Automatic tracking multi-pass welding control method and system for submerged arc welding of structural pipeline

Abstract

The invention discloses an automatic tracking multi-pass welding control method and system for structural pipeline submerged arc welding, and relates to the technical field of intelligent control, the method comprises the following steps: dynamically optimizing a welding path and parameters according to a real-time deviation value of a welding seam position fed back by an execution layer subsystem and actual current and voltage data uploaded by a submerged arc welding machine subsystem; a correction instruction is generated and issued to an execution layer through the industrial Ethernet, and a servo motor is driven to adjust the posture of a welding gun; and on the basis of the welding state data after the correction instruction is executed, abnormal event interruption signals are monitored in real time, if it is detected that arcing failure or tracking failure or welding flux recovery is abnormal, the welding process is paused according to preset emergency processing logic, the subsystem state is reset or the technological parameter threshold value is adjusted, and abnormal data are uploaded to an MES system. According to the invention, high-precision, high-stability and full-process intelligent control of multi-pass welding of submerged-arc welding of the structural pipeline is realized.

Description

Technical Field

[0001] This invention belongs to the field of intelligent control technology, and in particular relates to an automatic tracking multi-pass welding control method and system for submerged arc welding of structural pipelines. Background Technology

[0002] Traditional submerged arc welding (SAW) multi-pass welding technology for structural pipelines relies heavily on manual operation, which has several technical limitations. Operators need to manually adjust the welding torch position, welding current and voltage, and wire feed speed, resulting in high labor intensity and low efficiency. Especially in complex pipeline welding scenarios, human error can easily lead to unstable weld quality. In addition, traditional methods lack real-time weld tracking and dynamic parameter adjustment capabilities, failing to effectively compensate for interference factors such as thermal deformation and assembly errors during the welding process, resulting in poor weld consistency and high rework rates.

[0003] With the advancement of industrial automation technology, some research has attempted to introduce weld seam tracking systems and automated welding equipment to improve efficiency. For example, weld seam tracking technology based on photoelectric sensors or visual inspection can initially achieve welding torch position correction. However, existing systems mostly use a single control module and lack a hierarchical collaborative architecture, resulting in limited data processing capabilities, high response latency, and difficulty in meeting the real-time dynamic optimization requirements of multi-pass welding. At the same time, existing control algorithms (such as PID control) are prone to getting trapped in local optima when facing nonlinear, multivariable coupled welding parameter optimization problems, failing to achieve global optimization and limiting welding accuracy and process stability.

[0004] Therefore, there is an urgent need to design an automatic tracking multi-pass welding control method and system for submerged arc welding of structural pipelines to solve the problems mentioned above. Summary of the Invention

[0005] To address the technical problems mentioned in the background section, where existing systems often employ a single control module and lack a hierarchical collaborative architecture, resulting in limited data processing capabilities, high response latency, and difficulty in meeting the real-time dynamic optimization requirements of multi-pass welding, this paper provides an automatic tracking multi-pass welding control method and system for submerged arc welding of structural pipelines. This method aims to solve the problems of high precision, high stability, and intelligent control throughout the entire process of multi-pass submerged arc welding of structural pipelines.

[0006] To achieve the above objectives, the specific technical solution of the automatic tracking multi-pass welding control method for submerged arc welding of structural pipelines of the present invention is as follows: An automatic tracking multi-pass welding control method for submerged arc welding of structural pipelines, applied to the main control layer PLC, mainly includes the following steps: S1. Based on the pre-set process database or MES system work order parameters, generate multiple initial welding process instructions and send them to the execution layer subsystem via industrial Ethernet to trigger the laser weld seam tracking subsystem and submerged arc welding machine subsystem to execute the initial welding process. S2. Based on the real-time deviation of the weld position fed back by the execution layer subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, dynamically optimize the welding path and parameters, generate correction instructions, and send them to the execution layer via industrial Ethernet to drive the servo motor to adjust the welding torch posture. S3. Based on the welding status data after the execution of the correction command in S2, monitor the abnormal event interruption signal in real time. If arc ignition failure, tracking failure or flux recovery abnormality is detected, suspend the welding process, reset the subsystem status or adjust the process parameter threshold according to the preset emergency handling logic, and upload the abnormal data to the MES system. S4 integrates the welding process data generated from S1 to S3, including actual current and voltage, tracking deviation records and video monitoring images, and feeds them back to the MES system via industrial Ethernet to complete process parameter matching, quality traceability and closed-loop control command reception, forming a full-process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closed loop.

[0007] Furthermore, the initial process instructions for multi-pass welding include the welding current, voltage, and vertical and horizontal position coordinates for each weld pass.

[0008] Furthermore, S1 also includes: Parse the parameters of the pre-set process database or the work order of the MES system, extract the welding process requirements, and generate the initial process instructions containing the welding current, voltage, and vertical and horizontal position coordinates of each weld pass. The initial process command is sent to the execution layer subsystem via industrial Ethernet, triggering the laser weld seam tracking subsystem to start real-time weld seam position detection and simultaneously driving the submerged arc welding machine subsystem to execute welding actions according to the initial parameters. Based on the initial process command issuance result, the system receives the initial weld seam positioning data fed back by the laser weld seam tracking subsystem, verifies the matching between the initial position of the welding torch and the target coordinates, and generates a calibrated command to enable the submerged arc welding machine subsystem to perform the arc ignition operation according to the calibrated command.

[0009] Furthermore, the correction command includes the corrected welding torch target position coordinates and current and voltage settings; S2 also includes: constructing a dynamic optimization objective function based on the real-time deviation of the weld position fed back by the laser weld seam tracking subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem; The objective function is iteratively optimized using a genetic algorithm. The final candidate parameter set is generated through selection, crossover and mutation operations, including the welding torch target position coordinate correction, current and voltage adjustment and wire feed speed compensation value. Based on the final candidate parameter set, a correction instruction is generated and sent to the execution layer via industrial Ethernet to drive the servo motor to adjust the horizontal or vertical posture of the welding torch, and simultaneously adjust the wire feeding speed and welding voltage of the submerged arc welding machine.

[0010] Furthermore, based on the real-time deviation of the weld position, the lateral and longitudinal tracking errors of the welding torch are calculated, and the welding parameter error is determined based on the difference between the actual current and voltage data of the submerged arc welding machine subsystem and the set value of the initial process command. Combining the preset welding path smoothness constraints, the tracking error, welding parameter error, and path smoothness are used as optimization objectives, and the weight coefficients of each optimization term in the objective function are defined. Based on the range of initial process command parameters generated in S1, set the feasible domain constraints for welding torch position correction, current and voltage adjustment and wire feed speed compensation. Based on the aforementioned optimization objective, weight coefficients, and feasible region constraints, a dynamic optimization objective function is constructed.

[0011] Furthermore, based on the set feasible region constraints, the population of the genetic algorithm is initialized. The population consists of multiple candidate parameter individuals, each of which contains a random combination of welding torch position correction, current and voltage adjustment, and wire feed speed compensation value. Using the constructed dynamic optimization objective function, the fitness value of each candidate parameter is calculated. The fitness value is used to characterize the comprehensive optimization effect of the corresponding parameter combination on tracking error, welding parameter error and path smoothness. Based on the fitness value, a roulette wheel selection operation is performed to select the corresponding individuals as parents, and new offspring individuals are generated through a single-point crossover operation. Iteratively calculate the fitness value and perform single-point crossover operations until the preset number of iterations is met, and output the final candidate parameter set corresponding to the current population.

[0012] Furthermore, S3 also includes: Based on the welding status data after the execution of the correction command in S2, the actual values ​​of welding current and voltage of the submerged arc welding machine subsystem, the real-time deviation of the laser weld seam tracking subsystem, and the operating status signal of the flux recovery device are collected in real time. The actual values ​​of welding current and voltage are compared with the set threshold of the correction command. If the number of consecutive deviations exceeds the preset value, it is determined that the arc ignition has failed or the tracking has failed. At the same time, the deviation between the flux recovery flow rate and the preset value is monitored. If it exceeds the allowable range, it is determined that the flux recovery is abnormal. Based on the detected anomaly type, trigger the preset emergency handling logic; If arc initiation fails or tracking fails, the welding process is paused and the welding torch is reset to a safe position. At the same time, the current and voltage thresholds are adjusted to the range of the initial process instructions. If the flux recovery is abnormal, start the backup recovery device and reduce the wire feed speed until the flow rate is restored to the allowable range; The abnormal event type, occurrence time and handling result are uploaded to the MES system via industrial Ethernet, triggering the MES to issue process parameter re-optimization instructions or manual intervention requests. If the welding process is restarted after handling the anomaly, the dynamic optimization of step S2 will be re-executed based on the updated process parameters until the welding state returns to stability.

[0013] An automatic tracking multi-pass welding control system for submerged arc welding of structural pipes includes: The generation module is used to generate multiple initial welding process instructions based on the pre-set process database or MES system work order parameters, and send them to the execution layer subsystem via industrial Ethernet to trigger the laser weld seam tracking subsystem and submerged arc welding machine subsystem to execute the initial welding process. The correction module is used to dynamically optimize the welding path and parameters based on the real-time deviation of the weld position fed back by the execution layer subsystem and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, generate correction instructions, and send them to the execution layer via industrial Ethernet to drive the servo motor to adjust the welding torch posture. The detection module is used to monitor abnormal event interruption signals in real time based on the welding status data after the correction command is executed. If arc ignition failure, tracking failure or flux recovery abnormality is detected, the welding process is suspended, the subsystem status is reset or the process parameter threshold is adjusted according to the preset emergency handling logic, and the abnormal data is uploaded to the MES system. The processing module integrates the generated welding process data, including actual current and voltage, tracking deviation records, and video monitoring images, and feeds them back to the MES system via industrial Ethernet to complete process parameter matching, quality traceability, and closed-loop control command reception, forming a complete process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closure.

[0014] The automatic tracking multi-pass welding control method for submerged arc welding of structural pipelines of the present invention has the following advantages: By monitoring welding current, voltage, tracking deviation, and flux recovery status in real time, and combining this with preset emergency handling logic, the system automatically triggers pause, reset, or parameter adjustment actions, reducing downtime and manual intervention costs. Simultaneously, abnormal data and process parameters are fed back in a closed loop to the MES system, supporting remote process optimization and quality traceability, forming an intelligent decision-making chain. The flux control subsystem of this application adopts a negative pressure recovery and recycling design, reducing flux waste and avoiding environmental pollution; dynamic parameter optimization further reduces energy consumption, improving the economy and environmental friendliness of the welding process. Attached Figure Description

[0015] Figure 1This is a flowchart illustrating the automatic tracking multi-pass welding control method for submerged arc welding of pipes according to the present invention. Figure 2 This is a flowchart illustrating the automatic tracking multi-pass welding control system for submerged arc welding of pipes according to the present invention. Detailed Implementation

[0016] To make the objectives, technical solutions, and advantages of the embodiments of the present invention clearer, the technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] Those skilled in the art will understand that although some embodiments herein include certain features included in other embodiments but not others, combinations of features from different embodiments are intended to be within the scope of the invention and form different embodiments. For example, in the claims, any of the claimed embodiments can be used in any combination.

[0018] The following is a reference to the appendix. Figure 1 To be continued Figure 2 The present invention describes an automatic tracking multi-pass welding control method for submerged arc welding of structural pipes.

[0019] like Figure 1 As shown, the automatic tracking multi-pass welding control method for submerged arc welding of structural pipelines in this invention mainly includes the following steps: S1. Based on the pre-set process database or MES system work order parameters, generate multiple initial welding process instructions and send them to the execution layer subsystem via industrial Ethernet to trigger the laser weld seam tracking subsystem and submerged arc welding machine subsystem to execute the initial welding process. S2. Based on the real-time deviation of the weld position fed back by the execution layer subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, dynamically optimize the welding path and parameters, generate correction instructions, and send them to the execution layer via industrial Ethernet to drive the servo motor to adjust the welding torch posture. S3. Based on the welding status data after the execution of the correction command in S2, monitor the abnormal event interruption signal in real time. If arc ignition failure, tracking failure or flux recovery abnormality is detected, suspend the welding process, reset the subsystem status or adjust the process parameter threshold according to the preset emergency handling logic, and upload the abnormal data to the MES system. S4 integrates the welding process data generated from S1 to S3, including actual current and voltage, tracking deviation records and video monitoring images, and feeds them back to the MES system via industrial Ethernet to complete process parameter matching, quality traceability and closed-loop control command reception, forming a full-process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closed loop; the initial process commands for multiple welding passes include welding current, voltage, vertical and horizontal position coordinates of each weld pass.

[0020] In this embodiment of the invention, the seamless integration of a pre-set process database with the MES system enables the automated generation and distribution of welding process parameters, significantly reducing the need for manual intervention and greatly improving the efficiency of multi-pass welding operations. This is particularly suitable for mass production scenarios of complex pipeline structures. A multi-objective collaborative optimization strategy based on a genetic algorithm can globally optimize the welding torch position correction, current and voltage adjustment, and wire feed speed compensation values. This overcomes the shortcomings of traditional PID control, which is prone to getting trapped in local optima, effectively reducing weld tracking deviation and welding parameter errors, and improving the smoothness and consistency of the welding path.

[0021] Leveraging the high-speed communication capabilities of Industrial Ethernet (Profinet / ModbusTCP), real-time interaction of commands and data between the main control layer and the execution layer is achieved. This ensures coordinated response of subsystems such as laser tracking and welding machine control, reduces control errors caused by communication delays, and guarantees a stable and reliable welding process. By monitoring welding current, voltage, tracking deviation, and flux recovery status in real time, and combining this with preset emergency handling logic, pause, reset, or parameter adjustment actions are automatically triggered, reducing downtime and manual intervention costs.

[0022] Meanwhile, abnormal data and process parameters are fed back to the MES system in a closed loop, supporting remote process optimization and quality traceability, forming an intelligent decision-making chain. The flux control subsystem adopts a negative pressure recovery and recycling design to reduce flux waste and avoid environmental pollution; dynamic parameter optimization further reduces energy consumption and improves the economy and environmental friendliness of the welding process.

[0023] In a preferred embodiment, step S1: Based on the preset process database or MES system work order parameters, generate multiple initial welding process instructions and send them to the execution layer subsystem via industrial Ethernet to trigger the laser weld seam tracking subsystem and submerged arc welding machine subsystem to execute the initial welding process, including: Parse the parameters of the pre-set process database or the work order of the MES system, extract the welding process requirements, and generate the initial process instructions containing the welding current, voltage, and vertical and horizontal position coordinates of each weld pass. The initial process command is sent to the execution layer subsystem via industrial Ethernet, triggering the laser weld seam tracking subsystem to start real-time weld seam position detection and simultaneously driving the submerged arc welding machine subsystem to execute welding actions according to the initial parameters. Based on the initial process command issuance result, the system receives the initial weld seam positioning data fed back by the laser weld seam tracking subsystem, verifies the matching between the initial position of the welding torch and the target coordinates, and generates a calibrated command to enable the submerged arc welding machine subsystem to perform the arc ignition operation according to the calibrated command.

[0024] In this embodiment of the invention, structured data such as pipe geometry parameters (pipe diameter, wall thickness, material), welding standards (such as AWS / DIN), and weld types (butt joint / corner joint) are obtained from the MES system. Historical similar work orders are matched from the process database to extract typical parameter ranges (such as current 280-320A, voltage 28-32V). Based on the bevel size (e.g., a 60° V-groove angle) and penetration depth requirements, the system automatically calculates the number of weld layers (e.g., 4 weld layers for a 12mm wall thickness) and assigns parameters to each layer: the first layer uses a low current (e.g., 280A) to ensure fusion, while the filler layer uses a higher current (e.g., 320A) to improve efficiency. Based on the 3D model of the pipeline (e.g., imported from CAD drawings), the system calculates the horizontal offset (e.g., increasing by 2mm per layer) and vertical height (e.g., decreasing by 1.5mm per layer) for each weld layer. It generates a continuous sequence of coordinate points (e.g., X = [100, 102, 104...], Y = [50, 48.5, 47...]) to form a spiral welding path. This invention replaces manual experience-based programming, reduces parameter setting time (from hours to minutes), standardizes process parameters, and reduces welding defects caused by human error (e.g., incomplete fusion).

[0025] The process instructions are converted into Profinet data packets (e.g., PDU length 240 bytes), and parameters are mapped according to PLC addressing rules (e.g., DB1.DBW0 stores current values). The laser tracking system trigger delay is set (e.g., the laser is activated 200ms after the welding machine starts), ensuring that the light pattern is projected 10mm in front of the molten pool. The clock deviation of each subsystem is calibrated (accuracy ≤100μs) through a timestamp synchronization mechanism (e.g., IEEE1588). The laser system acquires weld seam images, and the deviation between the actual bevel center and the theoretical coordinates is calculated through a template matching algorithm (e.g., ΔX = 1.2mm, ΔY = 0.8mm). If the deviation exceeds the threshold (e.g., ±0.5mm), a compensation value is generated (e.g., X coordinate +1.2mm, Y coordinate +0.8mm). The multi-system collaborative error of this invention is ≤0.3mm, ensuring the initial position accuracy of the welding torch, detecting abnormalities such as tooling and fixture offsets in advance, and avoiding batch welding failures.

[0026] Based on the deviation value of laser feedback, the adjustment amount of the servo motor is calculated (e.g., a lead screw transmission ratio of 1:10, ΔX = 1.2mm corresponds to a 12° motor rotation), and a new coordinate command is generated (e.g., original X = 100mm → corrected X = 101.2mm). According to the compensated position, the arc-starting current is dynamically adjusted (e.g., when the position deviation is > 1mm, the arc-starting current is increased by 5%), and the arc-starting time gradient is set (e.g., 0.8s for thick-walled tubes and 0.5s for thin-walled tubes). The wire feeding is started 0.3s before arc-starting, and the wire feeding speed is accelerated from 0 to the set value (e.g., 5m / min) to avoid wire sticking.

[0027] In a preferred embodiment, the correction command includes the corrected welding torch target position coordinates and current and voltage setting values; In S2, based on the real-time weld position deviation fed back by the execution layer subsystem in step S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, the welding path and parameters are dynamically optimized, a correction command is generated, and it is sent to the execution layer via industrial Ethernet to drive the servo motor to adjust the welding torch posture, including: Based on the real-time deviation of the weld position fed back by the laser weld tracking subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, a dynamic optimization objective function is constructed. The objective function is iteratively optimized using a genetic algorithm. The final candidate parameter set is generated through selection, crossover and mutation operations, including the welding torch target position coordinate correction, current and voltage adjustment and wire feed speed compensation value. Based on the final candidate parameter set, a correction instruction is generated and sent to the execution layer via industrial Ethernet to drive the servo motor to adjust the horizontal or vertical posture of the welding torch, and simultaneously adjust the wire feeding speed and welding voltage of the submerged arc welding machine.

[0028] Specifically, the weld deviation data fed back by the laser tracking system (such as X-axis deviation of 0.8mm and Y-axis deviation of -0.5mm) is subjected to Kalman filtering to eliminate high-frequency noise (filtering window of 50ms). The current and voltage data uploaded by the submerged arc welding machine (such as actual current of 305A and set value of 300A) are normalized (normalization range of ±10%). The position deviation weight is set to 0.5, the current stability weight to 0.3, and the voltage stability weight to 0.2 (which can be adjusted according to the weld type).

[0029] Calculate the overall deviation value: Total deviation = 0.5 × position deviation + 0.3 × current deviation + 0.2 × voltage deviation.

[0030] Constraint definition: Parameter boundary constraints: current adjustment range ±15A, voltage adjustment range ±2V, position correction ±2mm.

[0031] Process constraints: Current-voltage matching relationship (e.g., for every 10A increase in current, the voltage needs to be increased by 0.5V).

[0032] Transforming multi-dimensional process parameters into a unified optimization objective avoids overall performance degradation caused by adjusting a single parameter, and adaptively adjusts weight allocation (such as the weight of positional accuracy for fillet weld reinforcement) to improve the adaptability of complex welds.

[0033] Compared to PID control, genetic algorithms can escape local optima (such as avoiding mistracking at weld seam edges), improve global optimization rate by 30%, and double dynamic response speed (from 200ms to 100ms), adapting to the requirements of high-speed welding (≥500mm / min).

[0034] Convert the optimized parameters into a device-recognizable format: Position correction amount → number of servo motor pulses (e.g., X correction 0.6mm → corresponding to 600 pulses), current and voltage adjustment amount → analog output value (e.g., +8A → add 4mA current signal), set command issuance delay: first send the position adjustment command (at time T0), 10ms later send the current and voltage adjustment command (T0+10ms), receive motor encoder feedback (e.g., actual movement 0.58mm, error 0.02mm), compare the adjusted current and voltage values ​​(e.g., adjusted current 312A, error 0.3% compared to the target value 313A).

[0035] The present invention achieves a position adjustment accuracy of ±0.05mm and a current and voltage control accuracy of ±1%. Through timing decoupling control, it avoids mutual interference during the adjustment of multiple parameters (such as changes in arc length caused by position adjustment).

[0036] As a preferred embodiment, based on the real-time weld position deviation fed back by the laser weld tracking subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, a dynamic optimization objective function is constructed, including: Based on the real-time deviation of the weld position, the lateral and longitudinal tracking errors of the welding torch are calculated, and the welding parameter error is determined based on the difference between the actual current and voltage data of the submerged arc welding machine subsystem and the set value of the initial process command. Combining the preset welding path smoothness constraints, the tracking error, welding parameter error, and path smoothness are used as optimization objectives, and the weight coefficients of each optimization term in the objective function are defined. Based on the range of initial process command parameters generated in S1, set the feasible domain constraints for welding torch position correction, current and voltage adjustment and wire feed speed compensation. Based on the aforementioned optimization objective, weight coefficients, and feasible region constraints, a dynamic optimization objective function is constructed.

[0037] In this embodiment of the invention, a real-time deviation vector (ΔX, ΔY) is obtained from the laser tracking system, for example: ΔX = Actual weld center X coordinate - Theoretical weld center X coordinate (unit: mm).

[0038] Calculate the lateral error component (left-right offset) and the longitudinal error component (height-low offset), for example: Lateral error = ΔX × cos(welding direction angle); Longitudinal error = ΔY × sin(welding direction angle).

[0039] Quantification of welding parameter errors: Current error = |Actual current - Set current| / Set current × 100%, for example: Actual current 310A, set current 300A → Current error = |310-300| / 300×100%≈3.3%.

[0040] Voltage error is calculated similarly, for example: Actual voltage 30V, set voltage 29V → voltage error = |30-29| / 29×100%≈3.4%.

[0041] This invention transforms multidimensional deviation data into quantifiable error indicators, providing a unified metric for multi-objective optimization. Distinguishing between lateral and longitudinal errors helps to adjust the welding torch posture in a targeted manner (e.g., adjusting the slider position first for lateral errors, and adjusting the robotic arm angle first for longitudinal errors).

[0042] Weights are assigned based on weld type, for example: Butt weld: tracking error weight 0.6, current error 0.2, voltage error 0.1, path smoothness 0.1.

[0043] Fillet weld: tracking error weight 0.7, current error 0.15, voltage error 0.05, path smoothness 0.1.

[0044] Dynamic adjustment mechanism: When the detected melt width fluctuation exceeds the threshold, the current error weight is temporarily increased to 0.3.

[0045] Path smoothness constraint: Calculate the gradient of the correction amount for adjacent welding torch positions: Gradient = |Current correction - Previous correction| / Time interval, for example: Within 100ms, the X-direction correction changes from 0.5mm to 0.8mm → gradient = |0.8-0.5| / 0.1 = 3mm / s.

[0046] Set a maximum allowable gradient (e.g., 5 mm / s), and add a penalty term if the gradient is exceeded.

[0047] The weighted adaptive mechanism of this invention ensures the optimal control strategy for different weld types (such as focusing more on positional accuracy for fillet welds), and the path smoothness constraint avoids frequent and large-scale adjustments, reducing spatter and poor forming caused by mechanical vibration.

[0048] Position correction constraint: Lateral correction range: ±2mm (limited by the maximum stroke of the slider).

[0049] Longitudinal correction range: ±1.5mm (limited by the joint angle of the robotic arm).

[0050] Current and voltage regulation constraints: Current adjustment range: ±15A (exceeding this range may result in insufficient penetration or burn-through).

[0051] Voltage adjustment range: ±2V (to maintain arc stability).

[0052] Wire feed speed compensation constraints: Compensation range: ±0.5m / min (the wire feeding speed and current must maintain a linear relationship: approximately 1A corresponds to 0.03m / min).

[0053] Parameter coupling constraints: Current-voltage linkage: For every 10A increase in current, the voltage must increase by 0.5V simultaneously.

[0054] Wire feeding speed and current matching: Wire feeding speed = reference speed + current adjustment amount × 0.03m / min / A.

[0055] The physical constraints of this invention prevent equipment damage or welding defects (such as burn-through caused by excessive current) due to parameter overruns. The parameter coupling constraints ensure that the coordinated changes of process parameters conform to the physical laws of welding, thereby improving control stability.

[0056] Objective function synthesis and construction process: Weighted error summation: Total error = Tracking error weight × Tracking error + Current error weight × Current error + Voltage error weight × Voltage error.

[0057] Smoothness penalty: If the path gradient exceeds the threshold (e.g., 5 mm / s), then the total error = gradient excess × penalty coefficient (e.g., 2).

[0058] Penalties for violating the restrictions: If any parameter exceeds the feasible region, the total error = violation amount × high penalty coefficient (e.g., 10).

[0059] This invention transforms multi-objective optimization into a single-valued optimization problem, which facilitates efficient solution by genetic algorithms. The penalty mechanism ensures that the search space is limited to the physically feasible range, thereby improving the effectiveness of the solution.

[0060] The objective function is iteratively optimized using a genetic algorithm. A final candidate parameter set is generated through selection, crossover, and mutation operations. This set includes the welding torch target position coordinate correction, current and voltage adjustment, and wire feed speed compensation values. Based on the set feasible region constraints, the population of the genetic algorithm is initialized. The population consists of multiple candidate parameter individuals, each of which contains a random combination of welding torch position correction, current and voltage adjustment and wire feed speed compensation value. Using the constructed dynamic optimization objective function, the fitness value of each candidate parameter is calculated. The fitness value is used to characterize the comprehensive optimization effect of the corresponding parameter combination on tracking error, welding parameter error and path smoothness. Based on the fitness value, a roulette wheel selection operation is performed to select the corresponding individuals as parents, and new offspring individuals are generated through a single-point crossover operation. Iteratively calculate the fitness value and perform single-point crossover operations until the preset number of iterations is met, and output the final candidate parameter set corresponding to the current population.

[0061] In this embodiment of the invention, continuous parameters (such as X correction amount [-2, 2] mm) are discretized into binary codes (such as 10-bit binary representation with an accuracy of 0.004 mm). Each individual contains 5 parameters: X / Y coordinate correction amount, current adjustment amount, voltage adjustment amount, and wire feeding speed compensation value, for a total of 50-bit binary codes.

[0062] Random initialization: Generate 100 individuals, with each parameter taking a random value within the feasible region: X correction amount: uniform sampling from [-2, 2] mm (e.g., 0.73 mm → binary code 0101110001).

[0063] Current adjustment amount: uniform sampling from [-15, 15]A (e.g., +8.2A → binary code 1010100010).

[0064] The binary encoding of this invention facilitates genetic operations while maintaining parameter accuracy, and random initialization ensures full coverage of the search space and avoids getting trapped in local optima.

[0065] Fitness calculation process: Decoding parameters: Restore the binary encoding to the actual parameter value: For example, the binary code of the X correction is 0101110001 → decimal value 377 → actual value = 377 × 4 / 1023 ≈ 1.48 mm.

[0066] Substitute into the objective function: Based on the objective function constructed in step S2.1, calculate the total error for each individual (e.g., total error = 0.32).

[0067] Fitness conversion: Fitness = 1 / (1 + total error), for example: total error 0.32 → fitness = 1 / (1 + 0.32) ≈ 0.758.

[0068] This invention transforms the problem of minimizing error into the problem of maximizing fitness, which conforms to the optimization logic of genetic algorithms. The nonlinear transformation (1 / (1+error)) amplifies the difference between high-quality and low-quality solutions, thus accelerating convergence.

[0069] Selection and crossover operation process: Roulette options: Calculate the selection probability of each individual = individual fitness / total population fitness. For example: individual A has a fitness of 0.758, and the total population fitness is 78.3 → selection probability = 0.758 / 78.3 ≈ 0.97%. Generate random numbers between 0 and 1, and select the parent generation according to the probability interval (e.g., random number 0.005 selects individual A). For the selected parent generation (e.g., individual A: 0101110001..., individual B: 1100101010...): Crossover occurs after the 25th position to generate offspring: Offspring C: 0101110001101010..., Offspring D: 1100101010011100..., directly retain the top 10% of individuals with the highest fitness (e.g., the first 10 individuals) to avoid losing the optimal solution. This invention uses a roulette wheel selection method to balance global search (exploration) and local development (utilization). Single-point crossover effectively combines high-quality genes from the parent generation, accelerating evolution towards the optimal solution. For newly generated offspring, a certain bit is randomly flipped with a 5% probability (e.g., the 37th bit changes from 0 to 1). The iteration stops after 10 generations, or terminates early when the change in optimal fitness is <0.1% for 3 consecutive generations. The individual with the highest fitness (e.g., binary code 0101110001101010...) is decoded as the actual parameters: X correction amount = 1.48mm, current adjustment +8.2A, voltage adjustment +0.4V, wire feed speed compensation +0.25m / min.

[0070] The mutation operation of this invention maintains population diversity, avoids premature convergence of the algorithm, and the adaptive termination condition balances computational efficiency and solution quality (10 iterations typically take <200ms).

[0071] In a preferred embodiment, S3, based on the welding status data after the execution of the correction command in S2, monitors abnormal event interruption signals in real time. If arc ignition failure, tracking failure, or flux recovery abnormality is detected, the welding process is paused, the subsystem status is reset, or the process parameter threshold is adjusted according to the preset emergency handling logic, and the abnormal data is uploaded to the MES system, including: Based on the welding status data after the execution of the correction command in S2, the actual values ​​of welding current and voltage of the submerged arc welding machine subsystem, the real-time deviation of the laser weld seam tracking subsystem, and the operating status signal of the flux recovery device are collected in real time. The actual values ​​of welding current and voltage are compared with the set threshold of the correction command. If the number of consecutive deviations exceeds the preset value, it is determined that the arc ignition has failed or the tracking has failed. At the same time, the deviation between the flux recovery flow rate and the preset value is monitored. If it exceeds the allowable range, it is determined that the flux recovery is abnormal. Based on the detected anomaly type, trigger the preset emergency handling logic: If arc initiation fails or tracking fails, the welding process is paused and the welding torch is reset to a safe position. At the same time, the current and voltage thresholds are adjusted to the range of the initial process instructions. If the flux recovery is abnormal, start the backup recovery device and reduce the wire feed speed until the flow rate is restored to the allowable range; The abnormal event type, occurrence time and handling result are uploaded to the MES system via industrial Ethernet, triggering the MES to issue process parameter re-optimization instructions or manual intervention requests. If the welding process is restarted after handling the anomaly, the dynamic optimization of step S2 will be re-executed based on the updated process parameters until the welding state returns to stability.

[0072] In this embodiment of the invention, welding current / voltage is sampled at 100Hz (period 10ms), using a sliding window filter (window size 5); weld deviation is sampled at 50Hz (period 20ms), using Kalman filtering to eliminate jitter. Flux recovery flow rate is sampled at 20Hz (period 50ms), using low-pass filtering to remove pulse noise; multi-source data is aligned based on timestamps (such as IEEE 1588 precise clock synchronization) to ensure parameter consistency during anomaly detection. This invention uses high-frequency sampling to capture transient anomalies (such as current surges caused by arc flash), and multi-sensor data synchronization avoids misjudgments caused by time differences (such as the causal relationship between current fluctuations and position deviations).

[0073] Logic process for determining abnormal events: Arc initiation failure judgment: Condition 1: The welding current is below the threshold (e.g., <200A) for more than 300ms.

[0074] Condition 2: Voltage fluctuations exceed the ±5V threshold and persist for 5 sampling periods.

[0075] If both conditions are met simultaneously, the arc initiation is deemed a failure.

[0076] Failure detection tracking: Condition 1: Absolute value of weld deviation > 2 mm for 200 ms.

[0077] Condition 2: The deviation reduction rate is less than 10% after three consecutive correction commands.

[0078] If any of the conditions are met, the tracking is deemed to have failed.

[0079] Flux recovery anomaly determination: Condition 1: The recovered flow rate is either less than 70% or more than 130% of the set value.

[0080] Condition 2: Standard deviation of flow rate fluctuation > 0.5m 3 / h lasts for 1 second.

[0081] If any condition is met, the recycling is deemed abnormal.

[0082] This invention reduces false alarms by using multi-dimensional judgment rules (such as a single current fluctuation may be a normal arc transition), quantifies the duration and rate of change indicators, and distinguishes between brief disturbances and true anomalies.

[0083] The process of tiered emergency response strategy: Arc initiation / tracking failure handling: Step 1: Send a pause command to the servo system (stop wire feeding within 3ms).

[0084] Step 2: Drive the welding torch back to a safe position at a speed of 50mm / s (e.g., 50mm before the arc starting point).

[0085] Step 3: Relax the current threshold to ±20A (originally ±10A) and the voltage threshold to ±1.5V (originally ±1V).

[0086] Handling of Flux Recovery Abnormalities: Step 1: Switch to the backup recovery channel (solenoid valve response time < 50ms).

[0087] Step 2: Reduce the wire feeding speed by 30% (e.g., from 5m / min to 3.5m / min).

[0088] Step 3: Check the flow rate every 500ms. After it returns to normal, gradually restore the wire feeding speed (increase by 0.2m / min each time).

[0089] The hierarchical response mechanism of this invention avoids a single anomaly from causing a complete shutdown (such as adjusting the wire feeding speed only when there is a recycling anomaly), and the dynamic adjustment of parameter thresholds enhances the robustness of the system (such as relaxing the tracking accuracy requirements when the weld is irregular).

[0090] Abnormal data closed-loop feedback process: Data packaging: Exception data packet format: [timestamp, exception type, parameter value, handling measures, handling result].

[0091] Example: [2023-04-24T10:30:45.123, Arc ignition failed, current 185A, voltage 32V, reset, successful].

[0092] Data is sent to MES via Profinet (data frame size 256 bytes, transmission period < 100ms).

[0093] MES response: If the frequency of anomalies exceeds 3 times per hour, a process re-optimization process will be triggered (such as adjusting beveling preparation parameters). If two consecutive anomalies of the same type occur, a manual intervention reminder will be pushed to the operation and maintenance terminal.

[0094] This invention provides a complete anomaly chain record to support quality traceability (e.g., batch arc ignition failure can be traced back to the welding wire batch), and MES dynamic decision-making reduces the cost of manual inspection (anomaly frequency analysis provides early warning of potential equipment hazards).

[0095] In another preferred embodiment of the present invention, step S4: integrates the welding process data generated in steps S1 to S3, including actual current and voltage, tracking deviation records, and video monitoring images, and feeds them back to the MES system via industrial Ethernet to complete process parameter matching, quality traceability, and closed-loop control command reception, forming a full-process control link from initial process planning, dynamic parameter adjustment, anomaly response to data closed loop; the multiple welding initial process commands include the welding current, voltage, vertical and horizontal position coordinates of each weld pass, including: Real-time parameters: current / voltage (100Hz), tracking deviation (50Hz), flux flow (20Hz), stored in time series.

[0096] Video data: 25 frames per second of high-definition video, compressed by an edge computing server (compression ratio 10:1) and then stored as key frames (such as arc initiation, arc termination, and abnormal moments).

[0097] Add metadata to each piece of data: Pipe number (e.g., P-20240524-001), weld sequence number (3rd layer, 5th weld), process stage (dynamic optimization stage).

[0098] Establishing associations: The current abrupt change point (such as 310A→280A) is bound to the video frame (showing the fluctuation of the molten pool) to form a "parameter-image" associated record.

[0099] This invention uses high-frequency data to accurately reproduce welding process details (such as 0.1-second-level current fluctuations that can be traced back to welding torch vibration), and links video and parameters for retrieval to quickly locate the cause of defects (such as the image of the moment when the welding torch is offset by 0.8mm corresponding to undercut).

[0100] Data transmission and MES system integration process: The fieldbus data (such as Profinet) is converted into the OPCUA format that the MES can recognize. The mapping relationship is as follows: Current → ns = 2; s = Welding.Current, Voltage → ns = 2; s = Welding.Voltage.

[0101] Batch transfer optimization A "change-triggered + timed aggregation" mechanism is adopted: Transmit immediately when parameter changes exceed 5%; otherwise, transmit in packets every minute (reducing data volume by approximately 60%). The receiving end ensures data integrity through CRC check and automatically retransmits abnormal packets (retransmission rate < 0.1%).

[0102] This invention reduces communication bandwidth usage to 40% of traditional full-volume transmission, making it suitable for concurrent data uploads on large-scale production lines. It balances real-time performance with reliability, and critical abnormal data (such as arc initiation failure) can be delivered to the MES within 200ms.

[0103] Process matching and quality traceability: Process parameter matching MES establishes a mapping table of "pipeline characteristics - process parameters": Input: Pipe diameter φ813mm, wall thickness 28mm, material X70 → Match historical successful process parameter set (current 290-320A, voltage 28-30V, compare actual parameters with matching parameters, such as a weld current average of 275A (below the lower limit of 5%), mark it as "parameter deviation".

[0104] Quality traceability logic: Defect tracing path: Defect type (incomplete fusion) → Weld sequence number (2nd layer, 3rd pass) → Timestamp (10:45:23) → Tracking deviation record (X=1.2mm, Y=0.9mm) → Operator (shift 03 group); Generate parameter trend reports weekly / monthly to identify potential risks (e.g., the standard deviation of current fluctuation is 20% higher than the mean when a welder is operating). The process matching efficiency of this invention is improved by 70%, the parameter generation time for new work orders is shortened from 2 hours to 15 minutes, quality traceability is accurate from "batch level" to "weld level", and the problem location time is shortened from 4 hours to 15 minutes.

[0105] Closed-loop control command generation and execution process: Command trigger conditions, automatically triggered: If the parameter deviation rate of the same weld bead in three consecutive pipes is greater than 10%, a process optimization instruction will be generated (e.g., the lower limit of current is adjusted to 285A).

[0106] Manual trigger: Quality inspectors manually issue parameter adjustment commands through the MES interface (e.g., for edge defects, reduce the voltage by 0.5V).

[0107] Instruction parsing and execution: Main control system interprets the instruction: Adjust the voltage of the third weld bead to 28.5V (original setting 29V), effective time: welding of the third layer of the next pipe begins.

[0108] Execution verification: 500ms before welding begins, the system compares the new parameters with the equipment capabilities (e.g., the voltage of 28.5V is within the allowable range of 26-32V), and sends the data after confirming that there are no errors.

[0109] The process optimization cycle of this invention changes from "post-process batch adjustment" to "real-time iteration during the process", improving the yield rate by 8%. Closed-loop control reduces reliance on human experience, and new employees can quickly reuse historically successful processes (shortening the training cycle by 50%).

[0110] like Figure 2 As shown, embodiments of the present invention also provide an automatic tracking multi-pass welding control system for submerged arc welding of structural pipelines, comprising: The generation module is used to generate multiple initial welding process instructions based on the pre-set process database or MES system work order parameters, and send them to the execution layer subsystem via industrial Ethernet to trigger the laser weld seam tracking subsystem and submerged arc welding machine subsystem to execute the initial welding process. The correction module is used to dynamically optimize the welding path and parameters based on the real-time deviation of the weld position fed back by the execution layer subsystem and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, generate correction instructions, and send them to the execution layer via industrial Ethernet to drive the servo motor to adjust the welding torch posture. The detection module is used to monitor abnormal event interruption signals in real time based on the welding status data after the correction command is executed. If arc ignition failure, tracking failure or flux recovery abnormality is detected, the welding process is suspended, the subsystem status is reset or the process parameter threshold is adjusted according to the preset emergency handling logic, and the abnormal data is uploaded to the MES system. The processing module integrates the generated welding process data, including actual current and voltage, tracking deviation records, and video monitoring images, and feeds them back to the MES system via industrial Ethernet to complete process parameter matching, quality traceability, and closed-loop control command reception, forming a complete process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closure.

[0111] It should be noted that this system is a system corresponding to the above method. All implementation methods in the above method embodiments are applicable to this embodiment and can achieve the same technical effect.

[0112] Embodiments of the present invention also provide a computing device, including: a processor and a memory storing a computer program, wherein the computer program, when executed by the processor, performs the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0113] Preferably, the submerged arc welding automatic tracking multi-pass welding control system proposed in this invention includes a welding host and tooling control main system, a laser weld seam tracking control subsystem, a submerged arc welding machine control subsystem, a flux control subsystem, a communication subsystem with the factory MES system, and a video monitoring system. The main system and subsystems communicate and interact with each other via industrial Ethernet, enabling real-time command parameter issuance and status data feedback.

[0114] The main control layer, with the welding host and tooling control main system (PLC / industrial computer) as the core, is responsible for multi-path welding planning, subsystem coordination, data aggregation and logic control.

[0115] The execution layer includes subsystems such as laser tracking, welding machine, flux, and video monitoring, which interact with the main system in real time via industrial Ethernet to exchange commands and status data.

[0116] The communication layer is based on industrial Ethernet (Profinet / ModbusTCP) to achieve high-speed communication, ensuring real-time performance and reliability.

[0117] Welding machine and tooling control main system: Multi-pass welding path planning generates parameters such as welding current, welding voltage, vertical position coordinates, and horizontal position coordinates for each pass of the multi-pass welding process based on the machine's preset process database or the command parameters issued by the MES.

[0118] Motion control, control and real-time position acquisition of various mechanisms of the welding host (manipulator) and tooling (roller frame).

[0119] Data fusion and decision-making integrates data such as laser tracking and welding machine current and voltage to optimize welding parameters in real time.

[0120] Laser weld seam tracking and control subsystem: Based on the principle of line laser triangulation, the weld position deviation is detected in real time. Lateral / longitudinal correction commands for the welding torch are generated to drive the servo motor to adjust the torch's posture, and communication with the main system is achieved via Modbus TCP.

[0121] Submerged arc welding machine control subsystem: The digital submerged arc welding machine allows for software-based adjustment of welding parameters (current, voltage, wire feed speed) and communicates with the main system via Profinet.

[0122] Flux control subsystem: The main control measures include the feeding and recovery of flux, flux recycling and reuse to reduce waste; and the negative pressure recovery device design prevents flux from scattering and polluting the environment.

[0123] MES Communication Subsystem: Work order management: Receives welding process cards issued by MES and automatically matches process library parameters.

[0124] Data traceability, uploading welding process data (welding current, welding voltage, heat input, and tracking deviation records, etc.).

[0125] Video surveillance system: A dedicated optical lens for submerged arc welding, covering the welding position on the welding head; enabling remote monitoring of the welding process.

[0126] Data interaction and collaborative control: Communication protocols and data streams: Once the main system issues commands, parameters such as the target position coordinates of the welding torch, current and voltage settings, and laser tracking are transmitted to each subsystem.

[0127] The subsystem feeds back data, including the real-time position coordinates of the welding torch, the actual current and voltage of the welding machine, and the laser tracking deviation, and uploads these status data to the main system. Abnormal events (such as arc ignition failure, tracking failure, etc.) are reported first through interrupt signals, triggering the main system to handle emergencies.

[0128] This invention enables paperless closed-loop production from MES work order to welding completion; from the start of welding to the completion of multiple welding passes, manual intervention is reduced by more than 90%; laser correction + welding gun servo control achieves a comprehensive tracking accuracy of ≤±0.1mm; it supports materials such as carbon steel and stainless steel, as well as various plate thicknesses and various bevel forms.

[0129] Embodiments of the present invention also provide a computer-readable storage medium storing instructions that, when executed on a computer, cause the computer to perform the method described above. All implementations in the above method embodiments are applicable to this embodiment and can achieve the same technical effects.

[0130] Obviously, the above embodiments of the present invention are merely examples for clearly illustrating the present invention, and are not intended to limit the implementation of the present invention. Those skilled in the art can make other variations or modifications based on the above description. It is neither necessary nor possible to exhaustively describe all embodiments here. Any modifications, equivalent substitutions, and improvements made within the spirit and principles of the present invention should be included within the scope of protection of the claims of the present invention.

Claims

1. A structural pipe submerged arc welding automatic tracking multi-pass welding control method applied to a master control layer PLC, characterized in that, Mainly includes the following steps: S1, based on the preset process database or MES system work order parameters, generate multiple pass welding initial process instructions, and issue to the execution layer subsystem through industrial Ethernet, trigger laser weld tracking subsystem and submerged arc welding machine subsystem to execute initial welding process; S2, according to the real-time deviation of the weld position fed back by the execution layer subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, dynamically optimize the welding path and parameters, generate correction instructions, and issue to the execution layer through industrial Ethernet, drive servo motor to adjust the welding gun posture; S3, based on the welding state data after executing the correction instruction in S2, real-time monitor abnormal event interrupt signal, if detecting arc striking failure, tracking failure or flux recovery abnormality, then according to the preset emergency processing logic, pause the welding process, reset the subsystem state or adjust the process parameter threshold, and upload the abnormal data to the MES system; S4, integrate the welding process data generated in S1 to S3, including actual current voltage, tracking deviation record and video monitoring image, feedback to the MES system through industrial Ethernet, complete process parameter matching, quality traceability and closed loop control instruction receiving, form the whole process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closed loop.

2. The control method of the automatic tracking multi-pass welding of the structural pipe submerged arc welding according to claim 1, characterized by, The multiple pass welding initial process instruction contains the welding current, voltage, vertical and horizontal position coordinates of each welding pass.

3. The control method of the automatic tracking multi-pass welding of the structural pipe submerged arc welding according to claim 2, characterized by, In S1 also includes: Analyzing the preset process database or MES system work order parameters, extracting the welding process requirements, generating initial process instructions containing welding current, voltage, vertical and horizontal position coordinates of each welding pass; Issue the initial process instructions to the execution layer subsystem through industrial Ethernet, trigger the laser weld tracking subsystem to start real-time weld position detection, and synchronously drive the submerged arc welding machine subsystem to execute welding action according to the initial parameters; Based on the initial process instruction issue result, receive the initial weld positioning data fed back by the laser weld tracking subsystem, verify the matching of the initial position of the welding gun and the target coordinates, if the deviation exceeds the preset threshold, generate the calibrated instruction to make the submerged arc welding machine subsystem execute the arc striking operation according to the calibrated instruction.

4. The control method of the automatic tracking multi-pass welding of the structural pipe submerged arc welding according to claim 3, characterized by, The correction instruction includes the corrected welding gun target position coordinates, current and voltage set values; In S2 also includes: based on the real-time deviation of the weld position fed back by the laser weld tracking subsystem in S1 and the actual current and voltage data uploaded by the submerged arc welding machine subsystem, construct a dynamic optimization objective function; Use genetic algorithm to iteratively optimize the objective function, generate the final candidate parameter set through selection, crossover and mutation operations, including welding gun target position coordinate correction amount, current and voltage adjustment amount and wire feeding speed compensation value; According to the final candidate parameter set, generate correction instructions and issue to the execution layer through industrial Ethernet, drive servo motor to adjust the welding gun horizontal or vertical posture, synchronously adjust the wire feeding speed and welding voltage of the submerged arc welding machine.

5. The control method of the automatic tracking multi-pass welding of a structural pipe submerged arc welding according to claim 4, wherein According to the real-time deviation of the weld position, calculate the horizontal and vertical tracking errors of the welding gun, and based on the difference between the actual current and voltage data of the submerged arc welding machine subsystem and the set values of the initial process instruction, determine the welding parameter error amount; The tracking error, the welding parameter error and the path smoothness are taken as optimization objectives in combination with a preset welding path smoothness constraint condition, and weight coefficients of optimization items in a target function are defined; Based on the initial process instruction parameter range generated in S1, feasible domain constraint conditions of the welding torch position correction amount, the current and voltage adjustment amount and the wire feeding speed compensation value are set; The optimization objectives, the weight coefficients and the feasible domain constraint conditions are comprehensively considered to construct a dynamic optimization target function.

6. The control method of the automatic tracking multi-pass welding of a structural pipe submerged arc welding according to claim 4, wherein Based on the set feasible domain constraint conditions, a population of a genetic algorithm is initialized, the population is composed of multiple candidate parameter individuals, and each individual includes a random combination of the welding torch position correction amount, the current and voltage adjustment amount and the wire feeding speed compensation value; The fitness value of each candidate parameter individual is calculated by using the constructed dynamic optimization target function, and the fitness value is used to represent the comprehensive optimization effect of the corresponding parameter combination on the tracking error, the welding parameter error and the path smoothness; According to the fitness value, a roulette selection operation is performed to select the corresponding individual as a parent, and a new child individual is generated through a single-point crossover operation; The fitness value and the single-point crossover operation are iteratively calculated until a preset iteration number is met, and a final candidate parameter set in the current population is output.

7. The control method for automatic tracking multi-pass welding of a structural pipe by submerged arc welding according to claim 1, wherein In S3, the following steps are further included: Based on the welding state data after the execution of the correction instruction in S2, welding current and voltage actual values of the submerged arc welding machine subsystem, real-time deviation amounts of the laser weld seam tracking subsystem and operation state signals of the flux recovery device are collected in real time; The welding current and voltage actual values are compared with set threshold values of the correction instruction, if the number of continuous overruns exceeds a preset value, it is determined that the arc striking fails or the tracking fails, and meanwhile, the deviation of the flux recovery flow from a preset value is monitored, if the deviation exceeds a permitted range, it is determined that the flux recovery is abnormal; According to the detected abnormal type, a preset emergency processing logic is triggered; If the arc striking fails or the tracking fails, the welding process is paused and the welding torch is reset to a safe position, and meanwhile, the current and voltage threshold values are adjusted to the initial process instruction range; If the flux recovery is abnormal, a backup recovery device is started and the wire feeding speed is reduced until the flow recovers to the permitted range; The abnormal event type, occurrence time and processing result are uploaded to the MES system through the industrial Ethernet, and a process parameter re-optimization instruction or a manual intervention request is triggered by the MES; If the welding process is restarted after the abnormal processing, the dynamic optimization of step S2 is re-executed based on the updated process parameters until the welding state recovers to be stable.

8. A control system for automatic tracking multi-pass submerged arc welding of a structural pipe, characterized by, The method is used for executing the method as claimed in any one of claims 1 to 7, comprising: A generation module is configured to generate initial process instructions of multi-pass welding based on preset process database or MES system work order parameters, and to issue the initial process instructions to an execution layer subsystem through an industrial Ethernet to trigger a laser weld seam tracking subsystem and a submerged arc welding machine subsystem to execute an initial welding process; A correction module is configured to dynamically optimize welding paths and parameters based on real-time deviation amounts of weld seam positions fed back by the execution layer subsystem and actual current and voltage data uploaded by the submerged arc welding machine subsystem, to generate correction instructions, and to issue the correction instructions to the execution layer through the industrial Ethernet to drive a servo motor to adjust a welding torch posture. The detection module is used for monitoring an abnormal event interrupt signal in real time based on the welding state data after the correction instruction is executed, and if an arc striking failure, a tracking failure or a flux recovery abnormality is detected, the welding process is paused, the subsystem state is reset or the process parameter threshold is adjusted according to a preset emergency processing logic, and the abnormal data is uploaded to the MES system; The processing module is used for integrating the generated welding process data, including actual current and voltage, tracking deviation record and video monitoring image, feeding back to the MES system through an industrial Ethernet, completing process parameter matching, quality traceability and closed-loop control instruction receiving, and forming a whole-process control link from initial process planning, dynamic parameter adjustment, abnormal response to data closed loop.

Images