Robot controller based on multi-welder protocol adaptation and its cooperative control system

Abstract

This invention relates to a robot controller and its collaborative control system based on multi-welding machine protocol adaptation, specifically in the field of welding machine robot control technology. The system includes: a protocol performance evaluation unit, which calculates protocol feature fusion coefficients based on parameter mapping distortion, protocol state activity, protocol response hysteresis, and channel coupling interference; a process matching control unit, which calculates process motion matching degree by combining the protocol feature fusion coefficients, welding speed, and protocol response hysteresis, for dynamically adjusting welding and motion parameters; and a global collaborative adjustment unit, which calculates collaborative control gain based on the geometric mean of process motion matching degree of multiple welding machines, average protocol state activity, maximum channel coupling interference, and maximum process motion matching degree, for dynamically evaluating weights, adjusting control timing, and achieving multi-machine command synchronization and trajectory compensation. This invention achieves graded matching of welding machines, precise control of thermal input, and global adaptive collaboration of the system by quantifying protocol conversion performance.

Description

Technical Field

[0001] This invention relates to the field of welding machine robot control technology, specifically to a robot controller and its collaborative control system based on multi-welding machine protocol adaptation. Background Technology

[0002] With the accelerated advancement of intelligent manufacturing, multi-robot collaborative welding has become a core production process in fields such as heavy equipment, shipbuilding, and oil and gas pipelines.

[0003] Currently, mainstream multi-welding machine collaborative systems generally adopt a two-layer architecture that combines protocol conversion and independent control: the bottom layer uses a multi-protocol gateway to convert protocol formats of welding machines from different brands, while the upper layer uses a general motion control algorithm to achieve trajectory synchronization of multiple robots and the distribution of process parameters. Although this architecture solves the compatibility problem of different brands of equipment, it has shown obvious technical bottlenecks in high-precision collaborative scenarios. Specifically, firstly, existing technologies only complete the basic function of protocol format conversion and only verify whether the data can be transmitted, lacking quantitative means for addressing the implicit problems in the conversion process; secondly, the calculation of welding heat input only uses three conventional parameters: current, voltage, and welding speed, thus ignoring implicit losses such as parameter distortion, data loss, and response delay during the protocol conversion process. Moreover, when there is a delay in protocol response, the welding torch has already moved a certain distance before outputting welding current, resulting in delayed arc ignition position, mismatch between heat input and position, and uneven weld formation; in addition, the global collaborative control of existing technologies uses fixed preset parameters, lacking quantitative assessment of the system performance upper limit and worst-case interference boundary, making the control strategy prone to exceeding the actual carrying capacity of the system, and easily leading to communication collapse and synchronization inaccuracy under severe interference.

[0004] Therefore, those skilled in the art have provided robot controllers and their collaborative control systems based on multi-welding machine protocol adaptation to solve the problems mentioned in the background art. Summary of the Invention

[0005] The technical problem solved by this invention is to provide a robot controller and its collaborative control system based on multi-welding machine protocol adaptation, so as to achieve precise matching of welding tasks, improve the accuracy of heat input control, and achieve global adaptive collaboration that balances accuracy and stability.

[0006] To address the above problems, the present invention provides the following technical solution:

[0007] In a first aspect, embodiments of this application provide a robot controller based on multi-welding machine protocol adaptation, including:

[0008] The protocol performance evaluation unit calculates and determines the protocol feature fusion coefficient for each access welding machine based on the parameter mapping distortion and protocol state activity during the multi-welding machine protocol conversion process, combined with the protocol response lag and channel coupling interference. The protocol feature fusion coefficient is related to the welding machine's parameter fidelity, communication quality, response performance, and anti-interference capability.

[0009] The process matching control unit calculates and determines the process motion matching degree of each welding machine based on the power loss effect caused by the protocol feature fusion coefficient of a single welding machine, combined with the welding speed and protocol response hysteresis. The process motion matching degree is used to indicate the dynamic adjustment of welding parameters and motion parameters.

[0010] The global collaborative adjustment unit calculates and determines the collaborative control gain of the entire collaborative operation cluster based on the overall system performance reflected by the geometric mean of the process motion matching degree of multiple welding machines and the average protocol state activity, combined with the maximum channel coupling interference degree and the maximum process motion matching degree. The collaborative control gain is used to dynamically calibrate the evaluation weight and adjust the control timing of the protocol performance evaluation unit.

[0011] Further: The process of determining the protocol feature fusion coefficients includes:

[0012] The product of parameter mapping distortion and protocol state activity is taken as the effective output component of the protocol.

[0013] The product of protocol response hysteresis and channel coupling interference is used as the protocol loss component.

[0014] Divide the effective output component of the protocol by the loss component of the protocol to obtain the protocol feature fusion coefficient.

[0015] Further: The process of obtaining the effective output components of the protocol includes:

[0016] Obtain the command current and command voltage issued by the welding machine and the actual feedback current and feedback voltage of the welding machine, and calculate the relative deviation of current and relative deviation of voltage;

[0017] The parameter mapping distortion is obtained by subtracting the product of the relative current deviation and the relative voltage deviation from the reference value. The greater the parameter mapping distortion, the higher the parameter fidelity.

[0018] The protocol activity level is obtained by obtaining the total number of data frames received by the protocol per unit time and the number of valid data frames that pass the verification. The ratio of the number of valid data frames to the total number of data frames is calculated.

[0019] The effective output component of the protocol is obtained by multiplying the parameter mapping distortion and the protocol state activity.

[0020] Further: The process of obtaining the protocol loss component includes:

[0021] Obtain the protocol response lag time of the welding machine and the system's preset protocol response reference time, calculate the ratio of the actual response delay time to the protocol response reference time, and record it as the protocol response lag degree;

[0022] Obtain the single-machine communication packet loss rate and the current packet loss rate when the welding machine is running alone and when it is running in multiple machines. Calculate the difference between the single-machine communication packet loss rate and the current packet loss rate. Based on the difference, obtain the channel coupling interference degree. The larger the difference, the higher the channel coupling interference degree.

[0023] The protocol loss component is obtained by multiplying the protocol response hysteresis and the channel coupling interference.

[0024] Secondly, embodiments of this application provide a robot collaborative control system based on multi-welding machine protocol adaptation, including a memory and a processor. The memory is connected to the processor, the memory is used to store program instructions, and the processor is used to execute the program instructions. When the processor executes the program instructions, it is also used to implement the functions of the controller described in any one of the above, specifically including performing the following operations:

[0025] Based on the hierarchical matching strategy and protocol feature fusion coefficient, welding machines are divided into different performance levels, and high-precision tasks are assigned to high-level welding machines;

[0026] Based on the deviation between the process motion matching degree and the preset process window, a dynamic adjustment strategy for welding parameters of a single welding machine is determined.

[0027] Based on the changes in cooperative control gain, the timing of multi-machine instruction synchronization and robot trajectory compensation strategies are determined.

[0028] Further: The process of determining the process motion matching degree includes:

[0029] The protocol feature fusion coefficient is used as a correction coefficient to correct the product of welding thermal efficiency, feedback current, and feedback voltage, thus obtaining the effective welding power after protocol correction.

[0030] The welding speed and protocol response hysteresis of the robot are obtained, and the product of the welding speed and the protocol response hysteresis is used as the motion displacement caused by the protocol delay.

[0031] Dividing the effective welding power by the amount of motion displacement yields the process motion matching degree; the larger the amount of motion displacement, the lower the corresponding value of the process motion matching degree.

[0032] Further: The process of determining the cooperative control gain includes:

[0033] Obtain the process motion matching degree of all welding machines in operation, calculate the geometric mean of the process motion matching degree of all operating welding machines, and obtain the geometric mean of the multi-machine matching degree.

[0034] Obtain the protocol status activity of all welding machines and calculate the average protocol status activity, which reflects the overall communication quality level.

[0035] The maximum value of the coupling interference of all welding machine channels is recorded as the maximum channel coupling interference, and the worst interference boundary of the system is obtained.

[0036] The maximum value of the process motion matching degree of all welding machines is recorded as the maximum process motion matching degree, and the system performance reference benchmark is obtained.

[0037] The cooperative control gain is obtained by dividing the product of the geometric mean of the multi-machine matching degree and the average protocol state activity by the product of the maximum channel coupling interference degree and the maximum process motion matching degree.

[0038] Further: The process of determining the multi-machine instruction synchronization timing includes:

[0039] Obtain the system's preset reference synchronization period and the cooperative control gain, calculate the ratio of the reference synchronization period to the cooperative control gain, and record it as the dynamic synchronization period;

[0040] Specifically, when the cooperative control gain is greater than 1, the dynamic synchronization period is less than the reference synchronization period, which shortens the instruction issuance interval and improves the time synchronization accuracy of multiple machines; when the cooperative control gain is less than a certain preset threshold, the dynamic synchronization period is greater than the reference synchronization period, which extends the instruction issuance interval and reduces the probability of communication conflicts.

[0041] Further: The process of determining the robot trajectory compensation strategy includes:

[0042] The cooperative control gain, the robot's welding speed, and the welding machine's protocol response lag time are obtained, and the product of the three is used as the welding torch trajectory advance compensation amount for a single robot.

[0043] The greater the collaborative control gain, the greater the advance compensation of the welding torch trajectory, and the more significant the arc ignition position lag effect caused by the response delay of the compensation protocol.

[0044] The relative error of the welding torch trajectory advance compensation of multiple collaborative robots is controlled within a preset range to ensure the consistency of the superposition of heat input from multiple welding machines.

[0045] The effects of the above solution are as follows:

[0046] 1. This invention converts existing independently monitored current and voltage deviations into deep-level features of protocol conversion accuracy through product fusion, intuitively reflecting the parameter fidelity; it converts the ratio of total data frames to valid data frames calculated by conventional link layer statistics into communication data validity features to distinguish between valid and invalid data; it normalizes conventional response time parameters through a benchmark value and converts them into deep-level features of relative response speed, enabling cross-model performance comparison; and it converts the difference between single-machine packet loss rate and multi-machine packet loss rate into deep-level features of multi-machine coupling interference to accurately locate additional interference when multiple machines are operating. Furthermore, by fusing these four features, the protocol feature fusion coefficient of a single welding machine is obtained, thus achieving a leap from simply ensuring protocol connectivity to tiered protocol performance.

[0047] In addition, the system automatically classifies welding machines based on scores, avoiding quality defects caused by low-performance equipment from the source, and improving the pass rate of key welds and the overall utilization rate of equipment.

[0048] 2. This invention corrects the traditional welding power calculation results through protocol feature fusion coefficients, transforming conventional welding power calculation into a deep feature of protocol-corrected effective power, thus solving the problem of traditional calculations neglecting protocol layer losses. The product of welding speed and relative response speed is packaged into a deep feature of process-motion misalignment, intuitively reflecting the heat input and position deviation caused by protocol delay. Finally, through the fusion calculation of effective power and misalignment, the process motion matching degree, as the core closed-loop index of single-machine control, is obtained, thus taking into account the impact of protocol losses and delay misalignment.

[0049] 3. This invention uses geometric mean instead of traditional arithmetic mean to calculate the geometric mean of multi-machine matching degree, which can be more sensitive to the performance changes of individual devices, thereby detecting the overall performance degradation of the system in advance; it extracts global features such as the validity of the average data of the whole network, the maximum interference level, and the highest performance benchmark, and obtains the collaborative control gain as the dynamic adjustment center of the whole system by the fusion calculation of the actual working conditions divided by the ideal working conditions.

[0050] In addition, the system adjusts the synchronization cycle of multi-machine commands in real time according to the cooperative control gain, so as to shorten the synchronization cycle when the working conditions are good and extend the synchronization cycle when the interference is severe, thereby reducing the probability of communication conflict and ensuring the stable operation of the system. At the same time, the trajectory compensation amount of each robot is dynamically adjusted according to the gain to offset the position lag caused by protocol delay. Attached Figure Description

[0051] Figure 1 This is an architecture diagram of the robot controller based on multi-welding machine protocol adaptation of the present invention;

[0052] Figure 2 This is a flowchart illustrating the robot collaborative control system based on multi-welding machine protocol adaptation of the present invention.

[0053] Figure 3 This is a schematic diagram illustrating the welding machine grading and matching strategy in this invention, as well as the high-precision task allocation achieved by combining the available quantity with actual needs. Detailed Implementation

[0054] The technical solutions of the present invention will be clearly and completely described below with reference to the accompanying drawings.

[0055] The following description, in conjunction with the accompanying drawings, details the specific solution provided in this application for a robot controller and its collaborative control system based on multi-welding machine protocol adaptation.

[0056] Please see Figures 1 to 3 This application first provides a robot controller based on multi-welding machine protocol adaptation. Addressing the technical problems of traditional protocol conversion focusing only on data connectivity while neglecting conversion quality, and the disconnect between welding control and motion control leading to heat input and position mismatches, when multiple brands and protocols of welding machines operate collaboratively, this controller proposes a three-layer progressive control architecture comprising a protocol performance evaluation unit, a process matching control unit, and a global collaborative adjustment unit.

[0057] In one exemplary embodiment, the controller is configured as the central brain of a distributed welding system, connecting to multiple heterogeneous welding machines and the robot body via industrial Ethernet. The controller first establishes communication links with each welding machine through a built-in multi-protocol adapter module that supports mainstream welding machine protocols.

[0058] Please see Figure 1 and Figure 2 For the protocol performance evaluation unit, its core lies in quantifying the comprehensive performance of the entire protocol conversion link from instruction issuance to action execution, rather than simply verifying whether data can be transmitted. In this embodiment, the protocol performance evaluation unit comprehensively determines the protocol feature fusion coefficient of each access welding machine by real-time monitoring and calculation of four key deep features: parameter mapping distortion, protocol state activity, protocol response lag, and channel coupling interference.

[0059] The parameter mapping distortion is used to characterize the welding machine's parameter fidelity. Specifically, the controller acquires the command current and voltage it sends to the welding machine, as well as the actual feedback current and voltage after the welding machine executes the command. First, the relative deviation of the current and the relative deviation of the voltage are calculated. To avoid misjudgment due to a single parameter deviation, the two are multiplied. Only when both the current and voltage parameters are accurately converted will this product approach zero. Therefore, the quantification method for parameter mapping distortion is as follows:

[0060] ;

[0061] Among them, PMDi Let I be the parameter mapping distortion of the i-th welding machine. i Let I be the feedback current of the i-th welding machine, representing the actual current value detected and reported by the i-th welding machine's own sensor after actual output. cmd,i U represents the command current for the i-th welding machine, indicating the target welding current setpoint issued by the controller to the i-th welding machine. i U represents the feedback voltage of the i-th welding machine, which indicates the actual voltage value detected and reported by the i-th welding machine's own sensor after the actual output. cmd,i The command voltage for the i-th welding machine represents the target arc voltage setting value issued by the controller to the i-th welding machine. The calculation measures the relative error of the current, which reflects the accuracy of the current parameter conversion. The calculation measures the voltage relative error, reflecting the accuracy of voltage parameter conversion. Two errors ( Multiplication reflects the requirement that both parameters must meet the standard simultaneously: the product will only be small when both current and voltage errors are very small. Subtracting 1 from... The product result is PMD i And PMD i The closer to 1, the higher the fidelity; when it equals 1, it means that the parameter is completely undistorted.

[0062] It should be noted that if the instruction value is 0, the welding machine will be skipped or PMD will be set. i =0.

[0063] Protocol state activity is used to characterize the effective data quality of a communication link. Simply having a connected link does not guarantee data availability; a large amount of invalid data can severely interfere with control decisions. In this embodiment, the protocol state activity is quantified as follows:

[0064] ;

[0065] Among them, PSA i N represents the protocol state activity of the i-th welding machine. total,i The protocol adaptation module receives the total number of frames from the physical link layer of the i-th welding machine per unit time. As long as the link layer receives a complete data packet, regardless of whether the content is correct, it is counted in N. total,i N valid,i N represents the number of valid feedback frames that need to undergo content verification at the protocol parsing layer and fully meet the requirements. valid,i From N total,i The conditions for content validation to be performed and fully compliant include:

[0066] 1. The frame header, frame trailer, and checksum are correct;

[0067] 2. The data fields are within the legal scope defined by the protocol;

[0068] 3. It is the business status feedback (current, voltage, temperature, etc.) required by the controller.

[0069] Protocol response hysteresis reflects the system's response performance. It does not simply measure absolute delay time, but rather compares performance across different machine types and operating conditions by contrasting it with a preset baseline time. Specifically, the quantification method for protocol response hysteresis is as follows:

[0070] ;

[0071] Among them, PRL i norm For the protocol response hysteresis of the i-th welding machine, PRL i The protocol response lag time for the i-th welding machine is specifically obtained by recording the system clock timestamp T when the controller issues the command. send,i Furthermore, upon receiving the response frame from the welding machine, the receiving timestamp T is recorded. recv,i Then use T recv,i -T send,i PRL was calculated i PRL base A preset protocol response baseline time for the system (e.g., based on optimal system conditions or industry standards). PRL i norm A value greater than 1 indicates a slower response than the baseline, while a value less than 1 indicates a faster response than the baseline.

[0072] Channel coupling interference is used to characterize the degree of mutual interference between channels in a multi-machine concurrent scenario. When multiple welding machines are working simultaneously, communication conflicts and signal crosstalk increase significantly. In this embodiment, the quantification method of channel coupling interference is as follows:

[0073] CCID i =max(1,1+(P multi,i -P single,i ));

[0074] Among them, CCID i The interference level of the channel of the i-th welding machine is obtained by individually testing the packet loss rate P of the i-th welding machine during the system initialization phase. single,i Then, when multiple machines are working simultaneously, the current packet loss rate P of the i-th welding machine is statistically analyzed in real time. multi,i CCID i The closer to 1, the less interference there is between multiple machines; when it equals 1, there is no additional interference; and when it is greater than 1, there is interference.

[0075] Furthermore, as a preferred implementation, after obtaining the above four features, the process for determining the protocol feature fusion coefficient is as follows: The parameter mapping distortion, representing effective output, is multiplied by the protocol state activity to obtain the protocol effective output component; the protocol response hysteresis, representing loss, is multiplied by the channel coupling interference to obtain the protocol loss component; subsequently, the protocol feature fusion coefficient is the ratio of the effective output component to the loss component, specifically:

[0076] ;

[0077] Where: PFFC i Let be the protocol feature fusion coefficient of the i-th welding machine, for The computational logic, in essence, is a quantitative indicator of the efficiency of effective data output: on the one hand... The calculations can reflect the accuracy and quality of protocol conversion. The overall quality of protocol conversion is determined by two necessary conditions: conversion accuracy and data validity. Neither can be lacking. The product relationship reflects the weakest link effect; if PMD i =0.9 (conversion accuracy 90%) but PSA i =0.5 (half the data is invalid), then the overall quality is only 0.9 × 0.5 = 0.45, meaning the effective conversion accuracy is only 45%. Therefore, only when both are close to 1 is the numerator close to 1, indicating excellent protocol conversion quality; on the other hand, The calculation can reflect the time loss and interference level, where communication performance is determined by two factors: response speed and anti-interference capability. The product relationship reflects the additive effect of the two. =2 (response speed is twice that of the baseline) and CCID i =1.5 (interference increases packet loss rate by 50%), then the overall communication loss is 2 × 1.5 = 3, meaning the communication efficiency is only 1 / 3 of the ideal state. Therefore, PRL base A larger denominator indicates a worse communication environment and higher uncertainty in data transmission. The higher the ratio, the higher the input-output ratio of the welding machine in terms of protocol adaptation, and the stronger the protocol adaptation capability.

[0078] Please see Figure 1 and Figure 2 Based on the deviation between the process motion matching degree and the preset process window, a dynamic adjustment strategy for the welding parameters of a single welding machine is determined.

[0079] For the process matching control unit, based on the protocol performance evaluation, it further addresses the mismatch between welding parameters and motion parameters caused by protocol losses. In this embodiment, the process matching control unit calculates and determines the process motion matching degree of each welding machine based on the power loss impact caused by the protocol characteristic fusion coefficient of a single welding machine, combined with the welding speed and protocol response hysteresis.

[0080] First, in order to correct the traditional welding power calculation Ignoring the drawbacks of protocol layer losses, this embodiment introduces a Protocol Feature Fusion Coefficient (PFFC). i As a correction factor, through The effective welding power obtained after protocol modification is the welding thermal efficiency of the i-th welding machine. This effective power better represents the actual energy input of the welding machine onto the weld.

[0081] Secondly, to quantify the misalignment caused by protocol delay, this embodiment combines welding speed with protocol response hysteresis. Specifically, the amount of motion displacement D caused by protocol delay... lag,i The calculation method is as follows:

[0082] ;

[0083] Among them, V i Let D be the welding speed of the i-th welding machine. lag,i This visually reflects the distance the welding torch has moved during the protocol response delay. A larger displacement indicates a greater deviation between the welding heat input and the target position.

[0084] Ultimately, the process motion matching degree is quantified as the ratio of effective welding power to motion displacement:

[0085] ;

[0086] Where: PMMD i Let η be the process motion matching degree of the i-th welding machine. i For the welding thermal efficiency of the i-th welding machine, PFFC i As a comprehensive quantitative indicator of protocol adaptability, PFFC perfectly compensates for the shortcomings of traditional computing. i It incorporates a comprehensive evaluation across four dimensions: Protocol Conversion Accuracy (PMD), Data Validity (PSA), Response Speed ​​(PRL), and Interference Immunity (CCID). The corrected effective power is... This represents the actual output power after protocol quality attenuation. For example, if a welding machine has PFFC... i=0.8 indicates that although the commanded power is 10kW, the actual effective output power is only 8kW. The remaining 2kW power loss is caused by factors such as protocol conversion error and communication delay.

[0087] Please see Figure 1 and Figure 2 Determine the collaborative control gain of the entire collaborative operation cluster.

[0088] The global collaborative adjustment unit dynamically calibrates the control strategy of the entire system from the overall perspective of the multi-machine cluster. In this embodiment, the global collaborative adjustment unit calculates and determines the collaborative control gain of the entire collaborative operation cluster based on the overall system performance reflected by the geometric mean of the process motion matching degree of multiple welding machines and the average protocol state activity, combined with the maximum channel coupling interference degree and the maximum process motion matching degree.

[0089] Specifically, the quantization method for the cooperative control gain is as follows:

[0090] ;

[0091] In the formula: CCG is the collaborative control gain, n is the actual number of welding machines executing welding tasks after scene-specific processing, and PSA is... avg CCID is the average protocol state activity. max For maximum channel coupling interference, PMMD max For maximum process motion matching, where PSA avg CCID max and PMMD max All calculations are based on n, for example, CCID max This reflects the CCIDs in n that are executed after scene-specific processing for welding tasks. i The maximum value.

[0092] On the one hand, compared to the arithmetic mean, the geometric mean ( It is more sensitive to outliers and can effectively suppress the impact of sudden changes in the performance of individual welding machines on the overall evaluation. avg It can directly reflect the overall data quality of the entire communication network. Overall performance is determined by the level of process matching. and communication data quality PSA avg Two necessary conditions jointly determine that only when both are at a high level will the molecule be large, indicating that the overall system is operating well; on the other hand, CCID max The key boundary condition determining system stability—the reliability of multi-machine collaboration—is determined by the worst-case communication link, not the average level. (PMMD) max As a benchmark reference value for system performance, ensuring the conservatism of the control gain, therefore, the denominator This is a quantification of the system's worst-case scenario. A larger value indicates more severe disturbances and a higher performance ceiling, requiring a more conservative control strategy. The closer the ratio is to 1, the closer the overall system operating condition is to the ideal state, allowing for a more aggressive collaborative control strategy.

[0093] Please see Figures 1 to 3 Based on the aforementioned controller, this application also provides a robot collaborative control system based on multi-welding machine protocol adaptation. The system includes a memory and a processor. The memory stores program instructions, and the processor executes the program instructions to implement the functions of the aforementioned controller and specifically performs subsequent dynamic management operations.

[0094] As a preferred implementation, the system executes a hierarchical matching strategy. Specifically, before initiating a welding task, the system first evaluates the minimum number of welding machines N required for the task based on the workpiece's three-dimensional model and process requirements. req And the optimal number of welding machines required, N opt :

[0095] N req : The minimum number of welding machines required to complete this task. For example, circumferential welding of large storage tanks requires at least 4 welding machines to operate simultaneously. In this case, N req =4;

[0096] N opt The number of welding machines required to achieve maximum production efficiency. For example, the optimal configuration for welding this storage tank is 6 welding machines, in which case N opt =6.

[0097] Subsequently, the system calculates the PFFC in real time. i The value is used to automatically assign welding tasks using a hierarchical matching strategy. The specific hierarchical matching strategy includes:

[0098] PFFC i Welding machines with a rating of ≥5 are classified as Grade S (Excellent) and are assigned welding tasks such as root pass welding, cover pass welding, and critical weld welding.

[0099] 2≤PFFC i Welding machines with a rating of <5 are classified as Grade A (Good) and are assigned welding tasks such as filler welding and secondary weld welding.

[0100] 0.5≤PFFC i Welding machines with a rating of <2 are classified as Grade B (qualified) and are assigned welding tasks such as spot welding, tack welding, and welding of non-stressed parts.

[0101] PFFC i Welding machines with a strength of <0.5 are classified as Grade C (unqualified). In this case, task transfer and standby inspection are required, and the task will be automatically transferred to other Grade S or A welding machines, while triggering a fault warning.

[0102] Furthermore, the system provides real-time statistics on all current Grade B and above (PMD) levels. i The number of available welding machines N (≥0.5) avail and Grade C (PMD) i The number of faulty welding machines N (<0.5) fault And perform scene-specific processing logic:

[0103] If N avail <N req This indicates a shortage of welding machines. The appropriate strategy in this situation is:

[0104] 1. Immediately trigger a Level 3 alarm, displaying the faulty welding machine number and the cause of the abnormality on the HMI interface;

[0105] 2. Suspend all task assignments and initiate the rapid detection process for faulty welding machines. Specifically, priority should be given to checking the communication link and protocol compatibility status of Class C welding machines.

[0106] 3. If the Class C welding machine cannot be restored to a usable state within 10 seconds, the task process requirements will be automatically reduced, and N will be reassessed. req ;

[0107] If N req ≤N avail ≤N opt If the number of welding machines is just enough to meet the requirements of the scenario, then the processing strategy is as follows:

[0108] 1. Prioritize PFFC i N, the highest rated avail The welding machine is performing its task.

[0109] 2. The Class C welding machine enters standby inspection mode, and the fault diagnosis program runs automatically in the background;

[0110] If N avail >N opt This indicates a scenario where there is an oversupply of welding machines. The appropriate strategy in this case is:

[0111] 1. Select N with the highest PFFC score. opt The welding machine performs its core tasks;

[0112] 2. Of the remaining available welding machines, PFFC i The top 20% of the rated welding machines are designated as hot backup welding machines, which synchronize task data in real time and can replace faulty welding machines at any time.

[0113] 3. The remaining available welding machines enter a low-power standby state and can be assigned other auxiliary welding tasks (such as spot welding and tack welding).

[0114] 4. Class C welding machines do not require immediate inspection; fault diagnosis is performed asynchronously in the background without affecting the execution of the current task.

[0115] This tiered matching strategy avoids welding quality defects caused by low-performance equipment from the source.

[0116] As a preferred implementation, the system executes a dynamic adjustment strategy for single-machine process parameters, and the system presets the PMMD required by the process. i The reasonable range is [PMMD] min PMMD max The real-time control logic is as follows:

[0117] If PMMD i PMMD max Then the welding current I needs to be reduced. i or voltage U i Or increase welding speed V i Increase the oscillation amplitude (e.g., from ±2mm to ±3mm) and increase the oscillation frequency (e.g., from 2Hz to 3Hz). In this way, when the effective heat input is too high, it is necessary to reduce the power input or speed up the movement of the welding torch to avoid burn-through or coarse grains, and to expand the heat-affected zone to avoid local overheating.

[0118] PMMD i <PMMD min Then the welding current I needs to be increased. i or voltage U i Or reduce welding speed V i Reduce the oscillation amplitude (e.g., from ±2mm to ±1.5mm) and the oscillation frequency (e.g., from 2Hz to 1.5Hz). In this way, when the effective heat input is insufficient, it is necessary to increase the power input or slow down the movement of the welding torch to avoid incomplete fusion or porosity defects, and concentrate the heat input to ensure sufficient penetration.

[0119] PMMD i When within a reasonable range, keep the current parameters stable, so that the heat input meets the process requirements and the weld quality is stable.

[0120] Please see Figure 2 Based on the change in cooperative control gain, the timing of multi-machine instruction synchronization and robot trajectory compensation strategy are determined;

[0121] As a preferred implementation, the system executes a multi-machine instruction synchronization and trajectory compensation strategy. Based on the calculated cooperative control gain (CCG), the system dynamically adjusts the global control strategy.

[0122] First, dynamically adjust the multi-machine instruction synchronization period T. sync :

[0123] Tsync =T sync,base / CCG;

[0124] Among them, T sync,base As the reference synchronization period, when CCG>1, the system is in good working condition, T sync Less than T sync,base The controller shortens the command issuance interval, improving the time synchronization accuracy of multiple machines. When CCG is less than a certain preset threshold (e.g., 0.5), the system operating conditions are poor, and T... sync Greater than T sync,base The controller extends the command issuance interval to reduce the probability of communication conflicts and ensure stable system operation. Additionally, it can adjust the T... sync Implement minimum value limits (e.g., ≥1ms) to avoid over-compression.

[0125] Secondly, dynamic robot trajectory compensation is performed. To address the issue of delayed arc initiation position caused by protocol response latency, the system calculates the lead compensation amount for the welding torch trajectory of each robot. The specific calculation logic is as follows:

[0126] ΔS i =CCG×V i ×PRL i ;

[0127] Where, ΔS i The advance compensation amount for the welding torch trajectory of the i-th robot is used to allow the robot to move a compensation distance in advance during the protocol response period. CCG serves as a gain factor, amplifying the compensation effect and accurately offsetting the delay when the system is operating well. Simultaneously, the system controls the relative error of the compensation amount of multiple collaborative robots within a preset range (e.g., within 5%) to ensure the consistency of the superimposed heat input from multiple welding machines, thereby achieving uniform weld formation.

[0128] In summary, this application achieves a complete control closed loop from single-machine protocol quantization to multi-machine global adaptation through a three-layer architecture of protocol performance evaluation, process matching control, and global collaborative adjustment, which significantly improves the accuracy, adaptability, and stability of multi-robot collaborative welding.

[0129] While the present invention has been disclosed above, it is not limited thereto. Any person skilled in the art can make various modifications and alterations without departing from the spirit and scope of the invention; therefore, the scope of protection of the present invention should be determined by the scope defined in the claims.

Claims

1. A robot controller based on multi-welding machine protocol adaptation, characterized in that, include: The protocol performance evaluation unit calculates and determines the protocol feature fusion coefficient for each access welding machine based on the parameter mapping distortion and protocol state activity during the multi-welding machine protocol conversion process, combined with the protocol response lag and channel coupling interference. The protocol feature fusion coefficient is related to the welding machine's parameter fidelity, communication quality, response performance, and anti-interference capability. The process matching control unit calculates and determines the process motion matching degree of each welding machine based on the power loss effect caused by the protocol feature fusion coefficient of a single welding machine, combined with the welding speed and protocol response hysteresis. The process motion matching degree is used to indicate the dynamic adjustment of welding parameters and motion parameters. The global collaborative adjustment unit calculates and determines the collaborative control gain of the entire collaborative operation cluster based on the overall system performance reflected by the geometric mean of the process motion matching degree of multiple welding machines and the average protocol state activity, combined with the maximum channel coupling interference degree and the maximum process motion matching degree. The collaborative control gain is used to dynamically calibrate the evaluation weight and adjust the control timing of the protocol performance evaluation unit.

2. The robot controller based on multi-welding machine protocol adaptation according to claim 1, characterized in that, The process of determining the protocol feature fusion coefficients includes: The product of parameter mapping distortion and protocol state activity is taken as the effective output component of the protocol. The product of protocol response hysteresis and channel coupling interference is used as the protocol loss component. Divide the effective output component of the protocol by the loss component of the protocol to obtain the protocol feature fusion coefficient.

3. The robot controller based on multi-welding machine protocol adaptation according to claim 2, characterized in that, The process of obtaining the effective output components of the protocol includes: Obtain the command current and command voltage issued by the welding machine and the actual feedback current and feedback voltage of the welding machine, and calculate the relative deviation of current and relative deviation of voltage; The parameter mapping distortion is obtained by subtracting the product of the relative current deviation and the relative voltage deviation from the reference value. The greater the parameter mapping distortion, the higher the parameter fidelity. The protocol activity level is obtained by obtaining the total number of data frames received by the protocol per unit time and the number of valid data frames that pass the verification. The ratio of the number of valid data frames to the total number of data frames is calculated. The effective output component of the protocol is obtained by multiplying the parameter mapping distortion and the protocol state activity.

4. The robot controller based on multi-welding machine protocol adaptation according to claim 2, characterized in that, The process of obtaining the protocol loss component includes: Obtain the protocol response lag time of the welding machine and the system's preset protocol response reference time, calculate the ratio of the actual response delay time to the protocol response reference time, and record it as the protocol response lag degree; Obtain the single-machine communication packet loss rate and the current packet loss rate when the welding machine is running alone and when it is running in multiple machines. Calculate the difference between the single-machine communication packet loss rate and the current packet loss rate. Based on the difference, obtain the channel coupling interference degree. The larger the difference, the higher the channel coupling interference degree. The protocol loss component is obtained by multiplying the protocol response hysteresis and the channel coupling interference.

5. A robot collaborative control system based on multi-welding machine protocol adaptation, characterized in that, The system includes a memory and a processor, wherein the memory is connected to the processor, the memory is used to store program instructions, and the processor is used to execute the program instructions. When the processor executes the program instructions, it is also used to implement the functions of the controller according to any one of claims 1-4, specifically including performing the following operations: Based on the hierarchical matching strategy and protocol feature fusion coefficient, welding machines are divided into different performance levels, and high-precision tasks are assigned to high-level welding machines; Based on the deviation between the process motion matching degree and the preset process window, a dynamic adjustment strategy for welding parameters of a single welding machine is determined. Based on the changes in cooperative control gain, the timing of multi-machine instruction synchronization and robot trajectory compensation strategies are determined.

6. The robot collaborative control system based on multi-welding machine protocol adaptation according to claim 5, characterized in that, The process of determining the process motion matching degree includes: The protocol feature fusion coefficient is used as a correction coefficient to correct the product of welding thermal efficiency, feedback current, and feedback voltage, thus obtaining the effective welding power after protocol correction. The welding speed and protocol response hysteresis of the robot are obtained, and the product of the welding speed and the protocol response hysteresis is used as the motion displacement caused by the protocol delay. Dividing the effective welding power by the amount of motion displacement yields the process motion matching degree; the larger the amount of motion displacement, the lower the corresponding value of the process motion matching degree.

7. The robot collaborative control system based on multi-welding machine protocol adaptation according to claim 5, characterized in that, The process of determining the cooperative control gain includes: Obtain the process motion matching degree of all welding machines in operation, calculate the geometric mean of the process motion matching degree of all operating welding machines, and obtain the geometric mean of the multi-machine matching degree. Obtain the protocol status activity of all welding machines and calculate the average protocol status activity, which reflects the overall communication quality level. The maximum value of the coupling interference of all welding machine channels is recorded as the maximum channel coupling interference, and the worst interference boundary of the system is obtained. The maximum value of the process motion matching degree of all welding machines is recorded as the maximum process motion matching degree, and the system performance reference benchmark is obtained. The cooperative control gain is obtained by dividing the product of the geometric mean of the multi-machine matching degree and the average protocol state activity by the product of the maximum channel coupling interference degree and the maximum process motion matching degree.

8. The robot collaborative control system based on multi-welding machine protocol adaptation according to claim 5, characterized in that, The process of determining the multi-machine instruction synchronization timing includes: Obtain the system's preset reference synchronization period and the cooperative control gain, calculate the ratio of the reference synchronization period to the cooperative control gain, and record it as the dynamic synchronization period; Specifically, when the cooperative control gain is greater than 1, the dynamic synchronization period is less than the reference synchronization period, which shortens the instruction issuance interval and improves the time synchronization accuracy of multiple machines; when the cooperative control gain is less than a certain preset threshold, the dynamic synchronization period is greater than the reference synchronization period, which extends the instruction issuance interval and reduces the probability of communication conflicts.

9. The robot collaborative control system based on multi-welding machine protocol adaptation according to claim 5, characterized in that, The process of determining the robot trajectory compensation strategy includes: The cooperative control gain, the robot's welding speed, and the welding machine's protocol response lag time are obtained, and the product of the three is used as the welding torch trajectory advance compensation amount for a single robot. The greater the collaborative control gain, the greater the advance compensation of the welding torch trajectory, and the more significant the arc ignition position lag effect caused by the response delay of the compensation protocol. The relative error of the welding torch trajectory advance compensation of multiple collaborative robots is controlled within a preset range to ensure the consistency of the superposition of heat input from multiple welding machines.

Images