Predictive maintenance method, system, medium and equipment for welding robot
By building digital twin models and communication connections of welding robots, the failure prediction and maintenance strategies are automated, and the problem of failure unvisible and maintenance strategies in the existing technology is solved, and the work efficiency and production progress of welding robots are improved.
Patent Information
- Application Number
- CN202411037697.X
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-07-30
- Publication Date
- 2024-11-19
AI Technical Summary
The existing welding robot control system cannot realize real-time fault visual linkage and timely maintenance strategy scheduling, resulting in low work efficiency and production progress.
By building a digital twin model of welding robot, establishing a communication connection between the real and digital twin models, obtaining welding data, combining the fault prediction model for fault prediction and self-judgment, mapping to a three-dimensional virtual model for linkage display, and generating according to the predicted fault type Maintenance strategies and control instructions.
It improves the visualization and work efficiency of welding robot work, and improves the overall production progress.
Smart Images

Figure CN118981848A_ABST
Abstract
Description
Technical Field
[0001] The present invention belongs to the technical field of digital twin robots, and specifically relates to a welding robot predictive maintenance method, system, medium and equipment. Background Art
[0002] The statements in this section merely provide background information related to the present invention and do not necessarily constitute prior art.
[0003] Since welding work has the characteristics of harsh working conditions and high repetitiveness, it is particularly suitable for industrial welding robots. Industrial welding robots have gradually replaced traditional manual positioning welding. Each industrial welding robot is composed of thousands of parts. In order to improve troubleshooting efficiency, improve the yield rate and avoid accidents during operation, it is a good means to monitor the operating status of industrial welding robots and diagnose faults, and predict and identify the faulty parts of industrial welding robots. If the industrial welding robot is repaired and maintained after a fault occurs, it will directly affect its work efficiency and overall production progress. Therefore, it is particularly important to perform predictive maintenance on industrial welding robots, so that parts that may fail can be maintained and replaced in advance.
[0004] The inventors found that in the existing welding robot control, although a digital twin model of the robot was constructed, on the one hand, the failures of the existing welding robots could not be realized in real-time visual linkage, and on the other hand, the corresponding maintenance strategies could not be dispatched in time according to the failures that occurred, resulting in low work efficiency and affecting the overall production progress. Summary of the invention
[0005] In order to solve the above problems, the present invention proposes a predictive maintenance method, system, medium and equipment for a welding robot. The present invention improves the visualization level and the working efficiency of the welding robot through the visual linkage between the real welding robot and the corresponding digital twin model, and is based on a maintenance strategy that can predict corresponding faults.
[0006] According to some embodiments, the present invention adopts the following technical solutions:
[0007] A welding robot predictive maintenance method comprises the following steps:
[0008] Build a digital twin model of the welding robot;
[0009] Establish a communication connection between the real welding robot and the digital twin model;
[0010] Obtain welding data during the welding process of the welding robot;
[0011] Based on the acquired welding data and combined with the fault prediction model, the fault prediction result is obtained;
[0012] The fault prediction results and the robot's spatial trajectory self-judgment results are mapped to the digital twin model. The key parts of the welding robot are marked in the digital twin model. When a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner.
[0013] The constructed maintenance strategy database is called according to the predicted fault type, and the corresponding control command signal is generated and sent to the welding robot to execute the command.
[0014] As an optional implementation, the establishing of a communication connection between the real welding robot and the digital twin model specifically includes:
[0015] The established digital twin model of the welding robot is lightweighted and rendered to obtain a three-dimensional model file in OBJ format;
[0016] Import the OBJ format 3D model file into Unity3D to build the scene and obtain the 3D virtual model of the welding robot;
[0017] Establish communication between the real welding robot and the 3D virtual model, read the data of the welding robot's welding process through the OPC UA protocol, and send the data of the real welding robot's welding process to the 3D virtual model of the welding robot, thereby realizing the connection with the robot's sensor data, realizing virtual-reality mapping, and driving the synchronous movement of the 3D virtual model of the welding robot.
[0018] As an optional implementation, the fault prediction result is obtained based on the acquired welding data and in combination with the fault prediction model, including:
[0019] Preprocess the acquired welding data;
[0020] Based on the preprocessed welding data, feature extraction is performed to obtain welding data features; wherein the welding data features include mean, standard deviation, root square amplitude, root mean square value, peak value, skewness and kurtosis;
[0021] The fault prediction results are obtained by combining the welding data features with the LSTM model for deep learning.
[0022] As an optional implementation, the robot's spatial trajectory self-judgment process includes:
[0023] The standard trajectory of each axis of the industrial welding robot is preset; the standard trajectory is determined by an algorithm that removes abnormal outliers after multiple operations and takes the average value;
[0024] Set axis motion range threshold;
[0025] The real-time acquired trajectory of each axis is compared with the preset standard trajectory, and situations that exceed the axis motion range threshold are marked and warned.
[0026] As an optional implementation method, when marking the key parts of the welding robot in the digital twin model, they are marked according to the importance ratio of the parts. The health status of the key parts of the robot is divided into: normal, warning, and fault. When marking, colors are used for marking. Green indicates normal, yellow indicates warning, and red indicates fault.
[0027] As an optional implementation, the welding data includes vibration signals of each axis, current signals of axis motors, angle distribution, welding current, welding temperature and wire feeding current;
[0028] As a further example, the specific details of the fault include the fault occurrence event, type, cause, phenomenon and location of the fault.
[0029] As an optional implementation, the specific details of the welding robot failure and the corresponding maintenance strategy specifically include:
[0030] If the robot water circulation system fails, the corresponding maintenance strategy is: check whether the cable harness is damaged, check whether the welding gun is damaged, check whether the welding gun locking system is normal, check whether the welding gun adapter is normal, and check whether the adapter one-way valve is damaged;
[0031] If the robot welding system fails, the corresponding maintenance strategy is: check whether the welding-related equipment and wiring harness are loose or damaged;
[0032] If the robot wire feeding system fails, the corresponding maintenance strategy is: check whether the cable bundle wire feeding tube and wire feeding wheel are damaged, replace the cable bundle wire feeding tube and wire feeding wheel and redo the push-pull synchronization;
[0033] If the robot control system fails, the corresponding maintenance strategy is: check whether the plugs, control panels and communication signals of each part are normal, and clean and plug them again;
[0034] If the robot's electrical system fails, the corresponding maintenance strategy is: check whether the corresponding components are damaged, replace them and restart the machine for testing.
[0035] A welding robot predictive maintenance system, comprising:
[0036] A digital twin model building module, which is used to build a digital twin model of a welding robot;
[0037] A communication connection establishment module, which is used to establish a communication connection between the real welding robot and the digital twin model;
[0038] A data acquisition module, which is used to acquire welding data during the welding process of the welding robot;
[0039] A fault prediction module is used to obtain a fault prediction result based on the acquired welding data and in combination with a fault prediction model;
[0040] The fault visualization module is used to map the fault prediction results and the robot's spatial trajectory self-judgment results to the digital twin model, mark the key parts of the welding robot in the digital twin model, and when a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner;
[0041] The maintenance strategy prediction module is used to call the constructed maintenance strategy database according to the predicted fault type, generate the corresponding control command signal, send it to the welding robot, and execute the command.
[0042] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps in the above-mentioned method for predictive maintenance of a welding robot.
[0043] A computer device comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein the processor implements the steps in the above-mentioned method for predictive maintenance of a welding robot when executing the program.
[0044] Compared with the prior art, the present invention has the following beneficial effects:
[0045] The present invention establishes a communication connection between a real welding robot and a digital twin model, maps fault prediction results and the robot's spatial trajectory self-judgment results to the digital twin model, marks key parts of the welding robot in the digital twin model, and when a key part fails, the corresponding fault location variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner, thereby improving the visualization of the industrial robot's working process, while also improving the working efficiency and overall production progress of the welding robot.
[0046] In order to make the above-mentioned objectives, features and advantages of the present disclosure more obvious and easy to understand, preferred embodiments are specifically cited below and described in detail with reference to the attached drawings. BRIEF DESCRIPTION OF THE DRAWINGS
[0047] The accompanying drawings constituting a part of the present disclosure are used to provide a further understanding of the present disclosure. The illustrative embodiments of the present disclosure and their descriptions are used to explain the present disclosure and do not constitute an improper limitation on the present disclosure.
[0048] Figure 1 This is a schematic diagram of the overall process of a welding robot predictive maintenance method of this embodiment. DETAILED DESCRIPTION
[0049] The present disclosure is further described below in conjunction with the accompanying drawings and embodiments.
[0050] It should be noted that the following detailed descriptions are all illustrative and are intended to provide further explanation of the present disclosure. Unless otherwise specified, all technical and scientific terms used herein have the same meanings as those commonly understood by those skilled in the art to which the present disclosure belongs.
[0051] It should be noted that the terms used herein are only for describing specific embodiments and are not intended to limit the exemplary embodiments according to the present disclosure. As used herein, unless the context clearly indicates otherwise, the singular form is also intended to include the plural form. In addition, it should be understood that when the terms "comprising" and / or "including" are used in this specification, it indicates the presence of features, steps, operations, devices, components and / or combinations thereof.
[0052] Explanation of terms
[0053] OPC UA (Open Platform Communications Unified Architecture) is an open standard protocol based on industrial communications, designed to achieve interoperability and data sharing between devices. It is a technical specification for communication between different devices and systems. Through OPC UA, various devices and systems can communicate and share data with each other, achieving more efficient industrial automation.
[0054] MQTT (Message Queue Telemetry Transport) is a message protocol based on the publish / subscribe paradigm under the ISO standard (ISO / IEC PRF 20922). It works on the TCP / IP protocol family and is a publish / subscribe message protocol designed for remote devices with low hardware performance and poor network conditions. For this purpose, it requires a message middleware.
[0055] MQTT is a client-server based message publish / subscribe transport protocol. The MQTT protocol is lightweight, simple, open and easy to implement, which makes it widely applicable. In many cases, including restricted environments such as machine-to-machine (M2M) communication and the Internet of Things (IoT). It has been widely used in sensors communicating via satellite links, medical devices that occasionally dial up, smart homes, and some miniaturized devices.
[0056] Welding robots are industrial robots that perform welding (including cutting and spraying). According to the definition of the International Organization for Standardization (ISO), industrial robots belong to standard welding robots. Industrial robots are multi-purpose, reprogrammable, automatically controlled manipulators with three or more programmable axes used in the field of industrial automation. In order to adapt to different uses, the mechanical interface of the last axis of the robot is usually a connecting flange that can be connected to different tools or end effectors. The welding robot is an industrial robot with a welding tongs or welding (cutting) gun installed on the end axis flange, so that it can perform welding, cutting or thermal spraying.
[0057] Each industrial welding robot is made up of thousands of parts. Predicting and identifying the faulty parts of the industrial welding robot can improve troubleshooting efficiency, increase the yield rate, and avoid accidents during operation. If the industrial welding robot is repaired and maintained after a fault occurs, it will directly affect its work efficiency and overall production progress.
[0058] In the existing welding robot control, although a digital twin model of the robot is constructed, on the one hand, the existing welding robot failure cannot achieve real-time visualization linkage, and on the other hand, the corresponding maintenance strategy cannot be scheduled in time according to the failure, resulting in low work efficiency and affecting the overall production progress. In order to solve the above problems, the present disclosure proposes a welding robot predictive maintenance method, system, medium and equipment.
[0059] The present disclosure improves the visualization level and the working efficiency of the welding robot by visually linking the real welding robot and the corresponding digital twin model, and is based on a maintenance strategy that can predict corresponding faults.
[0060] like Figure 1 As shown, the present invention is based on digital twins and is divided into a device layer, a platform layer, and an application layer. By acquiring the axis motion data of the industrial welding robot and processing the data at the platform layer, the synchronous driving of the three-dimensional virtual model of the computer application layer and the self-judgment of the trajectory of the industrial welding robot are realized. The collected signal data is transmitted to the platform layer through the above-mentioned sensor, and the data features are extracted by using algorithms such as data processing and deep learning to diagnose the fault of the industrial welding robot, reasonably predict the location of the fault, and automatically record the fault information into the system, and the fault information can be managed at the application layer.
[0061] Embodiment 1
[0062] This embodiment provides a welding robot predictive maintenance method, comprising the following steps:
[0063] Step 1: Build a digital twin model of the welding robot;
[0064] In this embodiment, when constructing a digital twin model of a welding robot, the following steps are included:
[0065] Step 101: Import the welding robot 3D model into the digital twin construction platform;
[0066] Step 102: Subscribe to the robot's axis data through MQTT, and create a data set for all axis data;
[0067] Step 103: select each axis of the robot to create new dynamic data, and associate the dynamic data from the above data set with each axis of the robot respectively;
[0068] Step 104: Read the real-time axis data stored in the database, and realize the synchronous movement of the real robot and the virtual robot through MQTT publishing.
[0069] Step 2: Establish a communication connection between the real welding robot and the digital twin model;
[0070] In step 2, the process of establishing the communication connection between the real welding robot and the digital twin model includes:
[0071] Step 201: lightweighting and rendering the established digital twin model of the welding robot to obtain a three-dimensional model file in OBJ format;
[0072] Step 202: Import the OBJ format 3D model file into Unity3D to build a scene and obtain a 3D virtual model of the welding robot;
[0073] Step 203: establishing communication between the real welding robot and the three-dimensional virtual model, reading the data of the welding process of the welding robot through the OPC UA protocol, and sending the data of the welding process of the real welding robot to the three-dimensional virtual model of the welding robot, thereby realizing connection with the robot sensor data, realizing virtual-real mapping, and driving the synchronous movement of the three-dimensional virtual model of the welding robot;
[0074] Through the above steps, the digital twin model of the welding robot and the driving mechanism of the digital twin model of the welding robot were established.
[0075] Step 3: Acquire welding data during the welding process of the welding robot;
[0076] In this embodiment, the welding data includes the vibration signal of each axis, the current signal of the axis motor, the angle distribution, the welding current, the welding temperature and the wire feeding current;
[0077] The above welding data is collected by sensors installed on the welding robot. The vibration signal of each axis is obtained by the vibration sensor, the angle is obtained by the axis motor angle sensor, and the welding temperature is obtained by the temperature sensor.
[0078] A variety of data during the welding process can be obtained through a variety of sensors, and the status of each component of the welding robot during its working process can be obtained more accurately, which is conducive to analyzing the occurrence of faults during the welding process.
[0079] Step 4: Based on the acquired welding data, the fault prediction results are obtained in combination with the fault prediction model; specifically including:
[0080] Step 401: preprocessing welding data;
[0081] Among them, preprocessing includes data cleaning, normalization and standardization, the purpose of which is to improve the learning effect and accuracy of the model;
[0082] The multidimensional data information is cleaned to obtain a first result, wherein the cleaning process includes filling in missing data and removing abnormal data;
[0083] The Min-max normalization method is used to normalize the first result to obtain the second result, which is calculated as follows:
[0084] The formula is as follows:
[0085]
[0086] In the formula, x i 、x max 、x min Respectively represent the sample value, the maximum value in the sample, and the minimum value in the sample, x i ' is the normalized result of the first cleaning result;
[0087] It should be noted that the collection of multi-dimensional data information can ensure the sufficiency and real-time nature of the original data. The collected equipment operation data is multi-dimensional time series data. In order to match the input conditions of the prediction model and speed up the training of the model, a deep processing of the multi-dimensional data is realized. Therefore, the collected data set needs to be preprocessed. Step 402: Extract features based on the preprocessed welding data to obtain welding data features;
[0088] Feature extraction of signals in the time domain includes mean, standard deviation, root amplitude, root mean square value, peak value, skewness and kurtosis;
[0089] Among them, the mean is:
[0090] The standard deviation is:
[0091] Square root amplitude:
[0092] RMS value:
[0093] Peak value: x p =max|x(n)|,
[0094] Skewness:
[0095] Kurtosis:
[0096] In the formula, x(n) is the welding data, N is the total number of welding data, and n is the nth welding data; is the cube of the standard deviation of the welding data, is the fourth power of the standard deviation of the welding data.
[0097] By extracting the above features, the characteristics of the data can be reflected from many aspects such as discreteness, deviation, range of variation, deviation, sharpness, etc.
[0098] Step 403: Combine the welding data features and the fault prediction model to perform deep learning to obtain a fault prediction result.
[0099] In this embodiment, the fault prediction model can adopt LSTM model, CNN, RNN, CAE, etc., and the LSTM model is preferred in this embodiment.
[0100] To build an LSTM model for deep learning, first define the model structure. LSTM is a special recurrent neural network. The key lies in its internal storage unit and three "gate" structures, which are convenient for processing long sequence data; then select the Adam optimizer to compile the model, define the loss function: mean square error (MSE), and select the evaluation indicator: accuracy. Finally, input the preprocessed data into the model for training, and use the test set to evaluate the performance of the model. The difference between the predicted value and the actual value can be calculated: mean square error (MSE) or root mean square error (RMSE). The multi-state time series model is set using the LSTM network model structure, wherein the multi-state time series model includes an input layer, a hidden layer, and an output layer; including the following steps:
[0101] Data partitioning: The preprocessed dataset is divided into a training dataset and a test dataset. The training dataset is used for the model training process, while the test dataset is used to evaluate the model's predictive ability on new data.
[0102] Model training: Use the training data set to train the set multi-state time series model, and optimize the model parameters through iteration so that the model can better fit the characteristics and patterns of historical data.
[0103] Model testing: Use the test data set to test the trained multi-state time series model to determine whether the model output meets the preset prediction indicators. Evaluate the prediction accuracy of the model by comparing the model's prediction results with the actual observations.
[0104] Model optimization: If the model output does not meet the preset prediction indicators, optimize the multi-state time series model. You can adjust the model structure, parameters or training strategy to improve the model's prediction performance until the prediction results output by the model in the test phase meet the prediction indicators.
[0105] The technical effect of the above technical solution is: the above technical solution of this embodiment predicts the operation status of robot welding. By using the LSTM network model, the historical data of the equipment can be modeled and analyzed to capture the time correlation and sequence characteristics in the data. After training and testing, a multi-state time series model with high prediction accuracy can be obtained. This model can be used to predict the future state and performance of the equipment and provide support for predictive maintenance and fault diagnosis. Step 5: Combine the fault prediction results and the robot's spatial trajectory self-judgment results, map the fault prediction results to the three-dimensional virtual model, mark the key parts of the welding robot in the three-dimensional virtual model, and when a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in linkage, and the corresponding control command signal is generated and sent to the welding robot to execute the corresponding command.
[0106] The robot's spatial trajectory self-judgment process includes:
[0107] First, the standard trajectory of each axis of the industrial welding robot is preset; the standard trajectory is determined by using an algorithm that removes abnormal outliers and takes the average value after multiple operations;
[0108] Then set the axis motion range threshold to meet the μ±nσ principle; where μ represents the mean of the data, σ represents the standard deviation of the data, and n is a parameter related to the robot accuracy. in It means rounding down, and s means the absolute position accuracy of the robot, which is usually given by the manufacturer.
[0109] The real-time acquired trajectory of each axis is compared with the preset standard trajectory, and situations that exceed the axis motion range threshold are marked and warned.
[0110] In this embodiment, when marking the key parts of the welding robot in the three-dimensional virtual model, the specific parts where the fault may occur are marked according to the importance ratio of the parts, such as the robot body axis motor failure, welding wire feeding failure, water circulation failure, welding system failure, etc., and the fault parts are marked on the three-dimensional virtual model to facilitate the staff to view, inspect and maintain;
[0111] In this embodiment, the health status of the key parts of the robot is divided into: normal, warning, and fault. The color and corresponding variables of the robot joint components are set, and the material color of the key parts of the robot is changed according to the acquired status data: green indicates normal, yellow indicates warning, and red indicates fault.
[0112] When the algorithm detects a fault, it outputs the fault location variable, which is then input into the twin model, and the color display of the corresponding fault part will change. Driven by twin data, a virtual-reality mapping relationship is formed, and the virtual three-dimensional model follows the movement of the real robot in real time.
[0113] Through the above scheme, the results of deep learning fault prediction are combined with the variables returned by the robot trajectory self-judgment system. The digital twin model is controlled by variables, and the faults are displayed on the interface. The key parts of the robot can be displayed in the three-dimensional virtual model in time, so as to issue corresponding control instructions, which greatly improves the visualization and efficiency of fault detection.
[0114] Step 6: Establish a database of specific fault details and corresponding maintenance strategies for industrial welding robots, and predict corresponding maintenance strategies based on the fault types corresponding to the 3D virtual model.
[0115] In this embodiment, the specific details of the fault include the fault occurrence event, category, cause, phenomenon and location of the fault; later users can search, add, delete and edit them for easy viewing at a later time.
[0116] The establishment of specific fault details and corresponding maintenance strategies of the industrial welding robot specifically includes:
[0117] If the robot water circulation system fails, the corresponding maintenance strategy is: check whether the cable harness is damaged, check whether the welding gun is damaged, check whether the welding gun locking system is normal, check whether the welding gun adapter is normal, and check whether the adapter check valve is damaged;
[0118] If the robot welding system fails, the corresponding maintenance strategy is: check whether the welding-related equipment and wiring harness are loose or damaged;
[0119] If the robot wire feeding system fails, the corresponding maintenance strategy is: check whether the cable bundle wire feeding tube and wire feeding wheel are damaged, replace the cable bundle wire feeding tube and wire feeding wheel and redo the push-pull synchronization;
[0120] If the robot control system fails, the corresponding maintenance strategy is: check whether the plugs, control panels and communication signals of each part are normal, and clean and plug them again;
[0121] If the robot's electrical system fails, the corresponding maintenance strategy is: check whether the corresponding components are damaged, replace them and restart the machine for testing.
[0122] Through the visual linkage between the real welding robot and the corresponding digital twin model, combined with a maintenance strategy that can predict corresponding faults, the visualization level and the work efficiency of the welding robot are improved.
[0123] Embodiment 2
[0124] This embodiment provides a welding robot predictive maintenance system, including:
[0125] A digital twin model building module, which is used to build a digital twin model of a welding robot;
[0126] A communication connection establishment module, which is used to establish a communication connection between the real welding robot and the digital twin model;
[0127] A data acquisition module, which is used to acquire welding data during the welding process of the welding robot;
[0128] A fault prediction module is used to obtain a fault prediction result based on the acquired welding data and in combination with a fault prediction model;
[0129] The fault visualization module is used to map the fault prediction results and the robot's spatial trajectory self-judgment results to the digital twin model, mark the key parts of the welding robot in the digital twin model, and when a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner;
[0130] The maintenance strategy prediction module is used to call the constructed maintenance strategy database according to the predicted fault type, generate the corresponding control command signal, send it to the welding robot, and execute the command.
[0131] Among them, in the digital twin model construction module, when building the digital twin model of the welding robot, the following steps are included:
[0132] Step 101: Import the welding robot 3D model into the digital twin construction platform;
[0133] Step 102: Subscribe to the robot's axis data through MQTT, and create a data set for all axis data;
[0134] Step 103: select each axis of the robot to create new dynamic data, and associate the dynamic data from the above data set with each axis of the robot respectively;
[0135] Step 104: Read the real-time axis data stored in the database, and realize the synchronous movement of the real robot and the virtual robot through MQTT publishing.
[0136] Among them, in the communication connection establishment module, the establishment of the communication connection between the real welding robot and the digital twin model specifically includes:
[0137] The established digital twin model of the welding robot is lightweighted and rendered to obtain a three-dimensional model file in OBJ format;
[0138] Import the OBJ format 3D model file into Unity3D to build the scene and obtain the 3D virtual model of the welding robot;
[0139] Establish communication between the real welding robot and the 3D virtual model, read the data of the welding robot's welding process through the OPC UA protocol, and send the data of the real welding robot's welding process to the 3D virtual model of the welding robot, thereby realizing the connection with the robot's sensor data, realizing virtual-reality mapping, and driving the synchronous movement of the 3D virtual model of the welding robot.
[0140] Among them, in the fault prediction module, the fault prediction result is obtained based on the acquired welding data and combined with the fault prediction model, including:
[0141] Preprocess the acquired welding data;
[0142] Based on the preprocessed welding data, feature extraction is performed to obtain welding data features; wherein the welding data features include mean, standard deviation, root square amplitude, root mean square value, peak value, skewness and kurtosis;
[0143] The fault prediction results are obtained by combining the welding data features with the LSTM model for deep learning.
[0144] By extracting the above features, the characteristics of the data can be reflected from multiple aspects such as discreteness, deviation, range of variation, deviation, sharpness, etc., and combined with the LSTM model for deep learning, welding failures can be better learned.
[0145] Among them, in the fault visualization module, the robot's spatial trajectory self-judgment process includes:
[0146] The standard trajectory of each axis of the industrial welding robot is preset; the standard trajectory is determined by an algorithm that removes abnormal outliers after multiple operations and takes the average value;
[0147] Set axis motion range threshold;
[0148] The real-time acquired trajectory of each axis is compared with the preset standard trajectory, and situations that exceed the axis motion range threshold are marked and warned.
[0149] When marking the key parts of the welding robot in the digital twin model, they are marked according to the importance ratio of the parts. The health status of the key parts of the robot is divided into: normal, warning, and fault. When marking, colors are used for marking. Green indicates normal, yellow indicates warning, and red indicates fault.
[0150] The welding data includes the vibration signal of each axis, the current signal of the axis motor, the angle distribution, the welding current, the welding temperature and the wire feeding current;
[0151] The above welding data is collected by sensors installed on the welding robot. The vibration signal of each axis is obtained by the vibration sensor, the angle is obtained by the axis motor angle sensor, and the welding temperature is obtained by the temperature sensor.
[0152] Alternatively, the specific details of the fault include the fault occurrence event, type, cause, phenomenon and location of the fault.
[0153] In the maintenance strategy prediction module, the maintenance strategy database constructed by calling according to the predicted fault type specifically includes:
[0154] If the robot water circulation system fails, the corresponding maintenance strategy is: check whether the cable harness is damaged, check whether the welding gun is damaged, check whether the welding gun locking system is normal, check whether the welding gun adapter is normal, and check whether the adapter one-way valve is damaged;
[0155] If the robot welding system fails, the corresponding maintenance strategy is: check whether the welding-related equipment and wiring harness are loose or damaged;
[0156] If the robot wire feeding system fails, the corresponding maintenance strategy is: check whether the cable bundle wire feeding tube and wire feeding wheel are damaged, replace the cable bundle wire feeding tube and wire feeding wheel and redo the push-pull synchronization;
[0157] If the robot control system fails, the corresponding maintenance strategy is: check whether the plugs, control panels and communication signals of each part are normal, and clean and plug them again;
[0158] If the robot's electrical system fails, the corresponding maintenance strategy is: check whether the corresponding components are damaged, replace them and restart the machine for testing.
[0159] Through the visual linkage between the real welding robot and the corresponding digital twin model, combined with a maintenance strategy that can predict corresponding faults, the visualization level and the work efficiency of the welding robot are improved.
[0160] Embodiment 3
[0161] This embodiment provides a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the steps in the predictive maintenance method for a welding robot as described above are implemented.
[0162] Embodiment 4
[0163] This embodiment provides a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, the steps in the predictive maintenance method for a welding robot as described above are implemented.
[0164] Embodiment 5
[0165] The present embodiment provides a program product, which is a computer program product, including a computer program, and when the computer program is executed by a processor, the steps in the predictive maintenance method of a welding robot as described above are implemented. It should be understood by those skilled in the art that the embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a hardware embodiment, a software embodiment, or an embodiment combining software and hardware. Moreover, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage and optical storage, etc.) containing computer-usable program codes.
[0166] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.
[0167] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.
[0168] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.
[0169] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. The storage medium can be a disk, an optical disk, a read-only memory (ROM) or a random access memory (RAM), etc.
[0170] The above description is only a preferred embodiment of the present disclosure and is not intended to limit the present disclosure. For those skilled in the art, the present disclosure may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present disclosure shall be included in the protection scope of the present disclosure.
[0171] Although the above describes the specific implementation methods of the present disclosure in conjunction with the accompanying drawings, it is not intended to limit the scope of protection of the present disclosure. Technical personnel in the relevant field should understand that on the basis of the technical solution of the present disclosure, various modifications or variations that can be made by those skilled in the art without creative work are still within the scope of protection of the present disclosure.
Claims
1. A welding robot predictive maintenance method, characterized in that: The following steps are involved: Build a digital twin model of the welding robot; Establish a communication connection between the real welding robot and the digital twin model; Obtain welding data during the welding process of the welding robot; Based on the acquired welding data and combined with the fault prediction model, the fault prediction result is obtained; The fault prediction results and the robot's spatial trajectory self-judgment results are mapped to the digital twin model. The key parts of the welding robot are marked in the digital twin model. When a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner. The constructed maintenance strategy database is called according to the predicted fault type, and the corresponding control command signal is generated and sent to the welding robot to execute the command.
2. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: The establishing of a communication connection between the real welding robot and the digital twin model specifically includes: The established digital twin model of the welding robot is lightweighted and rendered to obtain a three-dimensional model file in OBJ format; Import the OBJ format 3D model file into Unity3D to build the scene and obtain the 3D virtual model of the welding robot; Establish communication between the real welding robot and the 3D virtual model, read the data of the welding robot's welding process through the OPC UA protocol, and send the data of the real welding robot's welding process to the 3D virtual model of the welding robot, thereby realizing the connection with the robot's sensor data, realizing virtual-reality mapping, and driving the synchronous movement of the 3D virtual model of the welding robot.
3. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: The fault prediction result is obtained based on the acquired welding data and in combination with the fault prediction model, including: Preprocess the acquired welding data; Based on the preprocessed welding data, feature extraction is performed to obtain welding data features; wherein the welding data features include mean, standard deviation, root square amplitude, root mean square value, peak value, skewness and kurtosis; The fault prediction results are obtained by combining the welding data features with the LSTM model for deep learning.
4. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: The robot's spatial trajectory self-judgment process includes: The standard trajectory of each axis of the industrial welding robot is preset; the standard trajectory is determined by an algorithm that removes abnormal outliers after multiple operations and takes the average value; Set axis motion range threshold; The real-time acquired trajectory of each axis is compared with the preset standard trajectory, and situations that exceed the axis motion range threshold are marked and warned.
5. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: When marking the key parts of the welding robot in the digital twin model, they are marked according to the importance ratio of the parts. The health status of the key parts of the robot is divided into: normal, warning, and fault. When marking, colors are used for marking. Green indicates normal, yellow indicates warning, and red indicates fault.
6. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: The welding data includes the vibration signal of each axis, the current signal of the axis motor, the angle distribution, the welding current, the welding temperature and the wire feeding current; Alternatively, the specific details of the fault include the fault occurrence event, type, cause, phenomenon and location of the fault.
7. A welding robot predictive maintenance method as claimed in claim 1, characterized in that: The maintenance strategy database constructed by calling according to the predicted fault type specifically includes: If the robot water circulation system fails, the corresponding maintenance strategy is: check whether the cable harness is damaged, check whether the welding gun is damaged, check whether the welding gun locking system is normal, check whether the welding gun adapter is normal, and check whether the adapter one-way valve is damaged; If the robot welding system fails, the corresponding maintenance strategy is: check whether the welding-related equipment and wiring harness are loose or damaged; If the robot wire feeding system fails, the corresponding maintenance strategy is: check whether the cable bundle wire feeding tube and wire feeding wheel are damaged, replace the cable bundle wire feeding tube and wire feeding wheel and redo the push-pull synchronization; If the robot control system fails, the corresponding maintenance strategy is: check whether the plugs, control panels and communication signals of each part are normal, and clean and plug them again; If the robot's electrical system fails, the corresponding maintenance strategy is: check whether the corresponding components are damaged, replace them and restart the machine for testing.
8. A welding robot predictive maintenance system, characterized in that: include: A digital twin model building module, which is used to build a digital twin model of a welding robot; A communication connection establishment module, which is used to establish a communication connection between the real welding robot and the digital twin model; A data acquisition module, which is used to acquire welding data during the welding process of the welding robot; A fault prediction module is used to obtain a fault prediction result based on the acquired welding data and in combination with a fault prediction model; The fault visualization module is used to map the fault prediction results and the robot's spatial trajectory self-judgment results to the digital twin model, mark the key parts of the welding robot in the digital twin model, and when a key part fails, the corresponding fault position variable is fed back to the three-dimensional virtual model, and the corresponding fault part is displayed in a linked manner; The maintenance strategy prediction module is used to call the constructed maintenance strategy database according to the predicted fault type, generate the corresponding control command signal, send it to the welding robot, and execute the command.
9. A computer-readable storage medium having a computer program stored thereon, characterized in that: When the program is executed by a processor, the steps in a welding robot predictive maintenance method as described in any one of claims 1-7 are implemented.
10. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps in the predictive maintenance method for a welding robot as described in any one of claims 1-7 are implemented.
