An all-autonomous welding equipment adapting to unstructured environment and a welding seam tracking and process optimization method thereof

Abstract

The application discloses a kind of full autonomous welding equipment and its welding seam tracking and process optimization method suitable for unstructured environment.The application belongs to the technical field of welding robot.The equipment includes: omnidirectional mobile platform, for autonomously walking in unstructured environment;Welding robot arm, end integrated welding torch;Welding power supply, for providing welding current and welding voltage, and real-time acquisition arc voltage signal and arc current signal in welding process;Intelligent control unit, including: arc feature extraction module, time-frequency analysis is carried out, extracts arc inherent oscillation frequency and its variation characteristics, welding seam deviation calculation module, the lateral deviation and height deviation of welding torch end are calculated, process parameter adjustment module extracts molten pool oscillation phase, when penetration state deviates from the set target, adjust the output current waveform parameter of welding power supply.The application realizes welding seam tracking and penetration control using the physical characteristics of welding arc itself, and can realize autonomous welding in unstructured environment.

Description

Technical Field

[0001] This application belongs to the field of welding robot technology, specifically relating to a fully autonomous welding equipment adapted to unstructured environments and its weld seam tracking and process optimization method. Background Technology

[0002] Welding is an indispensable key process in modern manufacturing, widely used in shipbuilding, bridge construction, pressure vessels, and steel structure buildings. Currently, automated welding equipment is mainly used in structured environments such as fixed workstations. However, workpieces such as ship sections and large storage tanks are large in size and complex in shape, making them difficult to move to fixed workstations. Therefore, welding equipment with autonomous movement capabilities is urgently needed. However, existing technologies still have the following shortcomings: First, mobile platforms are not adaptable to unstructured environments. Traditional wheeled or tracked platforms have difficulty moving stably on curved, sloping, and vertical workpieces, and are prone to slippage or overturning; bionic hexapod platforms have limited load-bearing capacity and are complex to control; and track-based trolleys require the laying of tracks and have poor adaptability.

[0003] Second, weld seam tracking lacks reliability under strong interference environments. Arc light, spatter, and fumes during the welding process severely interfere with visual sensors, leading to signal loss or distortion. Existing dual-modal fusion strategies for vision and arc sensing are mostly fixed-weight fusion, lacking dynamic confidence assessment and switching mechanisms. When a certain modality fails, the fusion result is still affected by the degraded signal.

[0004] Third, welding process parameters lack real-time closed-loop control. Existing solutions mostly rely on pre-weld process experiments to set parameters, and cannot adjust them in real time according to the actual state of the molten pool during welding. Cloud analysis suffers from communication delays, and molten pool image acquisition is severely affected by arc light interference, making real-time process optimization difficult.

[0005] Fourth, the impact of welding thermal deformation on trajectory accuracy has not been effectively compensated. Welding heat input causes a non-uniform temperature field and thermal deformation in the workpiece, causing the actual position of the weld to deviate from the initial planned trajectory. Existing tracking methods do not consider feedforward compensation for thermal deformation.

[0006] In summary, existing technologies lack a fully autonomous welding equipment capable of simultaneously addressing issues such as adaptability to unstructured environments on mobile platforms, reliability of weld seam tracking under strong interference environments, real-time closed-loop control of welding process parameters, and thermal deformation compensation. Therefore, it is necessary to propose a new technical solution to overcome the aforementioned deficiencies in existing technologies. Summary of the Invention

[0007] This application provides a fully autonomous welding equipment adapted to unstructured environments and its weld seam tracking and process optimization method. The purpose is to solve the technical problems of poor stability of existing mobile welding equipment in unstructured environments such as curved surfaces, slopes and vertical surfaces, insufficient accuracy of weld seam tracking due to interference from arc light spatter, and inability to adjust welding process parameters in real time in a closed loop.

[0008] In a first aspect, embodiments of this application provide a fully autonomous welding equipment adapted to unstructured environments, characterized in that it includes: an omnidirectional mobile platform, a welding robotic arm, a welding power source, and an intelligent control unit; The omnidirectional mobile platform is used for autonomous movement in unstructured environments and includes multiple independently driven omnidirectional wheels or Mecanum wheels. The welding robotic arm is mounted on the omnidirectional moving platform and has a welding torch integrated at its end. The welding torch is electrically connected to the welding power source. The welding power source is used to provide welding current and welding voltage to the welding torch, and to collect arc voltage and arc current signals in real time during the welding process. The intelligent control unit is communicatively connected to the omnidirectional mobile platform, the welding robotic arm and the welding power source, and includes an arc feature extraction module, a weld deviation calculation module and a process parameter adjustment module. The arc feature extraction module is used to perform time-frequency analysis on the arc voltage signal and arc current signal to extract the arc's inherent oscillation frequency and its variation characteristics. The weld deviation calculation module is used to calculate the lateral and height deviations of the welding torch tip relative to the weld centerline based on the variation characteristics of the inherent oscillation frequency of the electric arc. The process parameter adjustment module is used to extract the molten pool oscillation phase based on the variation characteristics of the inherent oscillation frequency of the electric arc, and to determine the current penetration state based on the molten pool oscillation phase. When the penetration state deviates from the set target, the output current waveform parameters of the welding power supply are adjusted.

[0009] Furthermore, the weld deviation calculation module is specifically used for: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. When the fundamental frequency component increases and the harmonic component decreases, it is determined that the welding torch shifts to the left of the weld centerline. When the fundamental frequency component decreases and the harmonic component increases, it is determined that the welding torch is shifted to the right of the weld centerline. Based on the changes in the fundamental frequency component and harmonic components, the lateral deviation of the welding torch tip relative to the weld centerline is calculated.

[0010] Furthermore, the process parameter adjustment module is specifically used for: The arc current signal is bandpass filtered to extract the signal components within the molten pool oscillation frequency band; Perform a Hilbert transform on the signal components to obtain the instantaneous phase of the molten pool oscillation; The instantaneous phase is compared with a preset phase interval of the melt penetration state; When the instantaneous phase is in the incomplete penetration phase range, increase the pulse peak value of the welding current or prolong the pulse base value time; When the instantaneous phase is in the overpenetration phase range, reduce the pulse peak value of the welding current or shorten the pulse base value time.

[0011] Furthermore, it also includes a thermal deformation prediction module, which is used for: The welding heat input is calculated in real time based on the welding current, welding voltage, welding speed, and workpiece thickness. Based on a preset heat conduction model, the temperature field distribution of the workpiece under the current heat input is calculated. Based on the temperature field distribution, predict the thermal deformation trend of the workpiece during the welding process; Based on the thermal deformation trend, the welding trajectory of the welding robot arm is feedforward and the anti-deformation compensation amount is preset.

[0012] Furthermore, the thermal deformation prediction module is also used for: Based on the lateral and height deviations output by the weld deviation calculation module, the welding trajectory of the welding robot arm is corrected. The feedforward correction amount and the feedback correction amount are superimposed to generate the final welding trajectory control command.

[0013] Furthermore, the omnidirectional moving platform includes a magnetic adsorption component, which is disposed at the bottom of the omnidirectional moving platform and is used to provide adsorption force when the omnidirectional moving platform moves on the surface of a ferromagnetic workpiece. The magnetic adsorption assembly includes multiple permanent magnets and a lifting drive mechanism disposed below each of the permanent magnets. The lifting drive mechanism is used to drive each of the permanent magnets to switch between an adsorption position and a detachment position.

[0014] Furthermore, it also includes a visual positioning module, which is used for: Before welding begins, images of the weld start and end points are acquired to identify the macroscopic direction of the weld. Based on the macroscopic orientation of the weld seam, the initial welding trajectory of the welding robot arm is planned; During the welding process, the vision positioning module does not participate in weld seam tracking, but is only used to monitor the safety status of the welding area.

[0015] Furthermore, the intelligent control unit also includes an adaptive arc-starting module, which is used for: At the start of welding, the welding torch is brought close to the workpiece surface, and the arc voltage signal is detected in real time. When the arc voltage signal is detected to have arc ignition characteristics, the current welding torch position is locked as the welding start point. The welding trajectory is automatically generated based on the welding start point and the preset weld direction.

[0016] Secondly, embodiments of this application provide a weld seam tracking and process optimization method for fully autonomous welding equipment adapted to unstructured environments, executed using the welding equipment described above, and including the following steps: S1 controls the omnidirectional mobile platform to autonomously move to the target welding station, obtains the start and end points of the weld seam through the visual positioning module, and plans the initial welding trajectory. S2, control the welding robotic arm to move the welding torch to the weld start point, and use the adaptive arc start module to detect arc start characteristics and lock the welding start point; S3, start welding, and collect arc voltage and arc current signals in real time; S4, perform time-frequency analysis on the arc voltage signal, extract the arc's inherent oscillation frequency and its variation characteristics, and calculate the lateral and height deviations of the welding torch tip relative to the weld centerline. S5. Based on the lateral deviation and height deviation, adjust the welding gun position of the welding robot arm in real time so that the welding gun is always aligned with the center line of the weld. S6, perform bandpass filtering and Hilbert transform on the arc current signal, extract the molten pool oscillation phase, and determine the current melting state; S7, When the penetration state deviates from the set target, adjust the output current waveform parameters of the welding power supply; S8 calculates the welding heat input in real time based on the welding current, welding voltage, welding speed and workpiece thickness, predicts the workpiece thermal deformation trend, and feeds forward to correct the welding trajectory.

[0017] Furthermore, in step S4, the time-frequency analysis includes: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. Based on the variation law of the fundamental frequency component and harmonic components, the lateral deviation value of the welding torch tip relative to the weld centerline is calculated.

[0018] The technical solution provided in this application, through the coordinated control of magnetic adsorption and tracked composite mobile platform and adsorption force, enables the equipment to stably attach and move in unstructured environments such as curved surfaces, slopes, and vertical surfaces, solving the problem of poor adaptability of traditional wheeled or tracked platforms. Through the hierarchical fusion of vision, structured light, and force perception modes and the dynamic switching mechanism of confidence, it can still maintain highly reliable weld tracking under strong arc light and spatter interference. Through dynamic feature extraction of the molten pool and prediction by a reduced-order thermo-coupling model, it can realize the feedforward real-time adjustment of welding process parameters, significantly improving weld formation quality and welding efficiency, and realizing fully autonomous high-quality welding operations in unstructured environments. Attached Figure Description

[0019] Figure 1 This is a schematic diagram of the structure of the fully autonomous welding equipment adapted to unstructured environments provided in the embodiments of this application; Figure 2 This is a flowchart illustrating the operation method of the fully autonomous welding equipment adapted to unstructured environments provided in the embodiments of this application. Detailed Implementation

[0020] To make the objectives, technical solutions, and advantages of this application clearer, specific embodiments of this application will be described in further detail below with reference to the accompanying drawings. It should be understood that the specific embodiments described herein are merely for explaining this application and not for limiting it. It should also be noted that, for ease of description, only the parts relevant to this application are shown in the drawings, not all of them. Before discussing exemplary embodiments in more detail, it should be mentioned that some exemplary embodiments are described as processes or methods depicted as flowcharts. Although the flowcharts describe operations (or steps) as sequential processes, many of these operations can be performed in parallel, concurrently, or simultaneously. Furthermore, the order of the operations can be rearranged. The process can be terminated when its operation is completed, but may also have additional steps not included in the drawings. The process can correspond to a method, function, procedure, subroutine, subroutine, etc.

[0021] The technical solutions of the embodiments of this application will be clearly described below with reference to the accompanying drawings. Obviously, the described embodiments are only some, not all, of the embodiments of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application are within the scope of protection of this application.

[0022] The terms "first," "second," etc., used in the specification and claims of this application are used to distinguish similar objects and not to describe a specific order or sequence. It should be understood that such terms can be used interchangeably where appropriate so that embodiments of this application can be implemented in orders other than those illustrated or described herein, and the objects distinguished by "first," "second," etc., are generally of the same class and the number of objects is not limited; for example, a first object can be one or more. Furthermore, in the specification and claims, "and / or" indicates at least one of the connected objects, and the character " / " generally indicates that the preceding and following objects are in an "or" relationship.

[0023] The following detailed description, in conjunction with the accompanying drawings, of the specific embodiments and application scenarios provided in this application, illustrates the operating methods, apparatus, equipment, and media.

[0024] Example 1 Figure 1 This is a structural schematic diagram of the fully autonomous welding equipment adapted to unstructured environments provided in the embodiments of this application. For example... Figure 1 As shown, a fully autonomous welding equipment 100 adapted to unstructured environments is characterized by its ability to move autonomously and complete high-quality welding on complex terrains or large workpiece surfaces without the need for external tracks or fixed workstations. The equipment 100 mainly includes an omnidirectional moving platform 110, a welding robotic arm 120, a welding power source 130, and an intelligent control unit 140.

[0025] The omnidirectional mobile platform 110 can be a chassis structure with four or six independent drive wheels. Its wheels are omnidirectional wheels or Mecanum wheels, enabling omnidirectional movements such as straight movement, lateral movement, diagonal movement, and rotation in place within a plane. The platform 110 is used to support components such as the welding robotic arm 120, the welding power supply 130, and the intelligent control unit 140, and autonomously moves to the target welding position in an unstructured environment according to the instructions of the intelligent control unit 140.

[0026] The welding robotic arm 120, which may be a serial industrial robot with six or more degrees of freedom, integrates a welding torch at its end effector. The robotic arm 120 is fixedly mounted on an omnidirectional moving platform 110 and is used to adjust the spatial orientation of the welding torch according to control commands, aligning it with the weld seam and maintaining the set welding angle and extension. The welding torch is electrically connected to a welding power supply 130 via cables to receive welding current and voltage and perform welding operations.

[0027] The welding power supply 130 can be a digital inverter welding power supply with fast response and waveform control capabilities. On the one hand, the power supply 130 provides the current and voltage required for welding to the welding torch. On the other hand, its internal high-frequency sampling circuit can collect the arc voltage and arc current signals during the welding process in real time and transmit these signals to the intelligent control unit 140 for analysis.

[0028] The intelligent control unit 140 can be an industrial controller integrating a high-performance processor and multiple communication interfaces, such as a control board based on a field-programmable gate array (FPGA) or a digital signal processor (DSP). This control unit 140 is communicatively connected to the omnidirectional mobile platform 110, the welding robotic arm 120, and the welding power supply 130, respectively, for receiving sensor data, executing core algorithms, and issuing control commands. Internally, it integrates an arc feature extraction module 141, a weld deviation calculation module 142, and a process parameter adjustment module 143.

[0029] The arc feature extraction module 141 can be a signal processing algorithm module running within the intelligent control unit 140. This module 141 is used to perform time-frequency analysis on the received arc voltage signal and arc current signal. For example, by using short-time Fourier transform or wavelet transform, it extracts the inherent oscillation frequency of the arc and its dynamic change characteristics over time from the strong noise background, providing a data basis for subsequent deviation calculation and process adjustment.

[0030] The weld deviation calculation module 142 can be a software algorithm module based on a feature mapping model. The weld deviation calculation module 142 is used to receive the arc inherent oscillation frequency change characteristics output by the arc feature extraction module 141, and calculate the lateral deviation and height deviation of the welding torch tip relative to the weld centerline according to the preset frequency-deviation mapping relationship, and output the deviation values ​​in real time.

[0031] The process parameter adjustment module 143 can be an algorithm module with closed-loop control logic. The process parameter adjustment module 143 is used to further extract the molten pool oscillation phase based on the arc's inherent oscillation frequency change characteristics provided by the arc feature extraction module 141, and determine the current penetration state based on this phase. When the penetration state deviates from the set target, for example, if there is a trend of incomplete penetration or over-penetration, the process parameter adjustment module 143 will adjust the output current waveform parameters of the welding power supply 130 in real time, such as pulse peak value and base time, to correct the penetration state.

[0032] The technical solution provided in this embodiment integrates an omnidirectional mobile platform, a welding robotic arm, a welding power source, and an intelligent control unit into an organic whole. By utilizing the physical characteristics of the welding arc itself as a sensing source, it achieves autonomous movement in unstructured environments and weld seam tracking and adaptive process adjustment without the need for external sensors. This significantly reduces system costs and sensitivity to environmental interference, and improves the automation level and environmental adaptability of welding equipment.

[0033] In one embodiment, optionally, the weld deviation calculation module 142 is specifically used for: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. When the fundamental frequency component increases and the harmonic component decreases, it is determined that the welding torch shifts to the left of the weld centerline. When the fundamental frequency component decreases and the harmonic component increases, it is determined that the welding torch is shifted to the right of the weld centerline. Based on the changes in the fundamental frequency component and harmonic components, the lateral deviation of the welding torch tip relative to the weld centerline is calculated.

[0034] The weld deviation calculation module 142, as a core algorithm unit in the intelligent control unit 140, is implemented as follows: The weld deviation calculation module 142 first performs a short-time Fourier transform on the arc voltage signal. The short-time Fourier transform is a time-frequency analysis method that uses a fixed-length window function sliding along the time axis to perform a Fourier transform on the signal within the window, thereby obtaining a two-dimensional spectrum of the signal frequency changing over time. Through this spectrum, the dynamic changes of the arc's inherent oscillation frequency and its fundamental and harmonic components can be clearly observed.

[0035] The fundamental frequency component refers to the most dominant frequency component in the arc oscillation signal, typically corresponding to the arc's natural oscillation frequency. Harmonic components are frequency components whose frequencies are integer multiples of the fundamental frequency; their amplitudes are related to the arc's combustion stability. The weld deviation calculation module 142 determines the welding torch's offset direction by extracting the amplitude changes of the fundamental and harmonic components from the frequency spectrum. Specifically, when an increase in the fundamental frequency component and a decrease in the harmonic components are detected, the welding torch is determined to have shifted to the left of the weld centerline; conversely, when a decrease in the fundamental frequency component and an increase in the harmonic components are detected, the welding torch is determined to have shifted to the right of the weld centerline.

[0036] Furthermore, the weld deviation calculation module 142 calculates the specific lateral deviation value of the welding torch tip relative to the weld centerline based on the magnitude of the changes in the fundamental frequency component and harmonic components, using a built-in calibration curve or lookup table. For example, the deviation value can be obtained by multiplying the rate of change of the fundamental frequency component by a preset proportionality coefficient.

[0037] This technical solution provides a method for calculating weld deviation based on the frequency domain characteristics of electric arc signals. This method has low computational complexity and strong real-time performance. It does not require additional visual or structured light sensors. It can accurately determine the lateral offset direction and magnitude of the welding torch simply by analyzing the spectral changes of the arc voltage signal. This effectively reduces the system hardware cost and sensitivity to environmental interference such as arc light and spatter, and improves the reliability of weld tracking.

[0038] In one embodiment, optionally, the process parameter adjustment module 143 is specifically used for: The arc current signal is bandpass filtered to extract the signal components within the molten pool oscillation frequency band; Perform a Hilbert transform on the signal components to obtain the instantaneous phase of the molten pool oscillation; The instantaneous phase is compared with a preset phase interval of the melt penetration state; When the instantaneous phase is in the incomplete penetration phase range, increase the pulse peak value of the welding current or prolong the pulse base value time; When the instantaneous phase is in the overpenetration phase range, reduce the pulse peak value of the welding current or shorten the pulse base value time.

[0039] The process parameter adjustment module 143, as another core algorithm unit in the intelligent control unit 140, is implemented as follows. The process parameter adjustment module 143 first performs bandpass filtering on the arc current signal. Bandpass filtering is a signal processing technique that allows signals within a specific frequency band to pass through while attenuating signals in other frequency bands. By setting a passband that matches the molten pool oscillation frequency range, signal components reflecting the dynamic behavior of the molten pool can be extracted, and high-frequency noise and low-frequency interference can be filtered out.

[0040] Subsequently, the process parameter adjustment module 143 performs a Hilbert transform on the extracted signal components. The Hilbert transform is a mathematical transformation used to generate an analytic signal, thereby obtaining the instantaneous amplitude and instantaneous phase of the original signal. Through the Hilbert transform, the instantaneous phase of the molten pool oscillation can be obtained, and this phase information reflects the oscillation state of the molten pool at the current moment.

[0041] The process parameter adjustment module 143 compares the calculated instantaneous phase with the preset penetration state phase interval. The penetration state phase interval is pre-calibrated through experiments or simulations. For example, the incomplete penetration state corresponds to one phase interval, and the over-penetration state corresponds to another phase interval. When the instantaneous phase falls into the incomplete penetration phase interval, it indicates insufficient welding heat input. The process parameter adjustment module 143 will control the welding power supply 130 to increase the pulse peak value of the welding current or extend the pulse base value time to increase heat input. When the instantaneous phase falls into the over-penetration phase interval, it indicates excessive welding heat input. The process parameter adjustment module 143 will control the welding power supply 130 to decrease the pulse peak value of the welding current or shorten the pulse base value time to reduce heat input.

[0042] This technical solution provides a real-time judgment and closed-loop control method for the penetration state based on the oscillation phase of the molten pool. This method directly uses the arc current signal for phase analysis, has a fast response speed, can sense the dynamic changes of the molten pool in real time, and automatically adjusts the welding current waveform parameters according to the penetration state, effectively preventing welding defects such as incomplete penetration and burn-through, and significantly improving the stability and consistency of welding quality.

[0043] In one embodiment, optionally, a thermal deformation prediction module is further included, the thermal deformation prediction module being used for: The welding heat input is calculated in real time based on the welding current, welding voltage, welding speed, and workpiece thickness. Based on a preset heat conduction model, the temperature field distribution of the workpiece under the current heat input is calculated. Based on the temperature field distribution, predict the thermal deformation trend of the workpiece during the welding process; Based on the thermal deformation trend, the welding trajectory of the welding robot arm is feedforward and the anti-deformation compensation amount is preset.

[0044] The thermal deformation prediction module can be an independent algorithm module within the intelligent control unit 140, or it can be a software functional unit integrated within it. The thermal deformation prediction module is used to predict and compensate for thermal deformation during the welding process, thereby improving the accuracy of the welding trajectory.

[0045] The heat deformation prediction module first calculates the welding heat input in real time based on parameters such as the current welding current, welding voltage, welding speed, and workpiece thickness. Welding heat input is a physical quantity that measures the amount of heat input to the workpiece during the welding process. Its calculation formula is usually welding current multiplied by welding voltage and then divided by welding speed.

[0046] Then, the thermal deformation prediction module calculates the temperature field distribution of the workpiece under the current heat input based on a preset heat conduction model. The heat conduction model describes the heat transfer law inside the workpiece and can be established using methods such as finite element analysis or analytical solutions. Through this model, the temperature value of each point on the workpiece at any given time can be calculated.

[0047] Next, the thermal deformation prediction module predicts the thermal deformation trend of the workpiece during the welding process based on the calculated temperature field distribution. Due to the uneven temperature distribution, the workpiece will experience uneven thermal expansion, leading to deformation. This module predicts the possible direction and magnitude of deformation of the workpiece by analyzing the temperature gradient and constraint conditions.

[0048] Finally, the thermal deformation prediction module uses the predicted thermal deformation trend to feedforward correct the welding trajectory of the welding robot arm 120. Feedforward correction refers to applying compensation in advance based on the prediction results before the disturbance occurs. Specifically, a reverse deformation compensation amount can be preset in the welding trajectory, that is, the welding torch is shifted a certain distance in the opposite direction of deformation in advance to counteract the effects of thermal deformation generated during the welding process.

[0049] This technical solution establishes a thermal deformation prediction model, which enables proactive feedforward compensation for welding thermal deformation. It effectively solves the problem of weld offset caused by heat accumulation in long welds and thin plate welding, improves the accuracy of welding trajectory and the quality of weld formation, and is especially suitable for automated welding of large structural components.

[0050] In one embodiment, optionally, the thermal deformation prediction module is further configured to: Based on the lateral and height deviations output by the weld deviation calculation module, the welding trajectory of the welding robot arm is corrected. The feedforward correction amount and the feedback correction amount are superimposed to generate the final welding trajectory control command.

[0051] The thermal deformation prediction module, based on the aforementioned feedforward correction function, also integrates a feedback correction function to form a complete composite control loop. The thermal deformation prediction module first receives the lateral and height deviation signals output by the weld deviation calculation module. These deviation signals reflect the real-time error between the current position of the welding torch and the actual centerline of the weld.

[0052] Based on these deviation signals, the thermal deformation prediction module calculates a feedback correction amount using feedback control algorithms such as a proportional-integral-derivative controller to eliminate the current tracking error. The characteristic of feedback correction is its ability to correct existing deviations in real time, but it exhibits a certain degree of lag.

[0053] Finally, the thermal deformation prediction module superimposes the feedforward correction amount obtained from the thermal deformation prediction with the feedback correction amount calculated from the real-time deviation to generate the final welding trajectory control command. This command simultaneously includes predictive compensation and real-time correction, achieving composite control of feedforward and feedback.

[0054] This technical solution achieves composite control of the welding trajectory by combining feedforward compensation and feedback correction. Feedforward compensation can anticipate regular disturbances such as thermal deformation, while feedback correction can eliminate random errors in real time. The complementary advantages of the two significantly improve the tracking accuracy and dynamic response performance of the welding trajectory, ensuring welding quality under complex working conditions.

[0055] In one embodiment, optionally, the omnidirectional moving platform 110 includes a magnetic adsorption component disposed at the bottom of the omnidirectional moving platform 110, for providing adsorption force when the omnidirectional moving platform 110 moves on the surface of a ferromagnetic workpiece. The magnetic adsorption assembly includes multiple permanent magnets and a lifting drive mechanism disposed below each of the permanent magnets. The lifting drive mechanism is used to drive each of the permanent magnets to switch between an adsorption position and a detachment position.

[0056] The omnidirectional moving platform 110, in addition to having omnidirectional wheels or Mecanum wheels, also includes a magnetic adsorption assembly. The magnetic adsorption assembly is located at the bottom of the omnidirectional moving platform 110, and its function is to provide an adsorption force perpendicular to the surface of the workpiece when the platform moves on the surface of the ferromagnetic workpiece, thereby increasing the friction between the platform and the workpiece and preventing the platform from slipping or tipping over on inclined or vertical workpiece surfaces.

[0057] The magnetic adsorption assembly includes multiple permanent magnets and a lifting drive mechanism. The permanent magnets can be made of high-performance permanent magnet materials such as neodymium iron boron, which can provide strong magnetic attraction. The lifting drive mechanism can be a miniature electric actuator or cylinder, located below each permanent magnet, to drive the permanent magnet to move up and down in the vertical direction.

[0058] When the lifting drive mechanism lowers the permanent magnet to the adsorption position, the bottom surface of the permanent magnet approaches or adheres to the workpiece surface, generating a strong adsorption force that firmly attaches the platform to the workpiece. When it is necessary to move the platform, the lifting drive mechanism raises the permanent magnet to the detachment position, moving the permanent magnet away from the workpiece surface, reducing or eliminating the adsorption force, and the platform can then move freely.

[0059] This technical solution, by setting a controllable magnetic adsorption component at the bottom of the mobile platform, enables the welding equipment to move and operate stably on the surface of ferromagnetic workpieces such as inclined planes, vertical planes, and even overhead planes, greatly expanding the application range of welding equipment and solving the problem of difficult stable positioning in on-site welding of large structural components.

[0060] In one embodiment, optionally, a visual positioning module is further included, the visual positioning module being used for: Before welding begins, images of the weld start and end points are acquired to identify the macroscopic direction of the weld. Based on the macroscopic orientation of the weld seam, the initial welding trajectory of the welding robot arm is planned; During the welding process, the vision positioning module does not participate in weld seam tracking, but is only used to monitor the safety status of the welding area.

[0061] The vision positioning module can be an industrial camera and its associated image processing unit installed on the end effector of the welding robotic arm 120 or on the omnidirectional moving platform 110. This vision positioning module is mainly used for macroscopic positioning before welding and safety monitoring during the welding process, and does not participate in real-time weld seam tracking.

[0062] Before welding begins, the vision positioning module first acquires an image containing the start and end points of the weld, and then uses image processing algorithms to identify the macroscopic direction of the weld, such as extracting the centerline path of the weld through edge detection and line fitting algorithms. Based on the identified weld direction, the vision positioning module then plans the initial welding trajectory of the welding robot arm 120, providing an initial reference path for subsequent fine tracking.

[0063] During the welding process, this vision positioning module does not participate in weld seam tracking to avoid interference from arc light on the vision sensor. At this time, the vision positioning module is only used to monitor the safety status of the welding area; for example, it uses image recognition to detect whether personnel or objects have entered the danger zone and issues an alarm signal if necessary.

[0064] This technical solution, by limiting the function of the vision positioning module to initial positioning before welding and safety monitoring during welding, avoids the problem of its failure due to arc light interference during welding. At the same time, it utilizes the large field of view of vision to achieve rapid initial positioning and effective safety protection, forming a good functional complement to the arc-based tracking method, thereby improving the overall reliability and safety of the system.

[0065] In one embodiment, optionally, the intelligent control unit 140 further includes an adaptive arc-starting module, the adaptive arc-starting module being used for: At the start of welding, the welding torch is brought close to the workpiece surface, and the arc voltage signal is detected in real time. When the arc voltage signal is detected to have arc ignition characteristics, the current welding torch position is locked as the welding start point. The welding trajectory is automatically generated based on the welding start point and the preset weld direction.

[0066] The adaptive arc initiation module can be the algorithm module responsible for the welding start-up process in the intelligent control unit 140. The adaptive arc initiation module is used to automatically determine the welding start point and automatically generate the welding trajectory without manual teaching.

[0067] At the start of welding, the adaptive arc initiation module first controls the welding robotic arm 120 to slowly approach the welding torch towards the workpiece surface. During the approach, the adaptive arc initiation module monitors the arc voltage signal collected by the welding power source 130 in real time. When the distance between the welding torch and the workpiece is close enough, the arc is ignited, and the arc voltage signal undergoes a sudden change, exhibiting specific arc initiation characteristics, such as a sharp drop in voltage value from open-circuit voltage to arc voltage.

[0068] Upon detecting the arc-initiation feature, the adaptive arc-initiation module immediately locks onto the current welding torch position and uses it as the welding start point. Then, based on this welding start point and preset weld direction information, the adaptive arc-initiation module automatically generates a complete welding trajectory. The preset weld direction can be derived from the recognition results of the vision positioning module or from pre-input program instructions.

[0069] This technical solution achieves automatic locking of the welding start point and automatic generation of the welding trajectory through an adaptive arc initiation module, which simplifies the preparation work before welding, reduces the reliance on the operator's teaching skills, and improves the automation level and operation efficiency of welding equipment. It is especially suitable for automated welding scenarios with multiple varieties and small batches.

[0070] To enable those skilled in the art to better understand this solution, this application also provides a preferred embodiment.

[0071] This embodiment provides a fully autonomous welding equipment adapted to unstructured environments, the overall structure of which is as follows: Figure 1 As shown, it consists of four main parts: an omnidirectional mobile platform, a six-degree-of-freedom welding robotic arm, a digital welding power supply, and an intelligent control unit.

[0072] The omnidirectional mobile platform utilizes a four-wheel independently driven Mecanum wheel chassis, enabling omnidirectional movement within a plane. At each of the four corners of the platform's bottom is a magnetic adsorption assembly, each consisting of a neodymium iron boron permanent magnet and a miniature electric actuator. When the equipment needs to operate on inclined or vertical ferromagnetic workpiece surfaces, the electric actuator pushes the permanent magnet down to an adsorption position close to the workpiece surface, generating a strong magnetic force to firmly attach the platform to the workpiece. When the platform needs to move freely, the electric actuator raises the permanent magnet to a detached position, eliminating the adsorption force. Through independent lifting control of the four magnetic adsorption assemblies, the platform can smoothly transition between flat, inclined, and vertical surfaces.

[0073] A six-degree-of-freedom welding robotic arm is mounted above the platform, with a welding torch integrated at its end. The welding torch is connected to the welding power source via cables. The welding power source is a digital inverter power supply with high-frequency sampling capabilities, capable of acquiring arc voltage and arc current signals during the welding process in real time at a rate of tens of thousands of times per second, and transmitting these signals to the intelligent control unit.

[0074] The intelligent control unit is the core of the entire equipment, integrating an arc feature extraction module, a weld deviation calculation module, a process parameter adjustment module, a thermal deformation prediction module, a visual positioning module, and an adaptive arc initiation module.

[0075] Before welding begins, the vision positioning module first captures images of the workpiece surface using an industrial camera mounted at the end of the robotic arm, identifying the start and end points of the weld seam, obtaining the macroscopic direction of the weld seam, and planning an initial welding trajectory accordingly. Subsequently, the adaptive arc ignition module controls the robotic arm to move the welding torch towards the weld seam start point, continuously monitoring the arc voltage signal as it approaches the workpiece surface. When the welding torch is close enough to the workpiece, the arc is ignited, exhibiting a sudden drop in arc voltage characteristic. At this point, the module immediately locks the current welding torch position as the welding start point and automatically generates a complete welding trajectory based on this start point and the preset weld seam direction.

[0076] During welding, the arc feature extraction module performs a short-time Fourier transform on the arc voltage signal acquired in real time from the welding power source to obtain a spectrum of the arc's inherent oscillation frequency over time, and extracts the fundamental frequency component and harmonic components. The weld deviation calculation module determines the welding torch's offset direction based on the changing patterns of the fundamental frequency and harmonic components: when the fundamental frequency component increases and the harmonic components decrease, the welding torch is determined to shift to the left; when the fundamental frequency component decreases and the harmonic components increase, the welding torch is determined to shift to the right. Based on the magnitude of the changes in the fundamental frequency and harmonic components, the module calculates the specific lateral deviation value using a preset calibration curve, and adjusts the robotic arm's welding torch posture in real time accordingly to ensure the welding torch is always aligned with the weld centerline.

[0077] Meanwhile, the process parameter adjustment module performs bandpass filtering on the arc current signal, extracts the signal components within the molten pool oscillation frequency band, and then obtains the instantaneous phase of the molten pool oscillation through Hilbert transform. The module compares the instantaneous phase with a preset penetration state phase range: when the phase is in the incomplete penetration range, it increases the pulse peak value of the welding current or extends the pulse base time to increase heat input; when the phase is in the over-penetration range, it decreases the pulse peak value of the welding current or shortens the pulse base time to reduce heat input. This achieves real-time closed-loop control of the penetration state.

[0078] During the welding process, the thermal deformation prediction module operates synchronously. This module calculates the welding heat input in real time based on the current welding current, welding voltage, welding speed, and workpiece thickness, and calculates the internal temperature field distribution of the workpiece based on a preset heat conduction model. After predicting the thermal deformation trend of the workpiece based on the temperature field distribution, the module presets an anti-deformation compensation amount in the welding trajectory, using feedforward correction to adjust the robotic arm's motion path. Simultaneously, the module also receives the lateral and height deviations output by the weld deviation calculation module, calculates the feedback correction amount through a proportional-integral-derivative controller, and superimposes the feedforward and feedback correction amounts to generate the final welding trajectory control command, achieving composite control of feedforward and feedback.

[0079] Throughout the welding process, the vision positioning module does not participate in weld seam tracking, but only uses the camera to continuously monitor the safety status of the welding area. When it detects personnel or objects entering the danger zone, it immediately issues an alarm signal.

[0080] The modules mentioned above work together until all welds are completed.

[0081] The core invention of this solution lies in the fact that by using the physical characteristics of the welding arc itself as a sensing source, a closed-loop system for weld seam tracking and penetration control without the need for external sensors is constructed.

[0082] Specifically, this solution extracts the inherent oscillation frequency of the arc from the arc voltage signal, and uses the variation law of its fundamental frequency and harmonic components to determine the lateral offset direction and magnitude of the welding torch relative to the weld centerline, thus achieving weld tracking. Simultaneously, it extracts the molten pool oscillation phase from the arc current signal, and uses the phase interval to determine the penetration state and adjust the current waveform parameters in real time, achieving closed-loop penetration control. Both functions are based on the electrical signals generated by the welding process itself and do not rely on any external sensing devices.

[0083] Current technological approaches: Existing mobile welding equipment relies almost entirely on external sensors for weld seam tracking, including vision sensors, structured light sensors, lidar, and arc sensors. While arc sensors exist, current solutions only utilize arc signals for simple amplitude or mean analysis, failing to delve into the intrinsic relationship between the arc's inherent oscillation frequency and the welding torch's offset, nor do they utilize the molten pool oscillation phase for real-time assessment of penetration status.

[0084] The technological breakthrough of this solution lies in its pioneering use of the variation law of the fundamental frequency and harmonic components of the inherent oscillation frequency of the electric arc as a characterizing quantity of weld deviation, establishing a mapping relationship between frequency characteristics and spatial deviation. This represents a deep exploration and utilization of the physical characteristics of the electric arc. Simultaneously, by introducing the molten pool oscillation phase into the determination of the penetration state, a leap from post-weld detection to real-time control during welding is achieved.

[0085] The difference between this solution and existing technologies is not simply feature superposition or parameter optimization, but a fundamental shift in the technological approach, moving from external sensor perception to self-sensing during the welding process. This shift brings three significant and substantial improvements: First, it completely avoids the interference of arc light and spatter on the sensor, fundamentally improving tracking reliability; second, it eliminates expensive sensors such as vision and structured light, significantly reducing system costs; and third, it utilizes the high-frequency characteristics of the arc signal to achieve millisecond-level real-time response, far exceeding the processing frame rate of vision sensors.

[0086] Furthermore, this solution is the first to propose establishing a quantitative mapping relationship between the harmonic characteristics of its inherent oscillation frequency and the welding torch offset, and to use the molten pool oscillation phase for closed-loop control of penetration, which is not an obvious combination. This solution requires a deep understanding of the coupling relationship between arc physics, molten pool dynamics, and welding process parameters, and possesses a high technical threshold and a high degree of creativity.

[0087] The technical effects include, but are not limited to, the following: This preferred embodiment achieves the following beneficial effects through the above technical solution: First, by combining an omnidirectional mobile platform with a controllable magnetic adsorption component, the welding equipment can move and operate stably in various unstructured environments such as planes, slopes, and vertical surfaces. This solves the problem that traditional mobile platforms have difficulty in stably attaching to curved and vertical workpieces, and significantly expands the applicability of the equipment.

[0088] Secondly, by utilizing the inherent oscillation frequency of the electric arc as the sensing source for weld seam tracking, there is no need to install additional vision or structured light sensors, which fundamentally avoids the interference of arc light and spatter on the tracking signal, reduces the system hardware cost, and at the same time ensures the reliability and accuracy of tracking in strong interference environments.

[0089] Third, by judging the penetration state in real time through the oscillation phase of the molten pool and adjusting the current waveform parameters in a closed loop, the welding process parameters are optimized online adaptively, which effectively prevents defects such as incomplete penetration and burn-through, and significantly improves the quality consistency of weld formation.

[0090] Fourth, the composite control strategy of thermal deformation prediction, feedforward compensation, and real-time deviation feedback correction effectively offsets the trajectory deviation caused by welding heat accumulation, improves the trajectory accuracy of long welds and thin plate welding, and ensures welding quality under complex working conditions.

[0091] Fifth, the adaptive arc-starting module enables automatic locking of the welding start point and automatic generation of the welding trajectory, which simplifies pre-welding preparation, reduces reliance on operator training skills, and improves the automation level and operational efficiency of the equipment.

[0092] Example 2 Embodiment 2 of this application provides an operation method for a fully autonomous welding equipment adapted to unstructured environments. Figure 2 This is a flowchart illustrating the operation method of the fully autonomous welding equipment adapted to unstructured environments provided in the embodiments of this application, as shown below. Figure 2 As shown, the method includes: S1 controls the omnidirectional mobile platform to autonomously move to the target welding station, obtains the start and end points of the weld seam through the visual positioning module, and plans the initial welding trajectory. S2, control the welding robotic arm to move the welding torch to the weld start point, and use the adaptive arc start module to detect arc start characteristics and lock the welding start point; S3, start welding, and collect arc voltage and arc current signals in real time; S4, perform time-frequency analysis on the arc voltage signal, extract the arc's inherent oscillation frequency and its variation characteristics, and calculate the lateral and height deviations of the welding torch tip relative to the weld centerline. S5. Based on the lateral deviation and height deviation, adjust the welding gun position of the welding robot arm in real time so that the welding gun is always aligned with the center line of the weld. S6, perform bandpass filtering and Hilbert transform on the arc current signal, extract the molten pool oscillation phase, and determine the current melting state; S7, When the penetration state deviates from the set target, adjust the output current waveform parameters of the welding power supply; S8 calculates the welding heat input in real time based on the welding current, welding voltage, welding speed and workpiece thickness, predicts the workpiece thermal deformation trend, and feeds forward to correct the welding trajectory.

[0093] Furthermore, in step S4, the time-frequency analysis includes: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. Based on the variation law of the fundamental frequency component and harmonic components, the lateral deviation value of the welding torch tip relative to the weld centerline is calculated.

[0094] The operating method provided in this application can achieve the beneficial effects corresponding to the various functional modules of the above embodiments. To avoid repetition, it will not be described again here.

[0095] Example 3 This application also provides an electronic device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the operation method described above.

[0096] It should be noted that the system provided in this application embodiment can perform the operation method described above and has corresponding beneficial effects.

[0097] Example 4 This application also provides a readable storage medium storing a program or instructions. When the program or instructions are executed by a processor, they implement the various processes of the above-described operation method embodiments and achieve the same technical effect. To avoid repetition, they will not be described again here.

[0098] It should be noted that, in this document, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or system that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such a process, method, article, or system. Without further limitations, an element defined by the phrase "comprising one..." does not exclude the presence of other identical elements in the process, method, article, or system that includes that element. Furthermore, it should be noted that the scope of the methods and systems in the embodiments of this application is not limited to performing functions in the order shown or discussed, but may also include performing functions substantially simultaneously or in reverse order, depending on the functions involved. For example, the described methods may be performed in a different order than described, and various steps may be added, omitted, or combined. Additionally, features described with reference to some examples may be combined in other examples.

[0099] Through the above description of the embodiments, those skilled in the art can clearly understand that the methods of the above embodiments can be implemented by means of software plus necessary general-purpose hardware platforms, and of course, they can also be implemented by hardware, but in many cases the former is a better implementation method. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, can be embodied in the form of a computer software product. This computer software product is stored in a storage medium (such as ROM / RAM, disk, optical disk), and includes several instructions to cause a terminal (which may be a mobile phone, computer, server, or network device, etc.) to execute the methods described in the various embodiments of this application.

[0100] The embodiments of this application have been described above with reference to the accompanying drawings. However, this application is not limited to the specific embodiments described above, which are merely illustrative and not restrictive. Those skilled in the art, under the guidance of this application, can make many modifications without departing from the spirit and scope of the claims, all of which fall within the protection scope of this application.

[0101] The above description is merely a preferred embodiment and the technical principles employed in this application. This application is not limited to the specific embodiments described herein, and various obvious changes, readjustments, and substitutions that can be made by those skilled in the art will not depart from the scope of protection of this application. Therefore, although this application has been described in detail through the above embodiments, this application is not limited to the above embodiments, and may include more other equivalent embodiments without departing from the concept of this application, the scope of which is determined by the scope of the claims.

Claims

1. A fully autonomous welding equipment adapted to unstructured environments, characterized in that, include: Omnidirectional mobile platform, welding robotic arm, welding power supply, intelligent control unit; The omnidirectional mobile platform is used for autonomous movement in unstructured environments and includes multiple independently driven omnidirectional wheels or Mecanum wheels. The welding robotic arm is mounted on the omnidirectional moving platform and has a welding torch integrated at its end. The welding torch is electrically connected to the welding power source. The welding power source is used to provide welding current and welding voltage to the welding torch, and to collect arc voltage and arc current signals in real time during the welding process. The intelligent control unit is communicatively connected to the omnidirectional mobile platform, the welding robotic arm and the welding power source, and includes an arc feature extraction module, a weld deviation calculation module and a process parameter adjustment module. The arc feature extraction module is used to perform time-frequency analysis on the arc voltage signal and arc current signal to extract the arc's inherent oscillation frequency and its variation characteristics. The weld deviation calculation module is used to calculate the lateral and height deviations of the welding torch tip relative to the weld centerline based on the variation characteristics of the inherent oscillation frequency of the electric arc. The process parameter adjustment module is used to extract the molten pool oscillation phase based on the variation characteristics of the inherent oscillation frequency of the electric arc, and to determine the current penetration state based on the molten pool oscillation phase. When the penetration state deviates from the set target, the output current waveform parameters of the welding power supply are adjusted.

2. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, The weld deviation calculation module is specifically used for: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. When the fundamental frequency component increases and the harmonic component decreases, it is determined that the welding torch shifts to the left of the weld centerline. When the fundamental frequency component decreases and the harmonic component increases, it is determined that the welding torch is shifted to the right of the weld centerline. Based on the changes in the fundamental frequency component and harmonic components, the lateral deviation of the welding torch tip relative to the weld centerline is calculated.

3. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, The process parameter adjustment module is specifically used for: The arc current signal is bandpass filtered to extract the signal components within the molten pool oscillation frequency band; Perform a Hilbert transform on the signal components to obtain the instantaneous phase of the molten pool oscillation; The instantaneous phase is compared with a preset phase interval of the melt penetration state; When the instantaneous phase is in the incomplete penetration phase range, increase the pulse peak value of the welding current or prolong the pulse base value time; When the instantaneous phase is in the overpenetration phase range, reduce the pulse peak value of the welding current or shorten the pulse base value time.

4. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, It also includes a thermal deformation prediction module, which is used for: The welding heat input is calculated in real time based on the welding current, welding voltage, welding speed, and workpiece thickness. Based on a preset heat conduction model, the temperature field distribution of the workpiece under the current heat input is calculated. Based on the temperature field distribution, predict the thermal deformation trend of the workpiece during the welding process; Based on the thermal deformation trend, the welding trajectory of the welding robot arm is feedforward and the anti-deformation compensation amount is preset.

5. The fully autonomous welding equipment adapted to unstructured environments according to claim 4, characterized in that, The thermal deformation prediction module is also used for: Based on the lateral and height deviations output by the weld deviation calculation module, the welding trajectory of the welding robot arm is corrected. The feedforward correction amount and the feedback correction amount are superimposed to generate the final welding trajectory control command.

6. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, The omnidirectional moving platform includes a magnetic adsorption component, which is disposed at the bottom of the omnidirectional moving platform and is used to provide adsorption force when the omnidirectional moving platform moves on the surface of a ferromagnetic workpiece. The magnetic adsorption assembly includes multiple permanent magnets and a lifting drive mechanism disposed below each of the permanent magnets. The lifting drive mechanism is used to drive each of the permanent magnets to switch between an adsorption position and a detachment position.

7. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, It also includes a visual positioning module, which is used for: Before welding begins, images of the weld start and end points are acquired to identify the macroscopic direction of the weld. Based on the macroscopic orientation of the weld seam, the initial welding trajectory of the welding robot arm is planned; During the welding process, the vision positioning module does not participate in weld seam tracking, but is only used to monitor the safety status of the welding area.

8. The fully autonomous welding equipment adapted to unstructured environments according to claim 1, characterized in that, The intelligent control unit further includes an adaptive arc-starting module, which is used for: At the start of welding, the welding torch is brought close to the workpiece surface, and the arc voltage signal is detected in real time. When the arc voltage signal is detected to have arc ignition characteristics, the current welding torch position is locked as the welding start point. The welding trajectory is automatically generated based on the welding start point and the preset weld direction.

9. A method for weld seam tracking and process optimization of fully autonomous welding equipment adapted to unstructured environments, characterized in that, Using the welding equipment according to any one of claims 1-8, the method comprises the following steps: S1 controls the omnidirectional mobile platform to autonomously move to the target welding station, obtains the start and end points of the weld seam through the visual positioning module, and plans the initial welding trajectory. S2, control the welding robotic arm to move the welding torch to the weld start point, and use the adaptive arc start module to detect arc start characteristics and lock the welding start point; S3, start welding, and collect arc voltage and arc current signals in real time; S4, perform time-frequency analysis on the arc voltage signal, extract the arc's inherent oscillation frequency and its variation characteristics, and calculate the lateral and height deviations of the welding torch tip relative to the weld centerline. S5. Based on the lateral deviation and height deviation, adjust the welding gun position of the welding robot arm in real time so that the welding gun is always aligned with the center line of the weld. S6, perform bandpass filtering and Hilbert transform on the arc current signal, extract the molten pool oscillation phase, and determine the current melting state; S7, When the penetration state deviates from the set target, adjust the output current waveform parameters of the welding power supply; S8 calculates the welding heat input in real time based on the welding current, welding voltage, welding speed and workpiece thickness, predicts the workpiece thermal deformation trend, and feeds forward to correct the welding trajectory.

10. The weld seam tracking and process optimization method according to claim 9, characterized in that, In step S4, the time-frequency analysis includes: Perform a short-time Fourier transform on the arc voltage signal to obtain the spectrum of the arc's natural oscillation frequency as a function of time; Extract the fundamental frequency component and harmonic components of the arc's inherent oscillation frequency from the spectrum. Based on the variation law of the fundamental frequency component and harmonic components, the lateral deviation value of the welding torch tip relative to the weld centerline is calculated.

Images