Six-axis stamping robot control system

Abstract

The present application relates to the field of stamping robot control technology, and specifically discloses a six-axis stamping robot control system, which includes: an operation panel; a system program management module for managing system programs; a user management module for managing user accounts and permissions; a terminal module for monitoring stamping signals, online signals, and input and output signals, and configuring I / O points; a stamping rapid generation module for setting stamping-related parameters; a driver alarm information processing module for processing alarm information and generating solutions. The six-axis stamping robot control system not only has basic operating functions, parameter control and information management functions, but also has the ability to intelligently process alarm information, which helps to provide personalized solutions and processing measures that meet actual scenarios for different fault alarm information.

Description

Technical Field

[0001] The present application relates to the technical field of stamping robot control, and more specifically, to a six-axis stamping robot control system. Background Art

[0002] With the development of industrial automation, six-axis stamping manipulators have been widely used in manufacturing industries such as automobile manufacturing, home appliance production, and metal processing. This type of manipulator usually has six degrees of freedom (axes) and can flexibly complete various complex stamping tasks, such as grasping, handling, positioning, and assembling parts, so that various stamping operations can be completed more efficiently and accurately, greatly improving production efficiency and product quality.

[0003] Although the six-axis stamping robot plays an important role in industrial production, due to the complex structure and harsh working environment of the six-axis stamping robot, it may encounter various faults and abnormal conditions during operation. These faults will not only affect the production progress, but also cause equipment damage and safety hazards. However, the traditional fault diagnosis and six-axis stamping robot control mainly rely on the experience and skills of the operator. The intelligence level of this method is low and it cannot provide personalized solutions for the fault. At the same time, for experienced technicians, the traditional method may be able to quickly determine the problem, but for novices or inexperienced operators, the fault diagnosis process will become very difficult. In addition, in actual production, the same fault type may be caused by many different reasons. For example, when an overspeed alarm occurs, it may represent a circuit fault or a motor fault in the servo during the power-on process, while it may represent excessive load inertia or zero point error of the motor encoder when the motor starts. However, the traditional fault diagnosis and six-axis stamping robot control system often cannot accurately distinguish these different causes, and can only give general fault prompts, resulting in inaccurate treatment measures.

[0004] Therefore, an optimized six-axis stamping robot control system is desired. Summary of the invention

[0005] The present application provides a six-axis stamping robot control system, which not only has basic operating functions (such as start, stop, pause, etc.) and parameter control and information management functions (such as stamping related parameter settings, user management and system program management, etc.), but also has the ability to intelligently process alarm information, which helps to provide personalized solutions and processing measures that meet actual scenarios for different fault alarm information.

[0006] In the first aspect, a six-axis stamping robot control system is provided, including: an operation panel, the operation panel including a reset button, a stop button, a start button, a pause button, an axis control button, a speed adjustment button and a rotary encoder; a system program management module for managing system programs; a user management module for managing user accounts and permissions; a terminal module for monitoring stamping signals, online signals and input and output signals, and configuring I / O points; a stamping quick generation module for setting stamping related parameters; a driver alarm information processing module for processing alarm information and generating solution suggestions.

[0007] The present application provides a six-axis stamping manipulator control system, which includes: an operation panel, the operation panel includes a reset button, a stop button, a start button, a pause button, an axis control button, a speed control button and a rotary encoder; a system program management module for managing system programs; a user management module for managing user accounts and permissions; a terminal module for monitoring stamping signals, online signals and input and output signals, and configuring I / O points; a stamping rapid generation module for setting stamping related parameters; a driver alarm information processing module for processing alarm information and generating solution suggestions. The six-axis stamping manipulator control system not only has basic operating functions (such as start, stop, pause, etc.) and parameter control and information management functions (such as stamping related parameter settings, user management and system program management, etc.), but also has the ability to intelligently process alarm information, which helps to provide personalized solutions and processing measures that meet actual scenarios for different fault alarm information. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] In order to more clearly illustrate the technical solutions of the embodiments of the present application, the drawings of the embodiments of the present application are briefly introduced below. Obviously, the drawings described below only relate to some embodiments of the present application, and are not intended to limit the present application.

[0009] Figure 1 This is a schematic block diagram of a six-axis stamping robot control system according to an embodiment of the present application.

[0010] Figure 2 This is a schematic block diagram of a driver alarm information processing module in a six-axis stamping robot control system according to an embodiment of the present application.

[0011] Figure 3 This is a schematic diagram of the data flow of the driver alarm information processing module in the six-axis stamping robot control system of an embodiment of the present application.

[0012] Figure 4 This is a schematic block diagram of an alarm label encoding unit in a six-axis stamping robot control system according to an embodiment of the present application.

[0013] Figure 5 This is a schematic block diagram of an alarm content semantic enhancement unit guided by an alarm label in a six-axis stamping robot control system in an embodiment of the present application. DETAILED DESCRIPTION

[0014] The technical solutions in the embodiments of the present application will be described clearly and completely below in conjunction with the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present application, not all embodiments. Based on the embodiments of the present application, all other embodiments obtained by ordinary technicians in this field without creative work also fall within the scope of protection of the present application.

[0015] In view of the above technical problems, in the technical solution of this application, a six-axis stamping robot control system is proposed, such as Figure 1 As shown, the six-axis stamping manipulator control system includes: an operation panel 1, which includes a reset button, a stop button, a start button, a pause button, an axis control button, a speed control button and a rotary encoder; a system program management module 2, which is used to manage the system program; a user management module 3, which is used to manage user accounts and permissions; a terminal module 4, which is used to monitor stamping signals, online signals and input and output signals, and configure I / O points; a stamping fast generation module 5, which is used to set stamping related parameters; a driver alarm information processing module 6, which is used to process the alarm information and generate solution suggestions. The six-axis stamping manipulator control system not only has basic operating functions (such as start, stop, pause, etc.) and parameter control and information management functions (such as stamping related parameter settings, user management and system program management, etc.), but also has the ability to intelligently process alarm information, which helps to provide personalized solutions and processing measures that meet actual scenarios for different fault alarm information.

[0016] Exemplarily, in the operation panel 1, when the reset button is pressed first and then the start button is pressed, the system will run to the set standby point; when the system is running automatically, the stop button is pressed to stop the system, and the stop button is pressed again to switch the system to the manual state; in the manual state, the start button is pressed to switch the system to the automatic state, and in the automatic state, the start button is pressed to start the system; when the system is running automatically, the pause button is pressed to pause the system, and when the system has an alarm, the pause button is pressed to reset the alarm. The axis control button is used to manually control the corresponding axis; the speed adjustment button is used to adjust the system running speed; the rotary encoder is used to manually adjust the servo axis operation.

[0017] For example, the operation panel 1 further includes a status bar and a switching interface. The status bar is used to display the speed, system status, system time, current program and current user information in sequence; the bottom of the switching interface displays the current system version number and the motherboard version number, and the buttons at the bottom allow the user to switch between different interfaces.

[0018] Exemplarily, the system program management module 2 has multiple functions. For example, by clicking the current program above, the user can enter the editing interface of the program file. When creating a new program, just click the file name box, enter the program name, and then click the new button on the right to create it. The operation of copying the program is also very simple. The user only needs to select the program to be copied in the list and click Copy to generate a new file name. The load function allows the user to select the program to be loaded in the list. After clicking Load, the file will become the current program. The delete function allows the user to select the program that is no longer needed in the list and click Delete to remove it. In addition, the system also provides a program search function. The user can enter specific content in the search box in the upper left corner. After clicking Search, the system will automatically filter out qualified programs. Finally, the export to USB flash drive function allows the user to export the selected program to the USB flash drive. Just select the program to be exported in the list and click Export to USB flash drive to achieve it.

[0019] For example, in the user management module 3, by clicking the operator button, the system will pop up a login dialog box, and the user can select the corresponding authority level and enter the password. As an example, the default password of the administrator can be set to 222, while the default password of the senior administrator can be set to 123.

[0020] Exemplarily, in the terminal module 4, click the monitoring terminal button at the bottom of the screen to enter the monitoring interface. This interface is divided into several parts: the punching signal monitoring interface, which can be used to monitor the punching signal and set the punching machine status when entering the automatic mode, such as no material in the film, material not punched or material punched, etc.; the online signal monitoring interface, which is used to monitor the status of the online signal; the IO self-test interface, which can observe all IO port signals and the real-time torque, peak torque, origin angle and current angle of the servo; the I / O point configuration interface, which allows users to customize the name of the expansion port and the assigned port; the membrane button customization interface, which can set the specific functions of the O1, O2, O3, O4, and O5 buttons; the servo manual control interface, by clicking the stop button on the left, you can exit from the automatic operation state and enter the manual state. At this time, click the axis control button on the right side of the hand controller to manually control the operation of the servo. Click the number button in the status bar to call up the speed adjustment ratio interface, and press the rotary encoder to trigger this interface. By pressing the up and down buttons on the left side of the hand controller, the user can adjust the manual speed, and the adjusted speed will be displayed in the status bar. Note that the maximum speed in manual mode is limited to 50. The rotary encoder allows the user to manually fine-tune the movement of the servo axis, and the servo axis selection will change accordingly with each rotation of the scale.

[0021] Exemplarily, in the stamping rapid generation module 5, the user can jump directly to the stamping process interface by clicking the stamping process button on the function interface. On this interface, the user can set the relevant parameters in the stamping process in detail. The specific settings are as follows: In the online setting, the user can configure the machine ID and location mode of this machine. For example, the number of teaching points can be switched between six-point teaching and four-point teaching. When four-point teaching is selected, there is no need to teach the upper points. These points will be automatically generated by the standby points and the material collection / discharge points. In addition, the user can also choose whether to block the fixture detection. Once this option is checked, the system will no longer detect whether the fixture is fully opened or closed. The setting of the machine ID is crucial for online operation. It needs to be set to the upper station machine ID plus 1. For example, if the machine ID of the upper station is 1, the machine ID of this machine should be set to 2.

[0022] The point teaching function allows users to teach each key point in the stamping process. Enabling the transition point option allows users to add transition points between the standby point and the pick-up / discharge point. If you need to adjust or add transition points, the user can perform the corresponding operations in the teaching interface after generating the program. The position recording function can capture the exact position of each axis at the moment and write it to the selected point, so that users can make fine adjustments by dragging and dropping. By clicking the Next / Previous button, users can switch between two different pick-up and release schemes in stand-alone mode. The setting of the smoothing coefficient determines the smoothness of the transition between points. Setting it to 0 means that smoothing is not enabled. The larger the value, the more significant the smoothing effect. However, at the starting point and the end point, the smoothing coefficient will not take effect to ensure accurate arrival. After completing the point teaching, the user can directly click the Generate button to generate the teaching program.

[0023] The peripheral settings section covers the configuration of the feeder, intermediate machine, unloader, and fixture detection timeout. For example, the user can set the unloader to detect the release signal, choose to use the first machine for double sheet detection, select the unloader code bead, enable the punch press mold safety zone detection, determine whether to stop the punch press after an online emergency stop, select the feeder's incoming signal, specify the incoming type of the intermediate machine, and the start selection after an online emergency stop.

[0024] The punch press setting function allows users to configure according to the type of punch press actually used, including but not limited to pneumatic punch press, gear punch press, hydraulic press, transition table and air relay. In addition, users can also set the duration of the punch signal and the time of each punch. The in-mold area setting focuses on the configuration of the diagonal points in the material picking mold to ensure the accuracy and safety of the material picking process.

[0025] Furthermore, the six-axis stamping robot control system also includes a teaching program module. The user can enter the teaching program module by rotating the knob switch of the hand controller to the manual state and then clicking the teaching button at the bottom of the screen. This module specifically includes the following aspects: In terms of teaching instruction entry operations, users can perform the following operations: Move up - move the currently selected line up one line; Move down - move the currently selected line down one line; Edit - click the edit button, and you can modify more parameters of the currently selected line in the pop-up window; Combine - combine the currently selected line with the previous instruction (in automatic mode, these two instructions will start executing at the same time); Decompose - split the selected combined instruction into separate instructions (in automatic mode, these instructions will be executed in sequence); Trial - trial run the currently selected instruction, press the execute button to start, and release it to stop execution. It should be noted that not all instructions support the trial run function.

[0026] In the program instruction list section, the instructions currently provided by the system include: axis action instructions, output actions, waiting, variable operations, delays, palletizing, multi-axis positioning, IF instructions, conditions, labels, and detection. Among them, axis action instructions and output actions can be taught by checking the corresponding options and clicking the Insert button. For axis action instructions, users can also click the P point button to use the pre-entered P point coordinates as the target position. It should be noted that the delay time is performed before the instruction is executed, and supports the setting of early end and early speed change. The output action also supports the setting of delay time, which is also performed before the instruction is executed, and the time of pulse output is the action time.

[0027] The setting of the wait instruction is relatively simple. The user only needs to check the corresponding option and click the insert button to complete the creation of a wait instruction. The variable operation function allows the user to use variables in the range of 0-31. When it is not checked, the operand is a constant; after checking, the operand becomes a variable, and the range is also 0-31. The supported operators include "addition", "subtraction", "multiplication", "division", "remainder" and "assignment".

[0028] The use of the IF instruction involves three parts: variable 1 (0-31, when checked, variable 1 is a system variable, that is, the current position of each axis), variable 2 (optional user variable 0-31 or system variable, that is, the current position of each axis), and the three instructions IF, ELSE, and ENDIF, which must be used in pairs.

[0029] The multi-axis positioning function allows users to complete the teaching of multi-axis positioning by checking the corresponding options and clicking the insert button. Users can also click the "Use positioning point" button to use the pre-set position as the target position. The setting of the smoothing coefficient determines the degree of smoothness when transitioning to the next point. Setting it to 0 means that smoothing is not enabled. The larger the value, the more significant the smoothing effect. However, at the starting point and the end point, in order to ensure accurate arrival, the smoothing coefficient will not take effect.

[0030] In the condition setting, the user can specify the condition number. When unconditional execution is selected, it means there is no specific condition restriction. The call subroutine function allows the user to select the subroutine to be called. If no subroutine is to be called, select "None". The jump to label function allows the user to jump to a specific label after the subroutine is executed. Specifically, selecting a label means that the main program will stop and wait for the subroutine to complete before jumping to the specified label; not selecting a label means that after the subroutine is activated, the main program will continue to run and the subroutine will stop after executing one cycle.

[0031] The label function allows users to insert labels, which can be used in conjunction with conditional instructions to achieve more complex program logic. The subroutine function allows users to enter the corresponding channel or subroutine for teaching by clicking the main program button. The delay function allows users to insert a delay instruction, and the delay time can be freely selected between 0 seconds and 600 seconds. The detection function allows users to insert a detection instruction, which is usually used in conjunction with the output instruction to achieve monitoring of specific conditions.

[0032] In terms of position point setting, users can click the "position point" button to enter the position point editing interface, and click the standby point editing button to enter the standby point editing interface to set the return position and action sequence of each axis. This helps ensure that the robot can return to the preset safe position before each operation, thereby ensuring the safety and consistency of the operation.

[0033] After the teaching program module has completed the teaching, turn the selection switch to the automatic state to enter the automatic operation interface. In this interface, you can operate the system to run automatically, stop, run in single step, run in single cycle, adjust the automatic operation speed, etc. Edit: When the system is running automatically, if you want to adjust the instruction parameters, you can cancel the follow, select the corresponding instruction, and click Edit. In the pop-up editing window, you can fine-tune the instruction parameters. Single step: The system only executes the current line of instructions. Single cycle: The system executes the entire teaching instruction once. Speed ​​control prohibited: When speed control is prohibited, the speed cannot be adjusted by pressing the up and down buttons on the left side of the operator in the automatic interface. Click Speed ​​Control Prohibit, and when the icon changes to Speed ​​Control Allowed, you can press the up and down buttons on the left side of the operator in the automatic state to adjust the speed. Click the Clear button in the pop-up interface to clear the user variables and stack counts.

[0034] Exemplarily, the driver alarm information processing module 6 is an important component of the six-axis stamping robot control system, and its main function is to process the alarm information in the system and generate corresponding solution suggestions.

[0035] In particular, in the process of processing alarm information and generating solution opinions, the technical concept of the present application is to collect alarm information and the timestamp event when the alarm information is generated, and introduce data processing and semantic understanding algorithms based on artificial intelligence and natural language processing technology in the back end to analyze the alarm information and the timestamp event, so as to capture the alarm type-timestamp event joint coding semantics and alarm content semantics, and use the alarm type-timestamp event joint coding semantics as the query basic label to perform query-style cross-modal joint coding on the key information in the alarm content to obtain the alarm content semantic enhancement representation guided by the alarm label, so as to generate solution opinions based on this. In this way, personalized processing measures suitable for actual fault scenarios can be generated based on the joint semantics between the alarm information content and the label information, realizing more intelligent driver alarm information processing and providing a basis for the intelligent control process of the six-axis stamping robot.

[0036] In one embodiment, Figure 2 and Figure 3 As shown, the driver alarm information processing module 6 includes: an alarm information acquisition unit 61, which is used to acquire alarm information; a timestamp event extraction unit 62, which is used to extract the timestamp event when the alarm information is generated; an alarm-related data extraction unit 63, which is used to extract the alarm type and alarm content from the alarm information; an alarm label encoding unit 64, which is used to jointly encode the alarm type and the timestamp event to obtain an alarm type-timestamp event joint encoding vector; an alarm content semantic encoding unit 65, which is used to perform text semantic encoding on the alarm content to obtain an alarm content semantic encoding feature map; an alarm label guided alarm content semantic reinforcement unit 66, which is used to perform cross-modal local query guided semantic joint encoding on the alarm content semantic encoding feature map and the alarm type-timestamp event joint encoding vector to obtain an alarm label guided alarm content semantic reinforcement encoding feature; a solution opinion generation unit 67, which is used to generate the solution opinion based on the alarm label guided alarm content semantic reinforcement encoding feature.

[0037] Exemplarily, the alarm information acquisition unit 61 is used to acquire alarm information. The alarm information acquisition unit in the six-axis stamping manipulator control system is responsible for collecting alarm information in the system in real time. The unit monitors the status of each sensor, driver and other key components in the system. When an abnormality is detected, the alarm information acquisition unit sends the alarm information to the central processing unit for processing.

[0038] Exemplarily, the timestamp event extraction unit 62 is used to extract the timestamp event when the alarm information is generated. It should be understood that extracting the timestamp event when the alarm information is generated can help accurately locate the fault time, which is very useful for fault diagnosis and historical data analysis. Through the timestamp, the time point when the fault occurs can be quickly located, making it easier to find the root cause of the fault. In addition, timestamp events can also be used for fault trend analysis. By recording the timestamp of each alarm, the frequency and trend of fault occurrence can be analyzed to help predict future fault risks. At the same time, timestamp events can also be used for historical data recording to provide detailed data support for subsequent maintenance and optimization. Extract timestamp events from alarm information. Timestamp events are usually included in the alarm information, recording the specific time when the alarm occurs. For example, timestamp events under overspeed alarm include: when the power is turned on, during the motor operation process, and when the motor starts.

[0039] Exemplarily, the alarm-related data extraction unit 63 is used to extract the alarm type and alarm content from the alarm information. That is, to parse the alarm information. The alarm information is usually displayed in text form, including the alarm code, alarm type and alarm content. The alarm code is a unique identifier used to quickly find the specific information of the alarm. The alarm type refers to the category of the alarm, such as overspeed, main circuit overvoltage, main circuit undervoltage, position error and overheating. The alarm content is a detailed fault description, including the specific circumstances of the fault and possible causes.

[0040] In the alarm label encoding unit 64, the alarm type and the timestamp event are jointly encoded to obtain an alarm type-timestamp event joint encoding vector. It should be understood that the alarm type and the timestamp event provide two different types of information, namely the nature of the fault and the timing of its occurrence. Using one of the information alone may not fully reflect the actual situation of the fault. Through joint encoding, these two types of information can be combined to form a more complete fault description. Some alarm types may have different meanings under different timestamp events. For example, "overspeed" may mean different problems when "power is turned on" and "motor is running". Through joint encoding, this ambiguity can be eliminated, making the generation of fault descriptions and solutions more in line with actual scenarios and more accurate.

[0041] In one embodiment, Figure 4As shown, the alarm label encoding unit 64 includes: an alarm type one-hot encoding subunit 641, which is used to one-hot encode the alarm type to obtain an alarm type one-hot encoding vector; a timestamp event one-hot encoding subunit 642, which is used to one-hot encode the timestamp event to obtain a timestamp event one-hot encoding vector; an alarm type-timestamp event joint encoding subunit 643, which is used to cascade the alarm type one-hot encoding vector and the timestamp event one-hot encoding vector to obtain the alarm type-timestamp event joint encoding vector.

[0042] Exemplarily, the alarm type one-hot encoding subunit 641 and the timestamp event one-hot encoding subunit 642 are used to one-hot encode the alarm type to obtain an alarm type one-hot encoding vector, and to one-hot encode the timestamp event to obtain a timestamp event one-hot encoding vector. It should be understood that since the alarm type and the timestamp event are categorical data, for example, the alarm type includes: overspeed, main circuit overvoltage, main circuit undervoltage, position error and overheating, etc., and the timestamp events under each alarm type contain different specific events, for example, the timestamp events under the overspeed alarm include: when the power is turned on, during the motor operation process and when the motor starts, etc. Therefore, in order to convert the two categorical data of alarm type and timestamp event into a numerical form that can be processed by the machine learning model, in the technical solution of the present application, the alarm type is further one-hot encoded to obtain an alarm type one-hot encoding vector, and the timestamp event is one-hot encoded to obtain a timestamp event one-hot encoding vector. Through the one-hot encoding method, the alarm type and the timestamp event can be converted into vector form respectively, eliminating the implicit relationship between categories, improving the interpretability of the model, and providing a solid foundation for subsequent fault handling and personalized solution generation.

[0043] Exemplarily, the alarm type-timestamp event joint encoding subunit 643 is used to cascade the alarm type unique hot encoding vector and the timestamp event unique hot encoding vector to obtain the alarm type-timestamp event joint encoding vector. It should be understood that the alarm type and the timestamp event provide two different information, namely the nature of the fault and the timing of occurrence. And some alarm types may have different meanings under different timestamp events. For example, "overspeed" may represent different problems when "power is turned on" and "motor is running". Therefore, in order to more comprehensively reflect the fault characteristics, in the technical solution of the present application, the alarm type unique hot encoding vector and the timestamp event unique hot encoding vector are further cascaded to obtain the alarm type-timestamp event joint encoding vector. It is worth mentioning that the joint encoding representation between the alarm type unique hot encoding information and the timestamp event unique hot encoding information can be used as an alarm label to eliminate this ambiguity when a subsequent personalized solution is generated, so that the generation of fault description and solution is more in line with the scene and more accurate, thereby improving the overall reliability of the system.

[0044] In the alarm content semantic encoding unit 65, the alarm content is subjected to text semantic encoding to obtain an alarm content semantic encoding feature map. It should be understood that since the alarm content is usually described in the form of natural language and contains alarm semantic information, the alarm content is further subjected to semantic encoding based on a text convolutional neural network model to extract the semantic feature information in the alarm content, so as to understand the deep semantics of the alarm content to obtain the alarm content semantic encoding feature map. In one embodiment, the alarm content semantic encoding unit is used to: perform semantic encoding on the alarm content based on a text convolutional neural network model to obtain the alarm content semantic encoding feature map.

[0045] In the alarm label guided alarm content semantic reinforcement unit 66, the alarm content semantic coding feature map and the alarm type-timestamp event joint coding vector are cross-modal local query guided semantic joint coding to obtain the alarm label guided alarm content semantic reinforcement coding feature. It should be understood that since the alarm content semantic coding feature map contains content semantic features related to the alarm, such as the cause of the equipment failure, the description of the abnormal behavior, etc., these alarm content semantics provide specific information about the alarm; and the alarm type-timestamp event joint coding vector contains the joint semantic features between the type of alarm and the timestamp event at this time, such as overspeed when the power is turned on. This information can be used as a label for the alarm at this time, which is very important for understanding the background and context of the alarm. Therefore, in the technical solution of the present application, the alarm content semantic coding feature map and the alarm type-timestamp event joint coding vector are further cross-modal local query-guided semantic joint coding to obtain alarm content semantic enhancement coding features guided by alarm labels, so as to dynamically adjust the behavior of the model according to the two different modal features of alarm information, namely, alarm content semantics and alarm labels, thereby enhancing the fusion and understanding of cross-modal joint information, so as to more fully and comprehensively understand the alarm information from multiple angles, and provide a basis for generating personalized processing measures suitable for actual fault scenarios. Furthermore, through the continuous prompt-assisted learning process, the model can better capture the complex relationship between alarm label information and alarm content semantics, so as to overcome the problem of cross-modal information isolation, and can still maintain good performance when facing unseen data. In addition, through the processing of cross-modal local query-guided semantic joint coding, the expression of key semantic features in the alarm content can be strengthened under the guidance of alarm label information, providing a basis for subsequent alarm information recognition and personalized solution generation in different scenarios.

[0046] In one embodiment, Figure 5As shown, the alarm label guided alarm content semantic enhancement unit 66 includes: an alarm content local feature decomposition subunit 661, which is used to perform local feature decomposition along the channel dimension on the alarm content semantic coding feature map to obtain a set of alarm content local feature matrices; an alarm label-alarm content cross-modal local granularity interaction encoding subunit 662, which is used to use the alarm type-timestamp event joint coding vector as a query vector, and perform cross-modal local granularity interaction encoding based on query prompt semantics on the set of alarm content local feature matrices to obtain a set of alarm label-alarm content semantic cross-modal local granularity significant interaction matrices; an alarm label-alarm content feature aggregation subunit 663, which is used to perform feature aggregation on the set of alarm label-alarm content semantic cross-modal local granularity significant interaction matrices to obtain an alarm label guided alarm content semantic enhancement coding feature map as the alarm label guided alarm content semantic enhancement coding feature.

[0047] In one embodiment, the alarm label-alarm content cross-modal local granularity interaction encoding subunit is used to: use the alarm type-timestamp event joint encoding vector as a query vector, and use each alarm content local feature matrix in the set of alarm content local feature matrices as a key matrix, input the query vector and the key matrix into a prompt learning network based on a converter structure to obtain a set of alarm label-alarm content semantic local query prompt semantic encoding vectors; and use each alarm label-alarm content semantic local query prompt semantic encoding vector in the set of the alarm type-timestamp event joint encoding vector and the alarm label-alarm content semantic local query prompt semantic encoding vector. The warning content semantic local query prompt semantic encoding vector is subjected to cross-modal mask processing based on the prompt information to obtain a set of warning label-warning content semantic cross-modal local feature mask weight matrices based on the prompt information; the set of warning label-warning content semantic cross-modal local feature mask weight matrices based on the prompt information and the set of warning content local feature matrices are calculated, and the position point multiplication between each corresponding group of warning label-warning content semantic cross-modal local feature mask weight matrices based on the prompt information and the warning content local feature matrix is ​​obtained to obtain the set of warning label-warning content semantic cross-modal local granularity significant interaction matrices.

[0048] The alarm content semantic coding feature map and the alarm type-timestamp event joint coding vector are cross-modal locally query-guided semantic joint coding to obtain the alarm content semantic enhancement coding feature guided by the alarm tag; wherein the cross-modal joint coding formula is: ;in, is the decomposition function, is the semantic coding feature map of the alarm content, and are the first, second, and third local feature matrices of the alarm content. and The local feature matrix of alarm content, and They are The first, second, and and row vectors, for coding, is the alarm type-timestamp event joint encoding vector, yes The transposed vector of is matrix multiplication, is the one-norm of the vector, for and The semantic similarity score between the alarm label and the alarm content, represents the value of the natural exponential function with the natural constant e as the base, for and The semantic similarity score between alarm label and alarm content, for The corresponding alarm label-alarm content semantic weight coefficient, yes The number of row vectors in is the semantic encoding vector of alarm label-alarm content semantic local query prompt, for The transposed vector of is the scale of the alarm label-alarm content semantic prompt association matrix, that is, the length of the matrix multiplied by the width of the matrix, yes function, yes The corresponding alarm label-alarm content semantic cross-modal local feature mask weight matrix based on prompt information, and They are and The corresponding alarm label-alarm content semantic cross-modal local feature mask weight matrix based on prompt information, For point-by-position multiplication, is the feature concatenation along the channel dimension, The alarm tag guides the semantic enhancement coding feature map of the alarm content.

[0049] In the solution opinion generation unit 67, the solution opinion is generated based on the semantic enhancement coding feature of the alarm content guided by the alarm tag. In one embodiment, the solution opinion generation unit is used to: input the semantic enhancement coding feature map of the alarm content guided by the alarm tag into the solution generator based on the large model to obtain the solution opinion. In other words, the semantic enhancement representation of the alarm content guided by the alarm tag is used to generate the solution opinion through the large model. In this way, personalized processing measures suitable for actual fault scenarios can be generated based on the joint semantics between the alarm information content and the label information, realizing more intelligent driver alarm information processing, and providing a basis for the intelligent control process of the six-axis stamping robot.

[0050] Exemplarily, the large model-based solution generator uses the Transformer architecture. The model has been trained extensively and can understand complex text information and generate high-quality solution opinions. For example, 1. When the power is turned on, the following occurs: 1. Servo internal circuit failure, 2. Motor failure. The generated solution is: When this failure occurs, first check whether the motor encoder line is in good contact. If there is a new servo or other servos in the device, you can use swapping and other methods to verify. If the fault still exists after replacing the servo, it is very likely that the motor is faulty. 2. During the operation of the motor, the following occurs: 1. The input pulse frequency is too high, the acceleration and deceleration time is too small, and the electronic gear ratio is too large. The generated solution is: check the pulse frequency, increase the acceleration and deceleration time, and check whether the electronic gear ratio of PA-12.PA-13 is reasonable. 3. When the motor starts, the following occurs: 2. Encoder failure. The generated solution is: check whether the encoder connection line is in good contact, replace the encoder line, replace the servo motor, and check whether the relevant parameter settings are reasonable, such as whether there is overshoot in PA-6 and PA-63. 4. When the motor starts, the following occurs: 1. The load inertia is too large and the motor encoder zero point is wrong. The generated solution is: Check whether the load inertia ratio is over-adjusted, such as PA-5PA-6PA-9PA-63 and other parameters. 5. When the motor starts: 2 The motor UVW phase lead is wrong, and the motor encoder wiring is wrong. The generated solution is: Check whether the UVW phase leads of the motor power line are correctly sequenced. You can swap the positions of the UVW phases one by one, check whether the motor encoder line has a wrong connection, and whether the line sequence is correct. If the problem still exists, it needs to be returned to the factory for repair. 6. When the power is turned on: 1 The input power voltage is too high and the voltage is unstable. The solution is: check whether the voltage of the AC input power supply is too high. You can use a multimeter with AC750V gear to measure and see whether the measured voltage fluctuation is normal. For example, the result of 220V-230V-235V will represent that the voltage of the external network is extremely unstable. You can turn on the P-UDC in the servo db mode for monitoring, that is, 220X1.414=311V, and the same is true for the 380V drive, that is, 380X1.414=537V. If the value of P-UDC is not within the normal range or the value of P-UDC exceeds 400V during operation (the P-UDC of the 380V drive exceeds 800V), the internal voltage of the servo will gradually increase and generate an alarm. If the voltage is taken out of the 3-phase 380V in the control cabinet, the voltage values ​​of the other two phases can be measured, and the phase with the lowest voltage value can be taken as the voltage input of the servo. 7. The motor is running: the brake circuit capacity is insufficient, the brake resistor is burned out, and the internal circuit of the servo is faulty. The resulting solution is: check whether the brake resistor is burned out, and replace it with a higher power brake resistor, such as 25 ohm 2000W-30 ohm 2000W, which is generally determined according to the on-site load inertia.If it still cannot be used normally after replacement, it is very likely that the internal resistor of the servo is faulty. It is recommended to return to the factory for repair. 8. When the power is turned on, the following problems occur: 1. The input voltage is too low 2. Temporary power outage for more than 20MS. The generated solution is: Check the voltage value of the servo input power supply. You can use the AC750V gear of the multimeter to measure whether the voltage is normal. The normal range of the mains power is about 210-225V. If the fluctuation is large, it may be caused by the influence of other equipment and the unstable voltage of the external network. It is recommended to install an isolation transformer and an AC filter. If the above problems are confirmed to be correct and the problem still exists, it may be that the internal circuit of the servo is faulty. It is recommended to return to the factory for repair. 9. During the operation of the motor, the following problems occur: 1. The power capacity is insufficient 2. The radiator is overheated. The generated solution is: Check the power of the power supply. If it is converted through a transformer, check whether the power of the transformer is sufficient. Whether the driver power is insufficient, resulting in overheating of the radiator.

[0051] Here, when the alarm content semantic encoding feature map and the alarm type-timestamp event joint encoding vector respectively represent the text convolutional association encoding semantic features of the alarm content and the one-hot encoding features of the alarm type and timestamp event, considering the differences in their cross-modal prompt auxiliaryness, the alarm label-guided alarm content semantic enhancement encoding feature map obtained by cross-modal joint encoding will have semantic aggregation distribution richness, resulting in repetition of unstructured probability mapping, affecting the semantic quality of the generated solution opinions.

[0052] Therefore, preferably, when the alarm tag-guided alarm content semantic enhancement coding feature map is input into the large model-based solution generator to obtain the solution opinion, the alarm tag-guided alarm content semantic enhancement coding feature map is optimized. Specifically, the optimization includes the following steps:

[0053] Determine the number of zero eigenvalues ​​of the semantic enhancement coding feature map guided by the alarm tag , and the number of zero eigenvalues Subtract one to get the alarm label to guide the alarm content semantic enhancement encoding zero differentiable value .

[0054] The alarm tag guides the alarm content semantic enhancement encoding zero differentiable value Multiply and divide by the number of zero eigenvalues ​​respectively Get the first alarm tag to guide the alarm content semantic reinforcement encoding field fitting value The second alarm label guides the alarm content semantics to strengthen the encoding field fitting value .

[0055] Calculate the square root of the sum of squares of all feature values ​​of the alarm tag-guided alarm content semantic enhancement coding feature map to obtain the alarm tag-guided alarm content semantic enhancement coding mode representation value: ;in, The first character of the semantic enhancement coding feature map indicating the alarm content guided by the alarm tag eigenvalues, It indicates that the alarm tag guides the semantic enhancement encoding mode representation value of the alarm content.

[0056] Using the alarm tag to guide the alarm content semantic enhancement encoding zero differentiable value The power function of each feature value of the alarm tag guided alarm content semantic enhancement encoding feature map is calculated as an exponent , and multiply the first alarm label to guide the alarm content semantic reinforcement encoding field fitting value The first alarm label is used to guide the alarm content semantic enhancement encoding intermediate feature map .

[0057] The semantic enhancement coding feature map of the alarm content guided by the alarm tag and the fitting value of the semantic enhancement coding field of the alarm content guided by the second alarm tag are combined. and the alarm tag guides the alarm content semantic enhancement encoding mode representation value Perform point multiplication to obtain the second alarm label to guide the alarm content semantic enhancement encoding intermediate feature map .

[0058] Calculate the intermediate representation vector of the semantic enhancement encoding of the alarm content guided by the first alarm tag The second alarm tag guides the alarm content semantic enhancement encoding intermediate representation vector The weighted points minus To obtain the optimized alarm label-guided alarm content semantic enhancement encoding feature map, where: and represents the weighted hyperparameter, It means point multiplication by position. Indicates adding by position point.

[0059] That is, based on the vector field fitting law of high-dimensional manifold, the isolated zero in the semantic enhancement coding feature map of the alarm content guided by the alarm tag is used as the differentiable dimension of the high-dimensional manifold, so as to fix the normal direction pointing to the distribution center near the local coordinates represented by the eigenvalue of the semantic enhancement coding feature map guided by the alarm tag, thereby avoiding the invalid repetition in the probability mapping process caused by the unstructured characteristics of the rich distribution of the semantic enhancement coding feature map guided by the alarm tag by realizing the effective alignment of the feature distribution pattern of the semantic enhancement coding feature map guided by the alarm tag to the probability density differentiable field, and improving the semantic quality of the solution opinions generated by the solution generator based on the large model. In this way, the semantic expression of the alarm content can be optimized based on the alarm tag, and personalized processing measures suitable for the actual fault scenario can be generated based on this, realizing more intelligent driver alarm information processing, and providing a basis for the intelligent control process of the six-axis stamping robot.

[0060] In another example of this application, the following requirements are also expected for the existing six-axis stamping manipulator control system: 1. Transfer the existing PLC stamping system to the dual-bus motherboard; 2. Achieve smooth screen movement and improve the work efficiency of operators. 3. The absolute value of the pulse adopts 485 communication mode so that it can flexibly respond to different production scenarios and equipment requirements. There are plans to transfer to integrated drive and control in the future, and compatibility needs to be considered. The expected goals include: 1. Successfully transfer the PLC stamping system to the dual-bus motherboard to ensure stable system operation and significantly improve performance. 2. Achieve smooth screen movement, the human-computer interaction interface is friendly, intuitive, and easy to operate, which greatly improves the work efficiency of operators. 3. Use 485 to achieve precise control of the absolute value of the pulse, and the adaptability reaches a high level in the later stage, which can flexibly adapt to a variety of different production conditions and equipment configurations. 4. The interface design of the operator and the mainboard is clear and the functions are clear. The operator meets daily operation needs, and the mainboard realizes comprehensive system monitoring and parameter settings.

[0061] In summary, according to the embodiment of the present application, the six-axis stamping robot control system is explained, which not only has basic operating functions (such as start, stop, pause, etc.) and parameter control and information management functions (such as stamping related parameter settings, user management and system program management, etc.), but also has the ability to intelligently process alarm information, which helps to provide personalized solutions and processing measures that meet actual scenarios for different fault alarm information.

[0062] 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.

[0063] 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.

[0064] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0065] In the several embodiments provided in the present application, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. For example, the device 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 through some interfaces, indirect coupling or communication connection of devices or units, which can be electrical, mechanical or other forms.

[0066] 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.

[0067] 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.

[0068] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of the present application, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product, which is stored in a storage medium and includes several instructions for enabling a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the methods described in the various embodiments of the present application.

[0069] 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 six-axis stamping robot control system, characterized in that: include: An operation panel, the operation panel comprising a reset button, a stop button, a start button, a pause button, an axis control button, a speed adjustment button and a rotary encoder; A system program management module is used to manage system programs; a user management module is used to manage user accounts and permissions; a terminal module is used to monitor stamping signals, online signals and input and output signals, and configure I / O points; a stamping fast generation module is used to set stamping related parameters; a driver alarm information processing module is used to process alarm information and generate solution suggestions; wherein, the driver alarm information processing module includes: an alarm information acquisition unit, which is used to acquire alarm information; a timestamp event extraction unit, which is used to extract the timestamp event when the alarm information is generated; an alarm related data extraction unit, which is used to extract the alarm type and alarm from the alarm information content; an alarm label encoding unit, used to jointly encode the alarm type and the timestamp event to obtain an alarm type-timestamp event joint encoding vector; an alarm content semantic encoding unit, used to perform text semantic encoding on the alarm content to obtain an alarm content semantic encoding feature map; an alarm label-guided alarm content semantic reinforcement unit, used to perform cross-modal local query-guided semantic joint encoding on the alarm content semantic encoding feature map and the alarm type-timestamp event joint encoding vector to obtain an alarm label-guided alarm content semantic reinforcement encoding feature; a solution opinion generation unit, used to generate the solution opinion based on the alarm label-guided alarm content semantic reinforcement encoding feature.

2. The six-axis stamping robot control system according to claim 1 is characterized in that: When the reset button is pressed first and then the start button is pressed, the system will run to the set standby point; when the system is running automatically, the stop button is pressed to stop the system, and the stop button is pressed again to switch the system to the manual state; In the manual state, press the start button and the system will be switched to the automatic state. In the automatic state, press the start button and the system will start running. When the system is running automatically, press the pause button and the system will pause. When the system has an alarm, press the pause button to reset the alarm.

3. The six-axis stamping robot control system according to claim 1 is characterized in that: The alarm label encoding unit includes: an alarm type one-hot encoding subunit, which is used to one-hot encode the alarm type to obtain an alarm type one-hot encoding vector; a timestamp event one-hot encoding subunit, which is used to one-hot encode the timestamp event to obtain a timestamp event one-hot encoding vector; an alarm type-timestamp event joint encoding subunit, which is used to cascade the alarm type one-hot encoding vector and the timestamp event one-hot encoding vector to obtain the alarm type-timestamp event joint encoding vector.

4. The six-axis stamping robot control system according to claim 3 is characterized in that: The alarm content semantic encoding unit is used to: perform semantic encoding on the alarm content based on a text convolutional neural network model to obtain the alarm content semantic encoding feature map.

5. The six-axis stamping robot control system according to claim 4 is characterized in that: The alarm label guided alarm content semantic enhancement unit includes: an alarm content local feature decomposition subunit, which is used to perform local feature decomposition along the channel dimension on the alarm content semantic coding feature map to obtain a set of alarm content local feature matrices; an alarm label-alarm content cross-modal local granularity interaction encoding subunit, which is used to use the alarm type-timestamp event joint coding vector as a query vector, and perform cross-modal local granularity interaction encoding based on query prompt semantics on the set of alarm content local feature matrices to obtain a set of alarm label-alarm content semantic cross-modal local granularity significant interaction matrices; an alarm label-alarm content feature aggregation subunit, which is used to perform feature aggregation on the set of alarm label-alarm content semantic cross-modal local granularity significant interaction matrices to obtain an alarm label guided alarm content semantic enhancement coding feature map as the alarm label guided alarm content semantic enhancement coding feature.

6. The six-axis stamping robot control system according to claim 5 is characterized in that: The alarm label-alarm content cross-modal local granularity interactive encoding subunit is used to: use the alarm type-timestamp event joint encoding vector as a query vector, and use each alarm content local feature matrix in the set of alarm content local feature matrices as a key matrix, input the query vector and the key matrix into a prompt learning network based on a converter structure to obtain a set of alarm label-alarm content semantic local query prompt semantic encoding vectors; input each alarm label-alarm content semantic local query prompt semantic encoding vector in the set of the alarm type-timestamp event joint encoding vector and the alarm label-alarm content semantic local query prompt semantic encoding vector. The local query prompt semantic encoding vector is subjected to cross-modal mask processing based on the prompt information to obtain a set of alarm label-alarm content semantic cross-modal local feature mask weight matrices based on the prompt information; the point-by-point multiplication between each corresponding group of the alarm label-alarm content semantic cross-modal local feature mask weight matrices based on the prompt information and the alarm content local feature matrices in the set of the alarm label-alarm content semantic cross-modal local feature mask weight matrices based on the prompt information and the set of the alarm content local feature matrices is calculated to obtain the set of the alarm label-alarm content semantic cross-modal local granularity significant interaction matrices.

7. The six-axis stamping robot control system according to claim 6 is characterized in that: The solution opinion generating unit is used to: input the alarm tag-guided alarm content semantic enhancement coding feature map into the large model-based solution generator to obtain the solution opinion.

Images