Laser welding closed-loop control method based on visual guidance and molten pool monitoring

Abstract

The invention discloses a laser welding closed-loop control method and device based on visual guidance and molten pool monitoring, electronic equipment and a readable storage medium, and relates to the technical field of welding. The method comprises the steps that molten pool state scanning data and welding process radiation data are obtained; and positioning a welding product of welding equipment through a preset welding track obtained based on pre-welding product positioning scanning data, obtaining a current welding state through the molten pool state scanning data and the welding process radiation data, and adjusting welding parameters of the welding equipment on line through the current welding state. The method has the effects that the positioning precision of the pre-welding track is improved, and welding parameters are adjusted in time through analysis of welding pool state scanning data in combination with welding process radiation data.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of welding, in particular to a laser welding closed-loop control method based on visual guidance and molten pool monitoring. BACKGROUND

[0002] In the laser sealing welding process of vertical square shell batteries, in order to solve the assembly gap problem between the shell and the cover plate, a pre-point welding process is usually added to realize the fixture fixing, which increases the production process and time cost. In contrast, in the laser sealing welding of inverted square shell batteries, the shell can be stably pressed on the cover plate by its own gravity, effectively reducing the process and cost caused by additional pre-point welding. However, in the laser sealing welding of inverted square shell batteries, the welding path deviation is often caused by insufficient positioning accuracy of the clamp, and the process parameters cannot be adjusted in real time. SUMMARY

[0003] The present application aims to at least solve one of the technical problems in the prior art, and provides a laser welding closed-loop control method, device, electronic equipment and readable storage medium based on visual guidance and molten pool monitoring, which aims to improve the positioning accuracy of the pre-welding trajectory, and make timely adjustment of the welding parameters through the analysis of the molten pool state scanning data combined with the welding process radiation data.

[0004] In a first aspect, the embodiments of the present application provide a laser welding closed-loop control method based on visual guidance and molten pool monitoring, comprising: obtaining molten pool state scanning data and welding process radiation data; positioning the welding product of the welding equipment through the preset welding trajectory obtained based on the pre-welding product positioning scanning data, and obtaining the current welding state through the molten pool state scanning data and the welding process radiation data, and adjusting the welding parameters of the welding equipment in real time through the current welding state.

[0005] According to the technical scheme of the embodiments of the present application, at least the following beneficial effects are obtained: the preset welding trajectory obtained based on the pre-welding product positioning scanning data can improve the positioning accuracy of the pre-welding trajectory, while ensuring the welding accuracy; by obtaining the molten pool state scanning data and the welding process radiation data in real time during the welding process, the current welding state, i.e. whether the welding quality meets the standard, can be judged, so that the welding parameters of the welding equipment are adjusted in real time through the current welding state, and the stability of the welding quality and the product qualification rate are improved.

[0006] According to some embodiments of the present application, the method further comprises: controlling the welding position of the welding equipment through the molten pool state scanning data and the preset welding trajectory.

[0007] According to some embodiments of the present application, the controlling the welding position of the welding equipment based on the molten pool state scanning data and the preset welding track comprises: The predicted welding track is obtained based on the molten pool state scanning data, and the welding position of the welding equipment is controlled based on the predicted welding track and the preset welding track.

[0008] According to some embodiments of the present application, the current welding state comprises the shape and position of the current welding molten pool.

[0009] According to some embodiments of the present application, the current welding state further comprises welding heat input parameters and weld forming area. The current welding state is obtained based on the molten pool state scanning data and the welding process radiation data, comprising: The shape and position of the current welding molten pool are determined based on the molten pool state scanning data and the welding process radiation data, and the welding heat input and the weld forming area are determined based on the shape and position of the current welding molten pool.

[0010] According to some embodiments of the present application, the molten pool state scanning data is obtained by anti-interference scanning of the welding product by a 3D scanning device, and the anti-interference scanning comprises real-time filtering and focal point scanning.

[0011] According to some embodiments of the present application, the preset welding track is obtained by pre-scanning the to-be-welded area of the welding product by a 3D scanning device to obtain the geometric size parameters of the to-be-welded area, and then converting the geometric size parameters of the to-be-welded area into welding point coordinates, and then performing coordinate conversion on the welding point coordinates.

[0012] In a second aspect, the embodiments of the present application provide a running control device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, and the processor executes the program to implement the control method of the first aspect.

[0013] In a third aspect, the embodiments of the present application provide an electronic device comprising the running control device of the second aspect.

[0014] In a fourth aspect, the embodiments of the present application provide a computer readable storage medium, which stores computer executable instructions, and the computer executable instructions are used to make a computer execute the control method of the first aspect.

[0015] Other features and advantages of the present application will be set forth in the description that follows, and in part will be apparent from the description, or can be learned by practice of the application. The purposes and other advantages of the present application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings. BRIEF DESCRIPTION OF DRAWINGS

[0016] The accompanying drawings are included to provide a further understanding of the technical solution of the present application, and constitute a part of the specification, and are used together with the embodiments of the present application to explain the technical solution of the present application, and do not constitute a limitation on the technical solution of the present application.

[0017] The present application will be further described below in conjunction with the drawings and embodiments; Figure 1 is a flow chart of a control method provided by an embodiment of the present application; Figure 2 is a flow chart of a control method provided by another embodiment of the present application; Figure 3 is a flow chart of a control method provided by another embodiment of the present application; Figure 4 is a flow chart of a control method provided by another embodiment of the present application; Figure 5 is a schematic diagram of a running control device for executing the control method provided by an embodiment of the present application. DETAILED DESCRIPTION

[0018] This part will describe the specific embodiments of the present application in detail, the preferred embodiments of the present application are shown in the drawings, the role of the drawings is to supplement the description of the text part of the description, so that people can intuitively and visually understand each technical feature and the overall technical solution of the present application, but it cannot be understood as a limitation on the protection scope of the present application.

[0019] In the description of the present application, it should be understood that the orientation description, such as the orientation or position relationship indicated by up, down, front, back, left, right, etc. is based on the orientation or position relationship shown in the drawings, and is only for the convenience of describing the present application and simplifying the description, and therefore cannot be understood as indicating or implying that the device or element must have a specific orientation, be constructed and operated in a specific orientation, and therefore cannot be understood as a limitation on the present application.

[0020] In the description of the present application, the meaning of several is one or more, the meaning of multiple is more than two, greater than, less than, more than, etc. are understood as not including the number, above, below, etc. are understood as including the number. If it is described as first, second, it is only for the purpose of distinguishing technical features, and cannot be understood as indicating or implying relative importance or implicitly indicating the number of indicated technical features or implicitly indicating the sequence of indicated technical features.

[0021] In the description of this application, unless otherwise expressly defined, terms such as "setup," "installation," and "connection" should be interpreted broadly, and those skilled in the art can reasonably determine the specific meaning of the above terms in this application in conjunction with the specific content of the technical solution.

[0022] The various embodiments of the control method of this application will be further described below with reference to the accompanying drawings.

[0023] like Figure 1 As shown, Figure 1 This is a flowchart of a control method provided in one embodiment of this application. The control method may include, but is not limited to, steps S110 and S120.

[0024] Step S110: Obtain molten pool state scan data and welding process radiation data; Step S120: Position the welding product of the welding equipment using the preset welding trajectory obtained based on the pre-welding product positioning scanning data, and obtain the current welding state using the molten pool state scanning data and welding process radiation data, and adjust the welding parameters of the welding equipment online based on the current welding state.

[0025] It is understood that the embodiments of this application can be applied to welding scenarios of different types of products using various types of welding machines. For example, this application uses the laser sealing welding scenario of an inverted square-shell battery cell as an example to illustrate the specific implementation of this application.

[0026] It is understandable that both pre-weld product positioning scan data and molten pool state scan data are obtained by scanning the welded product. The difference is that pre-weld product positioning scan data is the welding scan data obtained by scanning the welded product before welding to determine the precise positioning of the welded product at the work station, while molten pool state scan data is the welding scan data obtained by scanning the welded product in real time during welding to identify the state of the molten pool. In other words, welding scan data includes pre-weld product positioning scan data and molten pool state scan data.

[0027] It can be understood that in the laser sealing welding process of the inverted square shell battery cell, the acquisition of the welding scanning data depends on high-precision optical scanning equipment. For example, the welding scanning data can be acquired by a galvanometer scanning system. The galvanometer system accurately controls the deflection angle of the laser beam through a high-speed rotating mirror, thereby realizing rapid and accurate positioning of the welding path. When selecting a specific scanning device for welding scanning data, the scanning accuracy, scanning speed, working distance, and collaborative control capability with the laser, etc. need to be considered. For example, in the vertical packaging of the square shell battery cell, the weld is usually located at the butt joint edge of the shell and the cover plate. The galvanometer system can realize tracking of straight lines, circular arcs, or complex curves through a pre-set two-dimensional or three-dimensional path planning. In addition, a scanning device integrated with a visual positioning module can be selected. The image of the weld position can be acquired by a CCD camera, and the workpiece assembly error can be compensated through image processing technology, further improving the accuracy of the scanning data. In addition, in addition to the galvanometer system, the scanning device of the welding scanning data can also be an optical profile scanner, such as a 3D profile instrument, which can reconstruct the three-dimensional geometric features of the weld by emitting a laser line beam and analyzing the deformed profile.

[0028] In some embodiments, the welding scanning data can include a multi-dimensional information set. The welding scanning data includes geometric trajectory data, which includes a two-dimensional or three-dimensional coordinate sequence of the welding path, a motion speed curve, an acceleration distribution, etc. The geometric trajectory data is directly related to the actual action position of the laser beam, and if there is a trajectory deviation, it may cause defects such as welding deviation and incomplete fusion. The welding scanning data can also include dynamic response data, which records the real-time state of the scanning device during execution, such as the control voltage, feedback current, mirror deflection angle, etc. of the galvanometer motor. These data can be used to diagnose device jitter, delay, and other abnormal conditions. The welding scanning data in the welding process, i.e. the molten pool state scanning data, can also include process-dependent modulation parameters, such as the curve of the laser power changing with the path, the dynamic adjustment record of the defocusing amount, the time sequence control signal of the protective gas flow, etc. These parameters together constitute a complete welding instruction set.

[0029] In some embodiments, the molten pool state scanning data can also include process monitoring data, such as the reflected light intensity signal acquired by a coaxial photodiode, or the temperature field distribution of the molten pool area acquired by an infrared thermal imager. Such data is recorded synchronously with the scanning trajectory and can be used to analyze the welding stability.

[0030] It can be understood that the detection device of the welding process radiation data can be a photoelectric sensing device, for example, a spectrometer and a photodiode detector. The spectrometer can perform wavelength-resolved measurement on the welding plasma radiation through grating spectrometry and array detection, including a CCD spectrometer and a CMOS spectrometer, etc. The spectrometer can obtain the wavelength distribution characteristics of the radiation, which is convenient for distinguishing different radiation mechanisms. The photodiode detector can convert the light signal into an electrical signal through a semiconductor element. According to the different sensitive wave bands, it can be divided into ultraviolet enhancement type, silicon-based standard type and infrared expansion type, etc. The selection of the detection device of the welding process radiation data needs to consider the dynamic range, linear response interval and anti-electromagnetic interference ability, etc., to ensure that reliable data can be obtained in the welding environment with strong arc light and high-frequency electromagnetic noise.

[0031] In some embodiments, the welding process radiation data can include data in three dimensions of spectral characteristics, intensity characteristics and space-time characteristics. Among them, the spectral characteristic data includes parameters such as continuous spectral intensity distribution, characteristic spectral line wavelength position and half-width, spectral line relative intensity ratio, etc. For example, in the aluminum alloy welding, the 394.4 nanometer and 396.15 nanometer aluminum atomic spectral lines, the intensity ratio can reflect the plasma electron temperature, and the continuous background radiation intensity is related to the plasma density; the intensity characteristic data can include total radiation intensity, specific waveband integrated intensity and its time-varying curve, which can be directly associated with the welding stability, for example, the sudden decay of radiation intensity can indicate the collapse of the molten pool or the perforation defect; the space-time characteristic data can include the spatial distribution form of the radiation source, the expansion speed of the radiation cloud cluster, the fluctuation frequency, etc. For example, if it is found that the radiation cloud cluster tilts to the rear of the keyhole during the welding process, it indicates that the keyhole is unstable and a porosity will be generated.

[0032] In some embodiments, the welding process radiation data can include voltage signals converted from infrared thermal radiation, laser back reflection radiation and plasma radiation signals.

[0033] It can be understood that the current welding state can be a comprehensive welding quality evaluation coefficient, or a highly comprehensive quantitative characterization of dynamic changes in physical and process conditions, that is, not a single parameter, but a whole description and evaluation of the instantaneous physical properties of the laser and material interaction zone, the weld forming process, and the potential defect possibility. Specifically, the current welding state can include the thermodynamic state, that is, the size of the molten pool, the existence of the life cycle, the internal flow pattern, and the overall temperature field distribution, etc., a stable and appropriately sized molten pool is the basis for forming a good weld; it can also include plasma / plume characteristics, including its electron density, temperature, stability of spatial form, and intensity and characteristics of emission spectrum, etc., which can reflect the efficiency of laser energy absorption and conversion; it can also include the dynamic behavior of the keyhole, in the deep penetration welding mode, the existence of the keyhole is crucial, its stability (including the fluctuation of the hole wall, necking phenomenon, periodic opening and closing) determines the consistency of the weld penetration depth and whether defects such as pores and spatter will occur; it can also include the real-time geometric forming quality of the weld, including the width of the weld, whether the reinforcement is uniform, and whether there are early signs of forming defects such as undercut, hump, and collapse; it can also include the process stability state, which is used to characterize whether the current welding process is in a stable and predictable steady state or a non-steady state transition period that is deviating from the normal parameters and may soon produce defects.

[0034] For example, the current welding state is obtained by molten pool state scanning data and welding process radiation data, which may involve multiple steps such as signal processing, feature extraction, and state reasoning.

[0035] For the molten pool state scanning data, first process the real-time coordinate sequence of the feedback, by calculating its instantaneous deviation from the preset trajectory, it can be judged whether the laser beam accurately tracks the weld, at the same time, the instantaneous fluctuation of the scanning speed is also an indicator of process stability; in addition, the vision or photoelectric sensor integrated coaxially in the scanning system can capture visible or near-infrared images of the molten pool area, by processing these images at high speed, the geometric features of the molten pool can be extracted, these geometric parameters are a direct reflection of the thermodynamic state of the molten pool, for example, a sudden increase in the length of the molten pool may indicate that the heat input is too high or the welding speed is too slow.

[0036] For the welding process radiation data, the plasma emission spectrum obtained by the spectrometer can be analyzed to calculate the electron temperature and electron density of the plasma, these parameters are related to the laser energy absorption rate and the stability of the keyhole; the radiation intensity and its fluctuation of a specific waveband detected by the photodiode can be used to indicate the process stability, for example, a sharp rise in ultraviolet radiation intensity is usually accompanied by the generation of metal spatter; in addition, a high-speed camera combined with a narrow-band filter can capture the two-dimensional form of the plasma / plume, and its contraction, expansion or jitter all indicate the dynamic behavior of the keyhole and the molten pool.

[0037] Finally, all these extracted multi-level features including geometric, spectral, intensity, and morphological ones are input into a state assessment model, which can be based on a predefined rule base or a classifier / regressor trained by machine learning algorithms. By fusing all these heterogeneous features, a comprehensive quantitative judgment about the current welding state is output.

[0038] It can be understood that after determining the current welding state, corresponding welding parameter adjustment will be triggered to achieve active intervention and optimization of the welding process, wherein the adjustable parameters include laser power, scanning speed, pulse characteristics (for pulse welding), defocusing amount, and protective gas flow, etc.

[0039] For example, when the state assessment indicates insufficient heat input, the laser power can be increased by a predetermined step or the scanning speed can be appropriately reduced to increase the energy input per unit length, and if an overheating state is diagnosed, the opposite operation can be performed, i.e., reducing the power or increasing the speed. For pulse welding, the pulse frequency, duty cycle, or pulse waveform can also be adjusted, for example, when welding thin plates or heat-sensitive materials, higher frequency and lower base power can reduce heat accumulation.

[0040] When the keyhole is unstable as judged by the radiation data or scanning data, the defocusing amount can be adjusted, i.e., the laser focal position is fine-tuned by a small distance inward or outward from the workpiece surface, which can change the spot size and energy density distribution, thereby stabilizing the keyhole dynamics. If signs of specific defects are detected, the adjustment strategy will be more targeted, for example, when the ultraviolet band radiation signal frequently appears as a sharp peak, indicating that spatter is intensifying, in addition to possibly reducing the power, power ramp control can also be used, i.e., slowly increasing and decreasing the power at the beginning and end of welding, or introducing beam oscillation technology. Although the adjustment of the protective gas flow and composition is relatively slow, it can also be implemented when dealing with persistent plasma transition expansion, for example, increasing the gas flow speed to suppress the plasma.

[0041] Understandably, the current welding state can be used as a standard for current welding quality. Welding trajectory and welding quality are complementary; welding quality can only be guaranteed when the welding trajectory deviation is small, and similarly, welding quality requires a relatively small welding trajectory deviation. Specifically, when the deviation between the actual scanning path of the laser beam and the pre-set ideal welding trajectory is controlled within a very small process-allowed range—for example, with positional deviation stable at the micrometer level, angular deviation less than 0.1 degrees, and speed following error less than 1 percent—the most fundamental geometric basis for forming a high-quality weld is laid. This is because a precise trajectory ensures that laser energy can be continuously, uniformly, and accurately delivered to the joint to be welded on the workpiece, avoiding a series of defects such as incomplete fusion due to welding deviation, uneven heat input due to speed fluctuations, or overheating or undercut due to improper corner handling. Thus, welding quality can be fundamentally guaranteed. Conversely, stable and excellent welding quality, characterized by uniform weld width and height, smooth and spatter-free weld surface, sufficient and stable penetration depth, and internal microstructure free of porosity and cracks, creates crucial prerequisites and a stable physical environment for maintaining small and stable welding trajectory deviations. This is because when the welding process is stable and of excellent quality, the molten pool dynamics are in equilibrium, the absorption and shielding effect of the plasma on laser energy is controllable, and the thermal deformation of the workpiece is suppressed to a minimum. This significantly reduces various interferences caused by process instability to high-precision trajectory tracking, making it easier and more accurate for the scanning system to achieve the preset ideal path. Therefore, similarly, only when the welding quality is good can the actual execution deviation of the welding trajectory be maintained within a relatively small and ideal range.

[0042] Understandably, the current welding status, comprehensively assessed by real-time monitoring and analysis of molten pool state scanning data and welding process radiation data, can serve as a real-time quantitative standard or evaluation criterion for measuring and judging the immediate quality level of the welding operation in progress. The current welding status integrates multi-dimensional information such as molten pool stability, plasma behavior regularity, appropriate heat input, and the risk level of potential defects. This allows it to transcend the limitations of post-weld inspection, providing a reliable assessment of the weld's internal quality and external shape at the very moment of weld formation. Furthermore, it can reveal the dynamic process and specific reasons for quality tending towards improvement or deterioration. For example, it can indicate whether slight overheating of the molten pool is due to a slight exceedance of energy density, or whether uneven local weld formation is caused by a delay in the scanning trajectory at a corner.

[0043] like Figure 2 As shown, Figure 2 This is a flowchart of a control method provided in another embodiment of this application; the control method described above may also include, but is not limited to, step S130.

[0044] Step S130, control the welding position of the welding equipment by the molten pool state scanning data and the preset welding track.

[0045] In some embodiments, in combination with the real-time acquired molten pool state scanning data and the preset welding track, information for guiding the welding equipment position control can be extracted, i.e. the difference between the actual welding state and the ideal process is quantified and a dynamic compensation instruction. For example, through the real-time acquired molten pool state scanning data and the preset welding track, a track tracking deviation for guiding the welding equipment position control can be obtained, including a position deviation, an angle deviation and a timing deviation. The position deviation is obtained by comparing the laser spot center coordinates fed back by the galvanometer system in real time with the target coordinates of the preset track, specifically a horizontal offset, a vertical offset and a vertical distance relative to the weld center line. The angle deviation is the change of the weld direction caused by the workpiece clamping error or thermal deformation, which can be obtained by calculating the included angle between the real-time scanning tangent and the preset track tangent. The timing deviation is the timestamp misalignment amount of the actual welding point and the corresponding point of the preset track.

[0046] In addition, through the real-time acquired molten pool state scanning data and the preset welding track, process state adaptation parameters for guiding the welding equipment position control can be obtained, for example, when the scanning data detects that the weld assembly has intermittent gaps, the system will generate modulation coefficients of laser power or scanning speed for re-planning of welding position energy input; similarly, when the plasma radiation intensity monitored coaxially abnormally increases, it may indicate that the focusing position deviates from the ideal plane, thereby generating a defocusing amount compensation signal to change the action position of the laser beam in the thickness direction of the workpiece.

[0047] Through the real-time acquired molten pool state scanning data and the preset welding track, predictive feedforward parameters for guiding the welding equipment position control can also be obtained. By performing short-term trend analysis on the continuous scanning data sequence, the track deformation trend or thermal accumulation area in the future hundreds of milliseconds can be predicted, and then a position adjustment instruction in advance can be generated, for example, actively reducing the speed and presetting the beam deflection angle before scanning to the corner area.

[0048] It can be understood that after obtaining the above information, the control of the welding device position can be realized through a closed loop or a feedforward feedback composite control system. Specifically, the control system first inputs the calculated trajectory deviation into the PID control algorithm of the galvanometer driver, which outputs the corresponding analog voltage signal to fine-tune the deflection angle of the X-axis and Y-axis galvanometer motors, thereby correcting the projection position of the laser beam in the two-dimensional plane in real time. The correction frequency is as high as several kilohertz, ensuring that the deviation can be eliminated in time during high-speed scanning. For spatial position adjustment, for example, to compensate for the defocusing phenomenon caused by uneven workpiece, the defocusing amount compensation signal can be sent to the dynamic focusing module. The module changes the relative distance between the collimating mirror and the focusing mirror to make the laser focal point always accurately fall on the workpiece surface, realizing Z-axis position control perpendicular to the workpiece plane. As for the process state adaptive parameters, a combination of look-up table and model reference adaptive control can be used, for example, when a gap is detected, the corresponding power and speed matching scheme will be called from the pre-set process parameter library, and the output power of the laser and the scanning speed of the galvanometer will be adjusted in real time to change the distribution of laser energy in the time and space dimensions. The entire control process distributes the processed information in real time to the galvanometer, laser, focusing module and other execution units through a high-speed data bus, forming a whole that works cooperatively, so as to realize accurate, dynamic and adaptive regulation and control of the welding position, ensuring that the weld formation of the inverted square shell battery laser sealing weld always follows the ideal path of high-quality and defect-free.

[0049] It can be understood that the current welding trajectory can be generated by the molten pool state scanning data, and the generation of the current welding trajectory by the molten pool state scanning data requires steps such as multi-sensor data fusion, coordinate transformation and trajectory reconstruction. The molten pool state scanning data is mainly derived from the position sensor integrated in the galvanometer scanning system and the possible auxiliary vision unit. Specifically, the encoder or capacitive position sensor inside the galvanometer system will feed back the accurate deflection angle of the X-axis and Y-axis galvanometer lenses at a very high frequency. Combined with the known optical system model, the theoretical coordinate position of the laser focal point on the two-dimensional working plane can be calculated in real time through a series of coordinate transformations to form the preliminary trajectory of the current welding trajectory. However, this is only the ideal beam position, and the actual trajectory will be affected by factors such as beam transmission delay, mechanical vibration, thermal lens effect and field curvature distortion.

[0050] Therefore, supplementary data from process monitoring is introduced. For example, a coaxially integrated high-speed CMOS or CCD camera captures visible / near-infrared images of the molten pool or keyhole area, and image processing algorithms are used to identify the actual location of the energy center. This is particularly useful when workpiece assembly errors or thermal deformation cause the weld position to deviate from the expected position, providing a position compensation reference. Furthermore, a contour scanner or confocal displacement sensor based on laser triangulation can provide three-dimensional contour information of the weld, thereby generating a current welding trajectory containing height information. In this way, all these multi-source heterogeneous scanning data are fused through data fusion algorithms to ultimately output an accurate and spatiotemporally synchronized current welding trajectory. This current welding trajectory can include not only the two-dimensional planar coordinate sequence of the laser beam, but also the angular information of the beam's relative attitude to the workpiece surface, as well as its dynamic characteristics changing over time.

[0051] For example, the control system continuously and rigorously compares each data point of the current welding trajectory with the corresponding point of the preset welding trajectory, thereby obtaining multi-dimensional errors, including positional errors in geometric space (the deviation of the straight-line distance between the current point and the target point in the X and Y directions), tangent angle errors in direction (i.e., the angle between the tangential direction of the current trajectory and the tangential direction of the preset trajectory), and phase errors in the time domain (i.e., the difference between the actual time it takes for the weld point to pass through a certain position and the preset timestamp). These multi-dimensional errors are then input into the PID controller in real time. The PID controller performs calculations based on the proportional, integral, and derivative terms of the error. The proportional term is responsible for responding to the current error immediately and quickly correcting the deviation; the integral term is responsible for accumulating historical errors to eliminate steady-state errors; and the derivative term makes predictive adjustments based on the trend of error changes to suppress system oscillations and improve stability. The controller's output is converted into physical signals that drive the welding orientation actuator.

[0052] For orientation control in a two-dimensional plane, the physical signal is an analog voltage command applied to the galvanometer motor driver. This command finely adjusts the deflection angle of the galvanometer lenses, thereby directly changing the projection point of the laser beam on the workpiece and pulling it back to the preset trajectory. For three-dimensional orientation control, especially when there is workpiece unevenness or the need for dynamic adjustment of the focusing position, the calculated defocusing error is converted into a control signal for the dynamic focusing module. By changing the position of the focusing lens group, it ensures that the laser focus always falls precisely on the changing workpiece surface.

[0053] like Figure 3 As shown, Figure 3 This is a flowchart of a control method provided in another embodiment of this application; regarding the above step S130, it may include, but is not limited to, step S230.

[0054] Step S230, obtaining a predicted welding trajectory through the molten pool state scanning data, and controlling the welding position of the welding equipment based on the predicted welding trajectory and the preset welding trajectory.

[0055] For example, the control system can continuously collect and cache the current welding trajectory data sequence within a very short time window, including the position coordinates and possible height information of the laser action point over time, and then analyze the historical data using a specific prediction algorithm to identify the dynamic trend of the trajectory change. The prediction algorithm can include a Kalman filter based on a linear model or its nonlinear variant, which can handle the inherent noise in the measurement data, take into account the dynamic characteristics of the welding system through an internal state space model, and thus make optimal estimates of the laser beam position for the next few milliseconds or even longer. In addition, machine learning models such as recurrent neural networks or long short-term memory networks can also be introduced, which can learn the complex evolution patterns of trajectory deviation under specific process conditions, such as progressive drift caused by heat accumulation or the common overshoot phenomenon at geometric feature points, by training on a large amount of historical welding data, thus making more physically realistic predictions. In addition to the geometric information of the trajectory itself, the prediction model will also integrate process monitoring data, for example, when the welding process radiation data indicates that the molten pool temperature is continuously rising, it can be predicted that the resulting workpiece thermal deformation will intensify, thus being converted into a compensation amount for the future trajectory position.

[0056] It can be understood that the welding position of the welding equipment is controlled based on the difference between the predicted welding trajectory and the preset welding trajectory, which changes from passive response to deviation to active prevention of deviation. For example, the control system compares the predicted welding trajectory with the preset ideal welding trajectory in real time, which also generates a multi-dimensional error, including predicted position error, predicted angle error, and predicted timing error. The prediction error is input to the feedforward control channel, and the feedforward controller can directly calculate the control amount required to eliminate these prediction errors based on the known dynamics model of the system, for example, when the prediction model shows that overshoot will occur at the upcoming trajectory corner, the feedforward controller will send a reverse compensation signal to the galvanometer motor driver in advance, instructing the light beam to slow down and pre-set a reverse deflection angle, so as to start offsetting before the error actually occurs. The feedback controller continues to work based on the current trajectory error to handle sudden disturbances or model uncertainties that are not predicted by the model, which is perfectly complementary to the PID feedback control.

[0057] In this way, the outputs of feedforward and feedback are superimposed in the control unit to generate the final control command together, which drives the actuator, that is, for the control of the welding position, the system not only receives the correction signal based on the current position deviation, but also contains the command for the future deviation.

[0058] In the control method provided in another embodiment of the present application, the current welding state includes the shape and position of the current welding molten pool.

[0059] It can be understood that, for the determination of the position of the molten pool, through special optical path design of the coaxial vision sensor integrated on the galvanometer scanning head, the light emission image of the molten pool area excited by the laser can be continuously captured, and through real-time processing of these continuous frames of images, the edge detection algorithm and the region contour extraction technology can be used to identify the boundary of the molten pool area with brightness significantly higher than the base material. Then, by calculating the pixel coordinates of the geometric center or the center of mass of the contour in the image coordinate system, and through the coordinate transformation of the pre-calibrated camera internal and external parameters, the pixel position can be converted into the two-dimensional spatial coordinates in the global coordinate system of the workpiece, representing the position of the current molten pool relative to the welding head. At the same time, the real-time beam deflection angle data fed back by the internal encoder of the galvanometer system, after forward kinematics calculation, can also provide a position of the theoretical focal point of the laser beam, which can be cross-verified with the molten pool position measured by vision and data fusion. For example, a Kalman filter is used to optimally estimate a more reliable and stronger anti-interference molten pool center position information, thereby effectively overcoming the problem of temporary overexposure of the image caused by strong plasma flash.

[0060] For the acquisition of the shape of the molten pool, deeper features need to be extracted from the vision image and the radiation spectrum. Specifically, from the coaxial vision image, not only the molten pool boundary can be identified, but also a series of key geometric parameters describing its shape can be calculated, including the length, width, area, aspect ratio of the molten pool, and the trailing angle of the tail of the molten pool. In addition, the shape of the molten pool is not static, and the boundary fluctuation is also an important indicator of the stability of the shape. The welding process radiation data provides a supplementary and verification dimension for the shape analysis. For example, the intensity distribution of the radiation signal collected by the multi-band photodiode or spectrometer is related to the size and temperature field of the molten pool.

[0061] It can be understood that the real-time position of the molten pool is the basis for high-precision weld tracking and deviation correction. By comparing the continuously calculated molten pool center position with the preset welding trajectory, the control system can immediately generate a position correction instruction for the galvanometer, ensuring that the laser energy is always accurately applied to the center line of the weld, and fundamentally avoiding defects such as incomplete fusion and undercut caused by welding deviation. The shape of the molten pool is a representation of the stability of the welding process and the quality of the final weld shape. The size of the molten pool reflects the size and distribution of the heat input. A suddenly lengthened and widened molten pool may indicate that the heat input is too high, which has the risk of burning through or generating a large heat-affected zone. A too small or irregular molten pool may mean that the energy is insufficient, which has the risk of incomplete penetration.

[0062] In the control method provided by another embodiment of the present application, the current welding state further includes a welding heat input parameter and a weld forming area; As Figure 4 shown, Figure 4 is a flow chart of the control method provided by another embodiment of the present application; regarding the above step S120, it can include but is not limited to step S220.

[0063] Step S220, by scanning the molten pool state data and the welding process radiation data, the shape and position of the current welding molten pool are determined, and according to the shape and position of the current welding molten pool, the welding heat input and the weld forming area are determined.

[0064] It can be understood that the determination of the welding heat input is not through direct measurement of energy, but through the inference of the shape parameters of the molten pool. The size of the molten pool has a direct positive correlation with the effective line energy acting on the workpiece. A wider and longer molten pool requires a higher instantaneous heat input. The system internally pre-stores a database or an empirical formula established through a large number of basic process tests. The database relates the geometric characteristics of the molten pool such as width and length to the theoretical heat input under the combination of process parameters such as laser power and welding speed. When the real-time image processing unit calculates the width and length of the current molten pool, the control algorithm will match it with the database or substitute it into the empirical model, so as to estimate the actual effective heat input value acting on the weld under the current condition.

[0065] In addition, the welding process radiation data, especially the plasma radiation intensity from the photodiode or the specific element spectral line intensity measured by the spectrometer, can be used as an auxiliary verification for heat input estimation, because higher energy input usually produces a more active plasma with higher radiation intensity.

[0066] It can be understood that the weld forming area is essentially the cross-sectional geometry and microstructure distribution of the final weld determined by the molten pool and its subsequent solidification process. According to the obtained heat input, combined with the known thermal physical properties of the material and the current welding speed, a simplified analytical model or a simplified numerical simulation empirical relationship can be called to predict the forming characteristics of the weld. For example, the penetration depth is related to the energy density and the absorption rate of the material to the laser. Higher heat input usually indicates deeper penetration; the width of the molten pool is associated with the observed molten pool width, but due to fluid flow and solidification shrinkage, the final weld width will be slightly smaller than the maximum width of the liquid molten pool; for the cross-sectional shape of the weld, such as nail head shape, wine cup shape or relatively ideal rectangular shape, the ratio of heat input to welding speed can be roughly judged. Too high or too low ratio will lead to an undesirable cross-sectional shape. In addition, the width of the heat-affected zone can also be estimated according to the heat input and the cooling conditions of the base material, because the size of the heat input directly determines the range of the area whose peak temperature exceeds the phase transition point of the material.

[0067] Therefore, the weld formation area can be obtained by predicting the weld penetration, weld width, cross-sectional morphology, and heat-affected zone range.

[0068] In another embodiment of the control method provided in this application, the molten pool state scanning data is obtained by performing anti-interference scanning on the welded product using a 3D scanning device. The anti-interference scanning includes real-time filtering and focus scanning.

[0069] Understandably, during the welding process, intense arc light, smoke, and spatter can severely interfere with the image acquisition quality of the vision system, potentially leading to positioning failures or incorrect corrections. Therefore, the quality of the molten pool state scanning data can be improved by performing anti-interference scanning on the welded products.

[0070] Understandably, real-time filtering and focused scanning can work collaboratively at both the signal processing and physical optics levels to prevent various interferences in the welding environment. Real-time filtering simultaneously cleans the signal at both the hardware and software ends of data acquisition, primarily targeting high-frequency random interference caused by electrical noise, ambient light fluctuations, and plasma flash. At the hardware level, this is achieved by designing and embedding analog filters, such as low-pass filters, at the signal output of the scanning sensor. At the software level, digital filtering algorithms smooth the discrete data sequence numerically after the data is acquired and processed. These algorithms include moving average filtering, which quickly suppresses random noise by calculating the arithmetic mean of data points within a sliding window; and Kalman filtering, which not only filters out noise but also provides optimal signal estimation based on the system's dynamic model. For extremely high-intensity instantaneous plasma flash interference during welding, a threshold-based outlier removal algorithm can be used. When the detected signal peak value far exceeds the normal welding radiation range, the data point is considered invalid and discarded, or replaced by interpolation based on valid data before and after it.

[0071] Focused scanning is an active, physical-level anti-interference method that primarily overcomes measurement errors caused by variations in workpiece surface reflectivity, dust and powder obstruction, and shallow optical penetration. It can be implemented using 3D scanners based on laser triangulation or confocal dispersion principles. In laser triangulation, a laser beam is projected onto the workpiece surface. A camera positioned at an angle relative to the emission axis captures the deformed image of the laser beam. Even in the presence of dust, as long as the main energy of the laser beam reaches the solid surface and is reflected, the camera can preferentially capture the clearest bright line. The diffuse dust, due to its scattering characteristics, appears as a blurred background in the image. Image processing algorithms can effectively separate the clear laser outline from the blurred background, thus suppressing dust interference.

[0072] In another embodiment of the control method provided in this application, the preset welding trajectory is obtained by scanning the area to be welded of the product in advance using a 3D scanning device to obtain the geometric dimension parameters of the area to be welded, then converting the geometric dimension parameters of the area to be welded into welding point coordinates, and then performing coordinate transformation on the welding point coordinates.

[0073] Understandably, the preset welding trajectory is generated by using a 3D scanning device to comprehensively scan the area to be welded of the assembled product before the welding process begins, obtaining detailed three-dimensional geometric parameters of that area. These three-dimensional geometric parameters are not simply a few length or height values, but a dense point cloud dataset containing the three-dimensional spatial coordinates of multiple points on the surface of the area to be welded. This dataset depicts microscopic geometric features such as the step height on the weld path, the width and uniformity of the butt joint gap, the relative flatness of the cover plate and the shell, and any possible assembly misalignment. After obtaining the raw point cloud data, point cloud processing algorithms can identify, filter, and convert these geometric parameters characterizing the macroscopic and microscopic morphology of the area to be welded into a series of ordered welding point coordinates that define the key nodes of the welding path.

[0074] After obtaining a series of welding point coordinates defining the path, coordinate transformation is required to finally generate a preset welding trajectory that the controller can recognize and execute. Coordinate transformation involves converting the welding point coordinates previously defined in the coordinate system of the 3D scanning device to the working coordinate system of the laser welding robot or galvanometer system. This transformation is achieved through a pre-calibrated transformation matrix, which contains rotation and translation parameters. This matrix aligns the visual measurement space of the scanner with the motion execution space of the laser processing head, thereby obtaining the preset welding trajectory.

[0075] Furthermore, based on the preset welding trajectory, an expert knowledge base or machine learning model can be used to intelligently match a highly coordinated set of optimized welding process parameters for the actual welding operation of the product. Because the system can identify the geometric features of the current trajectory—for example, when the trajectory contains sharp turns or small-radius arcs—it automatically matches a parameter set that reduces the scanning speed and may introduce beam oscillation to ensure welding quality at corners; when the measured gap value corresponding to a certain segment of the trajectory is detected to be too large, it matches a higher laser power or a slower welding speed to increase local heat input; and when the trajectory is displayed as a long straight line, it matches standard high-speed welding parameters to improve efficiency.

[0076] Based on the control methods of the above embodiments, the following presents various embodiments of the operation control device, electronic device, and computer-readable storage medium of this application.

[0077] like Figure 5 As shown,Figure 5 This is a schematic diagram of an operation control device for executing a control method according to an embodiment of this application. The operation control device 500 implemented in this application includes: a processor 520, a memory 510, and a computer program stored in the memory 510 and executable on the processor 520, wherein... Figure 5 The example uses a processor 520 and a memory 510.

[0078] Processor 520 and memory 510 can be connected via a bus or other means. Figure 5 Taking the example of a connection between China and Israel via a bus.

[0079] Memory 510, as a non-transitory computer-readable storage medium, can be used to store non-transitory software programs and non-transitory computer-executable programs. Furthermore, memory 510 may include high-speed random access memory, and may also include non-transitory memory, such as at least one disk storage device, flash memory device, or other non-transitory solid-state storage device. In some embodiments, memory 510 may optionally include remotely located memories 510 relative to processor 520, which can be connected to the operation control device 500 via a network. Examples of such networks include, but are not limited to, the Internet, intranets, local area networks, mobile communication networks, and combinations thereof.

[0080] Those skilled in the art will understand that Figure 5 The device structure shown does not constitute a limitation on the operation control device 500, and may include more or fewer components than shown, or combine certain components, or have different component arrangements.

[0081] exist Figure 5 In the illustrated operation control device 500, the processor 520 can be used to call the control program stored in the memory 510, thereby implementing the control method described above. Specifically, the non-transitory software program and instructions required to implement the control method of the above embodiment are stored in the memory 510, and when executed by the processor 520, the control method of the above embodiment is executed.

[0082] It is worth noting that, since the operation control device 500 of this application embodiment can execute the control method of any of the above embodiments, the specific implementation method and technical effect of the operation control device 500 of this application embodiment can refer to the specific implementation method and technical effect of the control method of any of the above embodiments.

[0083] Furthermore, one embodiment of this application also provides an electronic device that includes the operation control device described in the above embodiment.

[0084] It is worth noting that, since the electronic device of this application embodiment includes the operation control device of the above embodiments, and the operation control device of the above embodiments can execute the control method of any of the above embodiments, the specific implementation method and technical effect of the electronic device of this application embodiment can refer to the specific implementation method and technical effect of the control method of any of the above embodiments.

[0085] Furthermore, one embodiment of this application also provides a computer-readable storage medium storing computer-executable instructions for performing the control method described above. Exemplarily, the above-described control method is performed... Figures 1 to 5 The methods and steps in the text.

[0086] It is worth noting that, since the computer-readable storage medium of this application embodiment can execute the control method of any of the above embodiments, the specific implementation and technical effects of the computer-readable storage medium of this application embodiment can be referred to the specific implementation and technical effects of the control method of any of the above embodiments.

[0087] It will be understood by those skilled in the art that all or some of the steps and systems in the methods disclosed above can be implemented as software, firmware, hardware, and suitable combinations thereof. Some or all of the physical components can be implemented as software executed by a processor, such as a central processing unit, digital signal processor, or microprocessor, or as hardware, or as an integrated circuit, such as an application-specific integrated circuit. Such software can be distributed on a computer-readable medium, which may include computer storage media or non-transitory media and communication media or transient media. As is known to those skilled in the art, the term computer storage media includes volatile and non-volatile, removable and non-removable media implemented in any method or technology for storing information such as computer-readable instructions, data structures, program modules, or other data. Computer storage media includes, but is not limited to, RAM, ROM, EEPROM, flash memory or other memory technologies, CD-ROM, digital versatile disc DVD or other optical disc storage, magnetic cartridges, magnetic tape, disk storage or other magnetic storage devices, or any other medium that can be used to store desired information and is accessible to a computer. Furthermore, as is known to those skilled in the art, communication media typically contain computer-readable instructions, data structures, program modules, or other data in modulated data signals such as carrier waves or other transmission mechanisms, and may include any information delivery medium.

[0088] In the several embodiments provided in this application, it should be understood that the disclosed systems, instruments, and methods can be implemented in other ways. For example, the instrument embodiments described above are merely illustrative; for instance, the division of units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the shown or discussed mutual couplings, direct couplings, or communication connections may be through some interfaces; indirect couplings or communication connections between instruments or units may be electrical, mechanical, or other forms. Units described as separate components may or may not be physically separate, and components shown as units may or may not be physical units, i.e., they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0089] It should also be understood that the various implementation methods provided in this application can be combined arbitrarily to achieve different technical effects.

[0090] The embodiments of this application have been described in detail above with reference to the accompanying drawings. However, this application is not limited to the above embodiments. Within the scope of knowledge possessed by those skilled in the art, various changes can be made without departing from the spirit of this application.

Claims

1. A closed-loop control method for laser welding based on vision guidance and molten pool monitoring, characterized in that, include: Acquire molten pool condition scanning data and welding process radiation data; The welding equipment positions the product to be welded by means of a preset welding trajectory obtained from the pre-welding product positioning scanning data, and obtains the current welding state by means of the molten pool state scanning data and the welding process radiation data, and adjusts the welding parameters of the welding equipment online by means of the current welding state.

2. The control method according to claim 1, characterized in that, The method further includes: The welding orientation of the welding equipment is controlled by the molten pool state scan data and the preset welding trajectory.

3. The control method according to claim 2, characterized in that, The step of controlling the welding orientation of the welding equipment using the molten pool state scan data and the preset welding trajectory includes: The predicted welding trajectory is obtained by scanning the molten pool state data, and the welding orientation of the welding equipment is controlled based on the predicted welding trajectory and the preset welding trajectory.

4. The control method according to claim 1, characterized in that, The current welding state includes the shape and position of the current weld pool.

5. The control method according to claim 4, characterized in that, The current welding status also includes welding heat input parameters and weld formation area; The process of obtaining the current welding state through the molten pool state scanning data and the welding process radiation data includes: The shape and position of the current weld pool are determined by the molten pool state scanning data and the welding process radiation data. Based on the shape and position of the current weld pool, the welding heat input and weld formation area are determined.

6. The control method according to claim 1, characterized in that, The molten pool state scanning data is obtained by performing anti-interference scanning on the welded product using a 3D scanning device. The anti-interference scanning includes real-time filtering and focus scanning.

7. The control method according to claim 1, characterized in that, The preset welding trajectory is obtained by scanning the area to be welded of the product in advance using a 3D scanning device to obtain the geometric dimension parameters of the area to be welded, then converting the geometric dimension parameters of the area to be welded into welding point coordinates, and then performing coordinate transformation on the welding point coordinates.

8. An operation control device, characterized in that, It includes a memory, a processor, and a computer program stored in the memory and executable on the processor, the processor executing the program to implement the control method as described in any one of claims 1 to 7.

9. An electronic device, characterized in that, Includes the operation control device as described in claim 8.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores computer-executable instructions for causing a computer to perform the control method as described in any one of claims 1 to 7.

Images