resulting in insufficient penetration and poor weld formation. The weld may have defects such as depression, lack of fusion or incomplete filling, which will seriously affect the strength and durability of the weld joint. In some key structural parts, such welding defects may lead to catastrophic failure. In addition, if the wire feeding speed does not match the welding speed during welding, it will also lead to instability of the welding arc and irregular welding current and voltage fluctuations, which will further affect the quality and consistency of the weld.

[0067] The specific steps for abnormal analysis and processing of wire feeding speed and welding speed to generate wire feeding speed consistency index are as follows:

[0068] During a welding process, the wire feeding speed data is collected in real time, and the wire feeding speed is calculated as a function of the time. To express;

[0069] Calculate the difference between wire feeding speed and welding speed at each moment , the calculation expression is: ;

[0070] The absolute value of the speed difference is integrated to calculate the total difference D in one welding process. The calculation expression is: , where T is the total length of one welding time;

[0071] Calculate the wire feeding speed consistency index, the calculation expression is: , where represents the total difference after normalization, , , , , Indicates the wire feeding speed consistency index.

[0072] It can be seen from the wire feeding speed consistency index that the larger the wire feeding speed consistency index performance value generated by the industrial welding robot during a welding operation, the higher the matching degree between the wire feeding speed and the welding speed, the more stable the welding process, the less impact on the welding quality, and the better the consistency and integrity of the weld. On the contrary, if the consistency index performance value is small, it means that the wire feeding speed and the welding speed are poorly matched, and problems such as unstable molten pool and poor weld formation may occur during the welding process, which will have a greater negative impact on the welding quality and increase the risk of welding defects. Therefore, the wire feeding speed consistency index is an important parameter for measuring welding quality. The larger the performance value, the better the welding quality.

[0073] The machine learning model is not specifically limited here, and can achieve the heat input uneven index and wire feed speed consistency index Perform comprehensive analysis to generate welding coefficients In order to realize the technical solution of the present invention, the present invention provides a specific implementation method;

[0074] Welding coefficient The resulting calculation formula is: , where , Heat input unevenness index and wire feed speed consistency index The preset scaling factor of , Both are greater than 0.

[0075] It can be seen from the welding coefficient that the greater the performance value of the heat input unevenness index generated by the industrial welding robot during a welding operation, the smaller the performance value of the generated wire feeding speed consistency index. That is, the greater the performance value of the welding coefficient generated during a welding operation, the greater the potential abnormality risk in the welding process. Conversely, the smaller the potential abnormality risk in the welding process.

[0076] The welding coefficient generated by the industrial welding robot during a welding operation is compared and analyzed with the preset welding coefficient reference threshold to determine whether there is a potential abnormality in the current welding process. The judgment process is as follows:

[0077] If the welding coefficient is greater than or equal to the welding coefficient reference threshold, the welding process of the industrial welding robot is marked as a potential abnormality, indicating that there may be potential abnormalities when the industrial welding robot is welding;

[0078] If the welding coefficient is less than the welding coefficient reference threshold, the welding process of the industrial welding robot is marked as efficient welding, indicating that the industrial welding robot can achieve efficient welding.

[0079] If the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor;

[0080] First, install a high-resolution image sensor on the industrial welding robot. The image sensor should be installed in a position where it can clearly capture the weld marks, usually near the welding head. After installation, the sensor is calibrated to ensure the clarity and accuracy of the image. The calibration process involves adjusting the sensor's parameters such as focal length, exposure time, and resolution to adapt to the lighting conditions of the welding environment and the reflective characteristics of the welding material. After calibration, it can be ensured that the image sensor can accurately capture the details of the weld marks.

[0081] Image acquisition is performed before the welding process begins and after the welding is completed. Image acquisition before welding is to obtain the initial state of the workpiece, and image acquisition after welding is to record the final state of the welding mark. Image acquisition should be performed under the same ambient lighting conditions to ensure image consistency. During the acquisition process, the image sensor captures high-resolution images of the welding area, which will be used for subsequent analysis and comparison.

[0082] By comparing the difference between the last welding mark and the current welding mark through image processing technology, the quality of the current welding mark is evaluated;

[0083] When there is a potential abnormality in the welding process, a similarity analysis is performed on the welding traces of the previous welding and the welding traces of the current welding to generate a similarity difference index, and the quality of the current welding trace is evaluated by the similarity difference index.

[0084] The steps for generating the similarity difference index are as follows:

[0085] Before and after the welding process begins, the image sensor is used to obtain the welding mark image of the previous welding. And the welding mark image of the current welding ;

[0086] Capturing weld marks requires ensuring that the captured images have high resolution and consistent lighting conditions for accurate analysis.

[0087] Preprocess the acquired images, including grayscale, denoising and edge detection;

[0088] Grayscale can simplify image data, denoising can remove noise from the image, and edge detection can highlight the outline of welding marks.

[0089] Use shape descriptors (such as Hu moments and Zernike moments) to extract weld shape features from preprocessed images;

[0090] These parameters may include weld width, weld profile, surface quality, etc.

[0091] The cosine similarity between the previous welding mark image and the current welding mark image is calculated using cosine similarity. The calculation expression is: ,in, and are the feature vector components of the previous welding mark image and the current welding mark image, respectively, N is the dimension of the feature vector, Represents the cosine similarity between the previous welding mark image and the current welding mark image;

[0092] According to the cosine similarity calculation results, the similarity difference index is generated, and the generated expression is: , where Represents the similarity difference index, represents the weight factor, which is used to adjust the sensitivity of the similarity difference index;

[0093] The role of the weight factor is to amplify or reduce the impact of the similarity difference on the final index according to the specific application scenario or quality requirements. By properly setting the weight factor, the sensitivity to small changes can be increased, making the system more sensitive in detecting potential anomalies; conversely, reducing the weight factor can reduce the response to small differences and reduce false alarms. The selection of weight factors is usually based on experience or experimental data, and needs to be adjusted according to specific welding quality standards and process requirements to ensure that the similarity difference index has reasonable detection accuracy and reliability in practical applications. No specific restrictions are made here.

[0094] The similarity analysis of the welding traces of the last welding and the welding traces of the current welding is performed. The larger the performance value of the generated similarity difference index, the greater the difference between the welding trace image of the last welding and the welding trace image of the current welding; conversely, the smaller the performance value of the similarity difference index, the smaller the difference between the two welding trace images. Specifically, the similarity difference index is generated by comparing the characteristic parameters of the two welding traces to measure their consistency. A larger similarity difference index indicates that there are significant changes in the welding conditions, weld shape or surface quality during the two welding processes, which may affect the welding quality; a smaller similarity difference index indicates that the welding process is relatively stable, the weld characteristics remain consistent, and the welding quality is relatively reliable.

[0095] The similarity difference index generated after similarity analysis of the welding traces of the previous welding and the welding traces of the current welding is compared and analyzed with the pre-set similarity difference index reference threshold to evaluate the quality of the current welding trace. The specific steps are as follows:

[0096] If the similarity difference index is greater than or equal to the similarity difference index reference threshold, a low-quality welding signal is generated, and the current welding trace is marked as low-quality welding;

[0097] If the similarity difference index is less than the similarity difference index reference threshold, a high-quality welding signal is generated and the current welding trace is marked as high-quality welding.

[0098] Based on the results of the comparison analysis, determine whether the potential abnormalities in the welding process have an impact on the welding quality. If the comparison results show that there are obvious differences in the welding marks and they affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process, and the welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues;

[0099] When a low-quality welding signal is generated after a similarity analysis is performed on the welding traces of the previous welding and the welding traces of the current welding, it is determined that the potential abnormality has a negative impact on the welding process, the welding process is stopped and maintenance and management of the industrial welding robot is performed. When a high-quality welding signal is generated after a similarity analysis is performed on the welding traces of the previous welding and the welding traces of the current welding, it is determined that the abnormality has no substantial impact on the welding quality and the welding work continues.

[0100] The present invention can timely identify potential anomalies in the welding process by acquiring control information of the welding process in real time when the industrial welding robot controls the welding head, and analyzing the characteristic parameters in combination with a pre-trained machine learning model. If potential anomalies are found, the welding traces are further acquired through an image sensor for detailed comparison and evaluation. This process ensures that potential abnormalities can be discovered and marked in a timely manner during the welding process, preventing undetected anomalies from affecting the welding quality in subsequent batch production, thereby improving the reliability and consistency of the overall welding process.

[0101] When the present invention identifies the existence of potential abnormalities in the welding process, the last welding mark and the current welding mark are obtained by the image sensor, and the image processing technology is used to perform a detailed similarity analysis to generate a similarity difference index. The similarity difference index is compared with a preset reference threshold value, so as to accurately judge whether the potential abnormalities in the welding process have an impact on the welding quality. If the similarity difference index exceeds the threshold value, the current welding mark is marked as low-quality welding, the welding process is stopped in time and maintenance management is carried out to prevent unqualified products from entering subsequent production links. If the similarity difference index is within the threshold range, it indicates that the welding quality is not affected by the abnormality and the welding operation can continue, thereby providing an accurate quality assessment means by comparing the welding marks, ensuring the high quality of the welding products and the stability of the production process.

[0102] The present invention provides Figure 2 An industrial robot control system shown includes a data acquisition module, a data processing and anomaly detection module, an image acquisition module, an image analysis and quality assessment module, and a quality judgment and maintenance module;

[0103] The data acquisition module acquires the control information of the welding process in real time through sensors and controllers when the industrial welding robot controls the welding head, and transmits the acquired information to the central control system;

[0104] The data processing and anomaly detection module organizes the acquired information, extracts features from the organized information, performs anomaly analysis on the extracted features, and then inputs the extracted features into a pre-trained machine learning model for analysis to determine whether there are potential anomalies in the current welding process;

[0105] Image acquisition module, if the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor;

[0106] Image analysis and quality assessment module, which uses image processing technology to compare the difference between the previous welding mark and the current welding mark, and assess the quality of the current welding mark;

[0107] The quality judgment and maintenance module determines whether the potential abnormalities in the welding process have an impact on the welding quality based on the results of the comparison analysis. If the comparison results show that there are obvious differences in the welding marks and they affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process, the welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues;

[0108] An industrial robot control method provided in an embodiment of the present invention is implemented by the above-mentioned industrial robot control system. The specific method and process of an industrial robot control system are detailed in the embodiment of the above-mentioned industrial robot control method, which will not be repeated here.

[0109] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0110] The above embodiments may be implemented in whole or in part by software, hardware, firmware or any other combination thereof. When implemented by software, the above embodiments may be implemented in whole or in part in the form of a computer program product. The computer program product includes one or more computer instructions or computer programs. When the computer instructions or computer programs are loaded or executed on a computer, the process or function described in the embodiment of the present application is generated in whole or in part. The computer may be a general-purpose computer, a special-purpose computer, a computer network, or other programmable device. The computer instructions may be stored in a computer-readable storage medium, or transmitted from one computer-readable storage medium to another computer-readable storage medium. For example, the computer instructions may be transmitted from one website, computer, server or data center to another website, computer, server or data center by wired or wireless (e.g., infrared, wireless, microwave, etc.) means. The computer-readable storage medium may be any available medium that a computer can access or a data storage device such as a server or data center that contains one or more available media sets. The available medium may be a magnetic medium (e.g., a floppy disk, a hard disk, a tape), an optical medium (e.g., a DVD), or a semiconductor medium. The semiconductor medium may be a solid-state hard disk.

[0111] It should be understood that in the various embodiments of the present application, the size of the serial numbers of the above-mentioned processes does not mean the order of execution. The execution order of each process should be determined by its function and internal logic, and should not constitute any limitation on the implementation process of the embodiments of the present application.

[0112] Those of ordinary skill in the art will appreciate that the units and algorithm steps of each example described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are performed in hardware or software depends on the specific application and design constraints of the technical solution. Professional and technical personnel can use different methods to implement the described functions for each specific application, but such implementation should not be considered to be beyond the scope of this application.

[0113] In the several embodiments provided in the present application, it should be understood that the disclosed systems and methods can be implemented in other ways. For example, the embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation, such as multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be an indirect coupling or communication connection through some interfaces, devices or units, which can be electrical, mechanical or other forms.

[0114] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0115] In addition, each functional unit in each embodiment of the present application may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0116] The above is only a specific implementation of the present application, but the protection scope of the present application is not limited thereto. Any person skilled in the art who is familiar with the present technical field can easily think of changes or substitutions within the technical scope disclosed in the present application, which should be included in the protection scope of the present application. Therefore, the protection scope of the present application should be based on the protection scope of the claims.

Claims

1. A method for controlling an industrial robot, characterized in that: The following steps are involved: When the industrial welding robot controls the welding head, it obtains the control information of the welding process in real time through sensors and controllers, and transmits the acquired information to the central control system; The acquired information is sorted, and features are extracted from the sorted information. The extracted features are then analyzed for abnormalities. The extracted features are then input into a pre-trained machine learning model for analysis to determine whether there are potential abnormalities in the current welding process. If the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor; By comparing the difference between the last welding mark and the current welding mark through image processing technology, the quality of the current welding mark is evaluated; Based on the results of the comparison analysis, determine whether the potential abnormalities in the welding process have an impact on the welding quality. If the comparison results show that there are obvious differences in the welding marks and they affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process, and the welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues; Feature extraction is performed on the sorted information, wherein the extracted features include heat input, wire feeding speed and welding speed during welding. After extraction, abnormal analysis and processing is performed on the heat input during welding to generate a heat input unevenness index, and abnormal analysis and processing is performed on the wire feeding speed and welding speed to generate a wire feeding speed consistency index; The heat input unevenness index and wire feeding speed consistency index after abnormal analysis are input into the pre-trained machine learning model, and the welding coefficient is generated by the model. The welding coefficient is used to determine whether there is potential abnormality in the current welding process; The specific steps for analyzing the heat input during welding to generate the heat input unevenness index are as follows: In a welding process, the change function of welding current over time I(t) and the change function of welding voltage over time V(t) are obtained, and the instantaneous heat input power P(t) is calculated, which represents the heat input at time t. The calculation expression is: P(t) = I(t)·V(t); Calculate the cumulative heat input energy E(t) during a welding process, which represents the cumulative heat input energy at time t. The calculation expression is: In the formula, t1 and t2 represent the start time and end time of a welding process respectively, and the time period from t1 to t2 is the total duration of a welding process; Calculate the heat input density H(t), which represents the heat input density at time t. The calculation expression is: Where L(t) represents the cumulative welding length at time t, and the calculation expression is: Among them, v w (t) represents the variation function of welding speed with time; Calculate the local heat input non-uniformity, the calculation expression is: Where ΔH(t) represents the rate of change of heat input density, U i Indicates that at t i The local heat input non-uniformity at time ΔH(t i ) indicates that at t i The rate of change of heat input density at time instant, n represents the total number of time points; Calculate the heat input unevenness index, the calculation expression is: In the formula, HIUI Q Indicates the heat input unevenness index; The specific steps for abnormal analysis and processing of wire feeding speed and welding speed to generate wire feeding speed consistency index are as follows: During a welding process, the wire feeding speed data is collected in real time, and the function v of the wire feeding speed changing with time is plotted. s (t) to make an indication; Calculate the difference between the wire feeding speed and the welding speed at each moment, Δv(t), and the calculation expression is: Δv(t) = v s (t)-v w (t); The absolute value of the speed difference is integrated to calculate the total difference D in one welding process. The calculation expression is: Where T is the total length of one welding time; Calculate the wire feeding speed consistency index, the calculation expression is: Wire x =1-D norm , where D norm represents the total difference after normalization, Wire x Indicates the wire feeding speed consistency index; The welding coefficient generated by the industrial welding robot during a welding operation is compared and analyzed with the preset welding coefficient reference threshold to determine whether there is a potential abnormality in the current welding process. The judgment process is as follows: If the welding coefficient is greater than or equal to the welding coefficient reference threshold, the welding process of the industrial welding robot is marked as a potential abnormality; If the welding coefficient is less than the welding coefficient reference threshold, the process of the industrial welding robot welding is marked as efficient welding; When there is a potential abnormality in the welding process, the welding traces of the previous welding and the welding traces of the current welding are analyzed for similarity, and a similarity difference index is generated. The quality of the current welding trace is evaluated by the similarity difference index; The steps for generating the similarity difference index are as follows: Before and after the welding process begins, the welding mark image I of the previous welding is obtained through the image sensor perv and the welding trace image I during the current welding curr ; Preprocess the acquired images, including grayscale, denoising and edge detection; Use shape descriptors to extract weld shape features from preprocessed images; The cosine similarity between the previous welding mark image and the current welding mark image is calculated using cosine similarity. The calculation expression is: Among them, F perv，k and F curr，k are the feature vector components of the previous welding mark image and the current welding mark image, N is the dimension of the feature vector, Similarity(I perv , I curr ) represents the cosine similarity between the previous welding mark image and the current welding mark image; According to the cosine similarity calculation results, the similarity difference index is generated, and the generated expression is: In the formula, Difference Index Represents the similarity difference index, and ω represents the weight factor, which is used to adjust the sensitivity of the similarity difference index.

2. An industrial robot control system, used to implement the industrial robot control method according to claim 1, characterized in that: It includes data acquisition module, data processing and anomaly detection module, image acquisition module, image analysis and quality assessment module, and quality judgment and maintenance module; The data acquisition module acquires the control information of the welding process in real time through sensors and controllers when the industrial welding robot controls the welding head, and transmits the acquired information to the central control system; The data processing and anomaly detection module organizes the acquired information, extracts features from the organized information, performs anomaly analysis on the extracted features, and then inputs the extracted features into a pre-trained machine learning model for analysis to determine whether there are potential anomalies in the current welding process; Image acquisition module, if the model determines that there is a potential abnormality in the welding process, the welding system obtains the welding traces of the previous welding and the welding traces of the current welding through the image sensor; Image analysis and quality assessment module, which uses image processing technology to compare the difference between the previous welding mark and the current welding mark, and assess the quality of the current welding mark; The quality judgment and maintenance module determines whether the potential abnormalities in the welding process have an impact on the welding quality based on the results of the comparison analysis. If the comparison results show that there are obvious differences in the welding marks and they affect the welding quality, it is determined that the potential abnormalities have a negative impact on the welding process, the welding process is stopped and the industrial welding robot is maintained and managed. If the comparison results show that there are no obvious differences in the welding marks, it is determined that the abnormalities have no substantial impact on the welding quality and the welding work continues.

Images