Dissimilar metal laser welding device based on swing light beam and molten pool state online monitoring

Abstract

The invention discloses a dissimilar metal laser welding device based on swing light beams and molten pool state monitoring. The dissimilar metal laser welding device aims at improving the welding quality and the joint stability. The device comprises a laser welding head with a light beam swinging function, and the laser welding head can implement nonlinear energy scanning in a welding area according to a preset track type, frequency and amplitude; the acquisition module can synchronously acquire a visual image, an excitation spectrum and an infrared thermal imaging signal of the molten pool at a high frame rate, and extracts interface diffusion and metal mixing characteristics through multi-modal fusion; the calculation module performs dynamic feature modeling according to the collected physical feature information and the swing parameters, and quantifies multi-dimensional state indexes of the welding quality; the control module carries out combined adjustment on the laser power, the welding speed and the swing parameters according to the state indexes, closed-loop feedback control is constructed, and therefore real-time stable regulation and control and defect suppression in the dissimilar metal welding process are achieved.

Description

Technical Field

[0001] The present invention relates to the technical field of laser welding, and in particular to a dissimilar metal laser welding device based on swinging light beam and molten pool state monitoring. Background Art

[0002] Laser welding of dissimilar metals has been widely used in the automotive, aviation, and electronics manufacturing industries due to its high energy density, excellent directionality, and automation adaptability. Conventional laser welding equipment typically applies heat using a fixed laser beam or a simple linear scanning method. The welding process is controlled by static process parameter settings. Some systems integrate visual sensors or thermal imaging devices to identify weld seam paths and monitor the welding status.

[0003] Existing technologies still have numerous shortcomings in dissimilar metal welding. These include molten pool instability caused by significant differences in thermal properties, the easy formation of brittle phases at the interface, insufficient intermetallic mixing, and a lack of in-depth modeling and real-time feedback control capabilities for multi-source state data during the welding process, making it difficult to achieve highly reliable and consistent welding quality control. Furthermore, traditional state monitoring methods, which are mostly single-channel sensing, lack systematic modeling and response mechanisms for metal fusion behavior and interface diffusion characteristics, resulting in lagging control methods and limited adjustment accuracy.

[0004] In view of this, it is urgent to propose an integrated technical path of dynamic perception and real-time control for the dissimilar metal welding process to improve the stability of welding quality and the reliability of interface structure. Summary of the Invention

[0005] The present application provides a dissimilar metal laser welding device based on an oscillating beam and molten pool state monitoring to improve the stability of welding quality and the reliability of interface structure.

[0006] The present application provides a dissimilar metal laser welding device based on an oscillating beam and molten pool state monitoring, comprising:

[0007] A laser welding head for outputting a laser beam and applying welding energy to the dissimilar metal joint, the laser welding head having a beam swinging function for scanning the laser beam energy along a nonlinear trajectory within the welding area based on preset swing trajectory parameters, the swing trajectory parameters including trajectory type, swing frequency, and swing amplitude;

[0008] The acquisition module is used to synchronously collect multi-source data of the welding area at a high frame rate during the laser welding process, including visual images of the weld pool, plasma excitation spectra, and infrared thermal imaging signals. It also extracts physical characteristic information characterizing the degree of metal mixing, interface diffusion behavior, and dynamic fluctuations of the weld pool based on multimodal fusion methods;

[0009] a calculation module for performing dynamic feature modeling based on the physical feature information and the swing trajectory parameters, and calculating multidimensional state indicators characterizing welding quality using a time series feature extraction and spatial energy distribution analysis method, including a molten pool stability index, an instantaneous metal fusion uniformity index, and an interface temperature difference fluctuation index;

[0010] The control module is used to receive multi-dimensional status indicators and perform joint optimization adjustment of laser parameters and trajectory parameters based on the control weight matrix, including real-time adjustment of laser power, welding speed, oscillation frequency and trajectory shape.

[0011] This application has the following beneficial technical effects:

[0012] (1) The laser action range is expanded by nonlinear oscillating scanning of the laser beam, which promotes the enhancement of convection and diffusion in the molten pool, effectively improves the mixing degree between metals, and inhibits the formation of brittle compounds at the interface. (2) By fusion of multi-source data such as visual images, excitation spectra and thermal imaging, the system can comprehensively extract key physical characteristics such as molten pool stability, temperature distribution and interface diffusion, providing highly reliable state support for welding process control. (3) By calculating multi-dimensional welding quality indicators and implementing dynamic adjustment, the system can achieve joint optimization of laser power, welding speed and oscillation trajectory, significantly improving the stability and consistency of the welding process. (4) The device has good adaptive capabilities through dynamic modeling and parameter adjustment mechanisms, and can adapt to different material combinations and irregular joint paths, thereby improving the flexibility and reliability of the welding process. BRIEF DESCRIPTION OF THE DRAWINGS

[0013] Figure 1 This is a schematic diagram of a dissimilar metal laser welding device based on swinging beam and molten pool state monitoring provided in the first embodiment of the present application. DETAILED DESCRIPTION

[0014] The following description sets forth many specific details to facilitate a thorough understanding of the present application. However, the present application can be implemented in many other ways than those described herein, and those skilled in the art can make similar generalizations without violating the scope of the present application. Therefore, the present application is not limited to the specific implementations disclosed below.

[0015] The first embodiment of the present application provides a dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring. Figure 1 , which is a schematic diagram of the first embodiment of this application. Figure 1 The first embodiment of the present application provides a detailed description of a dissimilar metal laser welding device based on a swinging beam and molten pool state monitoring.

[0016] The dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring includes a laser welding head 101 , an acquisition module 102 , a calculation module 103 and a control module 104 .

[0017] The laser welding head 101 is used to output a laser beam and apply welding energy to the joint of dissimilar metals. The laser welding head has a beam swing function, which causes the laser beam to perform energy scanning along a nonlinear trajectory in the welding area based on preset swing trajectory parameters. The swing trajectory parameters include trajectory type, swing frequency and swing amplitude.

[0018] The laser welding head 101 is the core execution unit of the present invention, used to accurately apply laser beam energy to the dissimilar metal joining area. Its structure comprises a laser emitting assembly, a collimating and focusing assembly, an oscillation drive assembly, and a trajectory control interface, forming an integrated programmable execution device. The welding head 101 can drive its internal laser beam oscillation mechanism via an external control signal, so that the output laser beam within the welding area is no longer a linear fixed irradiation, but rather a dynamic energy scanning along a preset two-dimensional nonlinear trajectory. This expands the effective area of ​​the laser heat source and improves the heat input distribution at the interface between the dissimilar metals.

[0019] Specifically, the laser welding head 101 is equipped with a beam swing actuator consisting of a high-speed rotating mirror or a MEMS micro-vibration mirror. This mechanism can accurately swing according to the swing trajectory parameters issued by the control module 104, including three dimensions: trajectory type, swing frequency, and swing amplitude. The trajectory type can be any programmable path such as circular, elliptical, infinite, or spiral. The swing frequency controls the time period for the laser beam to complete a complete trajectory path and is usually adjustable in the range of 100Hz to 2kHz. The swing amplitude determines the maximum lateral displacement of the laser beam from the central axis. The typical value is between 0.1mm and 2mm, and is set according to the thermal conductivity of the metal, the joint width, and the expected weld width.

[0020] The welding head 101 maintains the thermal stability of its optical components through a built-in temperature-controlled cooling circuit to prevent focus drift or uneven output energy under the action of high-power lasers. Furthermore, to ensure beam focusing quality, the welding head 101 is equipped with a set of composite lenses or high-reflectivity laser windows at the front end, with adjustable focal lengths to accommodate dissimilar metal material combinations with varying thicknesses and thermal diffusion properties. During beam oscillation, this focusing system ensures that the laser energy remains concentrated on the workpiece surface or slightly sinks to the center of the metal pool throughout the entire oscillation trajectory, resulting in consistent penetration depth and stable pool morphology.

[0021] The laser welding head 101 also includes a trajectory control interface for receiving trajectory parameter update instructions from the control module 104 in real time and converting them into specific execution actions. This interface connects to the main control system based on an industry-standard communication protocol (such as EtherCAT or RS-485) and has interrupt response capabilities. It can quickly switch beam motion modes based on changes in the molten pool state, thereby preventing weld defects from expanding due to environmental disturbances or material unevenness.

[0022] During use, welding head 101 first receives the initially set oscillation trajectory parameters and focuses the laser output on the dissimilar metal joining area. Through continuous, high-frequency beam oscillation, uniform energy distribution and multi-point excitation are achieved during the welding process, effectively promoting convective mixing and interfacial diffusion between the metals. Furthermore, through the programmable adjustment of trajectory parameters, welding head 101 can adapt to metal materials of varying thicknesses and thermal conductivities, dynamically balancing and compensating for thermal asymmetry during dissimilar metal welding, thereby avoiding the formation of thermal cracks, pores, and brittle phases between the metals.

[0023] In short, the laser welding head 101 not only performs the energy transmission function, but also strengthens the control of metal-to-metal reactions through dynamic swinging, and is the basic physical execution component for achieving high-quality dissimilar metal connections.

[0024] The acquisition module 102 is used to synchronously collect multi-source data of the welding area at a high frame rate during the laser welding process, including visual images of the welding pool, plasma excitation spectra and infrared thermal imaging signals, and extract physical characteristic information characterizing the degree of metal mixing, interface diffusion behavior and dynamic fluctuations of the molten pool based on a multimodal fusion method.

[0025] The acquisition module 102 is a key component for realizing multi-dimensional perception of the molten pool state and real-time data support in the present invention. The core of its design is to be able to perform multi-source information acquisition of the molten pool and its surrounding areas with high temporal resolution and high spatial resolution during the laser welding of dissimilar metals, thereby providing a physical basis for subsequent welding quality modeling and control adjustment.

[0026] In terms of hardware composition, the acquisition module 102 includes at least three sub-devices: a high-speed industrial vision sensor, a spectral acquisition device, and an infrared thermal imager. These three sub-devices correspond to data acquisition capabilities in the visible light domain, the emission spectrum domain, and the infrared band, respectively. The high-speed industrial vision sensor preferably uses a global shutter CMOS image sensor with a frame rate of no less than 100 frames per second and a resolution of no less than 1280×1024 pixels. It can stably capture visual images of the molten pool contour, spatter behavior, and weld formation during laser welding. Combined with light source assistance, it can achieve clear imaging despite strong background light interference in the weld zone. The spectral acquisition device uses a highly sensitive linear spectrometer with an operating wavelength range of 350nm to 900nm. It is used to collect the plasma spectral signals generated by metal excitation during welding. By analyzing the intensity and wavelength distribution of characteristic spectral lines, key information such as element type, concentration changes, and excitation temperature can be extracted, thereby indirectly reflecting metal mixing and reactivity. The infrared thermal imager uses a non-contact high-precision thermal infrared camera with an operating band of 8-14μm, a thermal sensitivity better than 0.05℃, and a sampling frequency of not less than 50 frames / second. It can capture the temperature field distribution of the molten pool surface and the adjacent heat-affected zone in real time, and identify potential local overheating, cold crack sensitive areas and uneven energy input problems.

[0027] The above three types of sensors achieve unified timestamp alignment and synchronous sampling through industrial-grade data acquisition cards or edge computing units. After acquisition, all data are pre-processed by the fusion processing unit inside the acquisition module, including image denoising, spectral baseline correction, infrared thermal image grayscale temperature calibration and other operations. Based on the multimodal fusion method, space-time alignment and feature coupling are performed, and finally a unified data packet format is formed and output to the computing module 103 for use. In terms of fusion methods, a feature-level fusion strategy is adopted to coordinate the visual edge information with the thermal imaging gradient change, and the spectral peak intensity is superimposed as a weighting factor to improve the boundary recognition accuracy of the metal mixing area and the dynamic response capability of the state change. Through this multimodal data structure, the following physical feature information can be extracted synchronously: molten pool geometry, center temperature, interface temperature difference distribution, spectral peak shift trend, element excitation intensity, welding spatter frequency and edge disturbance amplitude, etc.

[0028] All sensors in the acquisition module 102 are mounted on a structural support coaxially or on the same platform as the laser welding head 101. Their optical axis adjustment and focal length settings are calibrated to ensure that the acquisition area accurately covers the laser scanning area, avoiding blind spots and measurement errors. Furthermore, the acquisition module features built-in temperature control and optical filtering protection systems to ensure long-term stable operation in strong laser and high-temperature environments.

[0029] For example, acquisition module 102 synchronously monitors the laser welding process of aluminum and steel dissimilar metals, collecting three types of data: visual images of the weld area, plasma spectral signals, and infrared thermal images. To perform multimodal fusion, these three types of data are first time-synchronized to ensure that each image frame, each set of spectral signals, and each frame of thermal image corresponds to the welding state at the same moment.

[0030] At one welding point in time, the visual image showed that the edge of the molten pool was slightly irregular, indicating possible local instability; the corresponding thermal image showed that the temperature at the edge of the area fluctuated greatly, while the center temperature was relatively low; at the same time, the emission intensity of the aluminum element in the spectral signal increased abnormally, while the emission signal of the iron element weakened, indicating that the metal mixing ratio had changed and aluminum segregation might have occurred.

[0031] In this case, the three types of information are analyzed through "feature superposition and fusion." Specifically, the melt pool boundaries identified in the visual image are overlapped and compared with the temperature distribution in the thermal map to determine whether corresponding temperature unevenness exists. This is then combined with changes in the proportion of metal elements in the spectral signal to confirm whether uneven metal diffusion or insufficient fusion has occurred in that area.

[0032] This fusion approach extracts three distinct physical features: first, the degree of metal mixing, as reflected by the change in the ratio of the spectral intensities of aluminum and iron; second, interface diffusion behavior, as determined by the uniformity of the temperature distribution combined with changes in spectral peaks; and third, dynamic fluctuations in the weld pool, as assessed by a comprehensive assessment of changes in boundary shape and temperature jumps in the thermal image. These features are transmitted as structured information to the computation module 103 for subsequent welding quality analysis and parameter adjustment.

[0033] In summary, the acquisition module 102 not only realizes high-frequency, low-latency, and physically coupled data acquisition of the state of the molten pool of dissimilar metal laser welding from multiple dimensions, but also establishes a standardized physical feature set that can be used for subsequent calculation and analysis through fusion processing, so that the system has the basic capabilities to support dynamic modeling and real-time feedback control.

[0034] Furthermore, the acquisition module is specifically used to:

[0035] In the initial stage of welding, a weld area stability prediction model based on the oscillation trajectory parameters of the laser welding head and the preset metal combination characteristics is established to predict the potential welding quality fluctuation risk level at each trajectory position point throughout the oscillation cycle;

[0036] Based on the output of the stability prediction model, the swing trajectory is divided into multiple dynamic acquisition areas, and enhanced acquisition parameters are automatically configured for areas with high predicted risk levels, including improved image resolution, increased thermal image refresh rate, or extended spectral integration time;

[0037] During the actual welding process, the acquisition module performs local enhanced sampling on high-risk areas according to the dynamic acquisition area configuration, and the spatial resolution or time density of the acquired visual images, plasma excitation spectra, and infrared thermal imaging signals is higher than that of the standard area;

[0038] The data obtained by enhanced sampling is marked as high-confidence feature fragments, and its corresponding trajectory position label and risk prediction level are attached, which are used as key frame priority input in subsequent modeling;

[0039] The physical feature information extracted from the standard area and the high-risk area is uniformly structured and output as physical feature information for characterizing the degree of metal mixing, interface diffusion behavior and dynamic fluctuations of the molten pool, and output to the calculation module for dynamic feature modeling based on the physical feature information.

[0040] In this embodiment, the functional design of the acquisition module is not limited to conventional data acquisition, but integrates multi-level technologies such as predictive modeling, regional dynamic configuration, enhanced sampling strategy and data structured output to achieve real-time identification and refined monitoring of potential fluctuation risks in welding quality during the laser welding of dissimilar metals, ensuring that the basic data for subsequent modeling and control have high confidence and high responsiveness.

[0041] At the initial stage of welding, the system first calls the built-in metal material database according to the task parameters to extract the combined characteristics of dissimilar metals involved in the current welding. This characteristic set includes but is not limited to the thermal conductivity, specific heat capacity, melting point, surface tension, laser reflectivity and intermetallic reactivity of the two metals. These characteristics determine the impact of heat input on the interface area and the stability of the mixing behavior during the welding process. At the same time, the current oscillation trajectory parameters of the laser welding head are also read synchronously, including the oscillation trajectory type (such as "∞" shape, sawtooth shape, Lissajous curve, etc.), oscillation frequency (in Hertz) and oscillation amplitude (in millimeters). These parameters describe how the laser beam moves on the weld surface, thereby affecting the local energy distribution.

[0042] Based on the above inputs, the system constructs a model for predicting the stability of the welding area. This model can use a structure such as a multi-layer perceptron (MLP) or a lightweight random forest. During the training phase, the model inputs quality fluctuation data under different swing trajectories and metal combinations in historical welding experiments, and outputs the welding quality fluctuation risk level at each trajectory position point in the welding area. The risk level is a discrete label, usually divided into low risk (level 0), medium-low risk (level 1), medium-high risk (level 2), and high risk (level 3). In actual welding tasks, the model is used to predict and output the risk level of each trajectory position point throughout the entire swing cycle, forming a risk distribution map.

[0043] Based on this risk distribution, the system then divides the swing trajectory into multiple dynamic acquisition zones, each corresponding to one or more trajectory locations. For trajectory zones predicted to be high-risk, the system automatically configures enhanced acquisition parameters to enhance data acquisition capabilities in that area. The acquisition parameter configuration includes three optimization strategies: first, increasing the resolution of the visual image, for example, from 640×480 to 1280×960, to more precisely capture changes in the morphology of the melt pool boundary; second, increasing the infrared thermal imaging frame rate, for example, from 30fps to 90fps, to more frequently capture temperature changes and identify short-term heat flow disturbances; and third, extending the integration time of the spectral sensor to more fully receive the weak excitation light signal, thereby improving the signal-to-noise ratio of alloy evaporation characteristics or intermetallic reaction spectral lines.

[0044] During the actual welding process, the acquisition module uses the aforementioned dynamic acquisition area divisions to perform enhanced sampling in high-risk areas. Enhanced sampling is not limited to increasing the sampling frequency or resolution but also includes changes in sampling methods, such as enabling multi-angle simultaneous camera capture within high-risk areas or increasing the spatial sampling density of thermal imaging points. This allows the system to capture changes in physical behavior within high-risk trajectory segments with greater temporal and spatial resolution, providing sufficient data support for subsequent feature extraction.

[0045] All data obtained from enhanced sampling will be marked during the acquisition phase. First, they will be labeled with a "high-confidence feature segment" label to indicate that the data is highly reliable. At the same time, a "trajectory position label" will be attached, which is indexed by the normalized position identifier within the swing period (such as 0.15T, 0.60T, etc.). In addition, each piece of data will also be labeled with a "risk prediction level" to indicate the priority setting of the data in the modeling process.

[0046] This data is structured and organized within the acquisition module. The extracted physical characteristics primarily include the frequency and amplitude characteristics of the melt pool boundary vibration in the visual image, the relative intensity variations of key elements (such as Fe, Ni, and Cu) in the spectral signal, and the temperature gradient and local temperature difference variation patterns in the infrared thermal image. All of this data from high-risk and standard areas is uniformly formatted into three types of physical characteristics, corresponding to the degree of metal mixing, interface diffusion behavior, and dynamic fluctuations in the melt pool.

[0047] For example, the degree of metal mixing can be reflected by the ratio of the intensity of different metal spectral lines in the spectrum and the consistency of their distribution along the welding path; the interface diffusion behavior can be measured by the width and gradient of the interface temperature transition region in the high-resolution infrared image; and the dynamic fluctuation of the molten pool can be represented by the frequency response changes of the molten pool boundary contour in the image sequence.

[0048] Ultimately, the acquisition module fully outputs all of this structured physical feature information to the computation module, serving as the input data source for subsequent dynamic modeling and multi-dimensional state indicator calculations. The entire acquisition module process is centered on the physical logic of the welding process. Through risk-aware data enhancement and labeling strategies, the system significantly improves monitoring accuracy and responsiveness in complex welding scenarios.

[0049] Furthermore, the acquisition module is further configured to:

[0050] When the risk prediction level at a certain swing trajectory position reaches a set high-risk threshold, the visual image enhancement acquisition process is automatically initiated. The visual image enhancement acquisition process includes: performing multi-scale wavelet reconstruction processing on the visual image collected at the high-risk trajectory position, and combining edge-guided filtering with a local contrast enhancement algorithm to enhance the texture saliency of the weld pool edge, interface disturbance, and metal-to-metal detail structure contained in the visual image;

[0051] Based on the grayscale histogram distribution information of the infrared thermal imaging signal collected at the high-risk trajectory location, the hot spot concentration index of the thermal imaging channel is dynamically calculated. When the hot spot concentration exceeds the adaptive saturation threshold, the slope of the response curve and the saturation upper limit parameters of the infrared thermal imaging signal are adjusted in real time, so that the infrared thermal imaging signal maintains the details of the high-temperature area while improving the overall contrast of the low-temperature area.

[0052] Perform pixel-level alignment and inter-channel geometric registration on the enhanced visual image and the infrared thermal imaging signal after response curve adjustment to form composite enhanced frame data with spatial consistency and modal complementarity. The composite enhanced frame data synchronously inherits its corresponding trajectory position label and risk prediction level information.

[0053] Structurally marking the composite enhanced frame data as high-confidence feature segments, and adding a texture complexity index and a hot spot adjustment parameter as auxiliary modeling factors during the structuring process to provide a basis for regional refinement weights when subsequently output to the calculation module;

[0054] The high-confidence feature fragments are uniformly output as part of the physical feature information used to characterize the degree of metal mixing, interface diffusion behavior and dynamic fluctuations of the molten pool, and are fused with the standard area acquisition results and input into the calculation module to enhance the accuracy and response sensitivity of the modeling results at the high-risk trajectory position points.

[0055] In the proposed dissimilar metal laser welding device based on an oscillating beam and melt pool state monitoring, upon detecting that the risk prediction level at a specific oscillating trajectory location reaches a set high-risk threshold, the acquisition module performs a series of image enhancement operations to obtain higher-quality physical feature information, providing the computation module with input data with enhanced discriminative capabilities, thereby improving the perception and response accuracy of weld quality fluctuations. To achieve this goal, the acquisition module's design incorporates multi-scale signal processing, image enhancement, adaptive adjustment of infrared response curves, and multimodal data registration, forming a complete, feasible, and technically sound visual image enhancement acquisition process.

[0056] Specifically, when the risk prediction model determines that there is a significant risk of unstable welding quality at a certain swing trajectory position point based on the previous modeling results or real-time data analysis, that is, the prediction level reaches the set high-risk threshold (for example, level 3, the full score), the system triggers the visual image enhancement acquisition process. The first step of this process is to perform multi-scale wavelet reconstruction processing on the visual image collected at the trajectory position point. Wavelet transform is a tool that can simultaneously analyze the local frequency and time domain information of the signal at different scales. It is suitable for multi-resolution decomposition of complex details such as boundary changes and molten pool disturbances in welding images. By performing two-dimensional discrete wavelet decomposition on the original image, the system can extract low-frequency background information and high-frequency edge information, and amplify the influence weight of the high-frequency component in the reconstruction stage, so that the boundary contour and texture changes are clearer. In this device, it is preferred to use db4 wavelet as the mother wavelet function, and set the number of decomposition layers to 3 layers to take into account both edge detail retention and computational efficiency.

[0057] After completing the wavelet reconstruction, in order to further highlight the detailed structure between metals and the disturbance of the molten pool boundary in the image, the system sends the image to the edge-guided filter for denoising and detail preservation. Edge-guided filtering is different from traditional mean filtering or Gaussian filtering. It does not blur the image edges while smoothing the noise. Instead, it combines the image gradient direction information and only performs limited smoothing in the vertical direction of the edge, thereby retaining the edge sharpness. This process significantly improves the local clarity of areas such as the edge of the laser molten pool and the metal overlap line. In addition, in order to improve the overall contrast, especially when the grayscale distribution of the image is concentrated in the medium and low range, the system introduces a local contrast enhancement algorithm (such as CLAHE, contrast-limited adaptive histogram equalization), which stretches the grayscale dynamic range according to the local window of the image, so that the dark details can be restored and the accuracy of the subsequent texture complexity calculation can be ensured.

[0058] At the same time, the acquisition module also synchronously processes the infrared thermal imaging signal at the high-risk trajectory position point. Unlike visible light images, infrared images use temperature distribution as the signal source. The key is to reflect the spatial uniformity and local concentration of heat input in the welding area. In order to identify whether there are problems such as excessive concentration of heat input and uneven energy distribution, the system first extracts the grayscale histogram distribution of the current thermal imaging frame, and calculates the hot spot concentration index on this basis. This index can be defined as the product of the proportion of pixels above the 80% grayscale threshold and its aggregation density function, which is used to measure whether the distribution of high-temperature areas in the image is concentrated in a small area. If the hot spot concentration exceeds the set adaptive saturation threshold (for example, set to 0.35), it indicates that the welding heat is excessively concentrated, which may cause abnormal local penetration or metal burn-through.

[0059] In the case of excessively high hot spot concentration, the system performs real-time adjustment processing on the response curve of the infrared image in order to preserve the detailed information of the high-temperature area and enhance the visibility of the temperature distribution in the low-temperature area. The adjustment process includes two parameters: the slope of the response curve and the saturation upper limit. The slope of the response curve controls the response sensitivity of the low and medium grayscale ranges. An increase in the slope means that the temperature difference response in the low-temperature area is amplified, while the saturation upper limit parameter controls the upper limit of the image brightness to prevent grayscale compression in the high-temperature area, resulting in information loss. This dynamic adjustment mechanism is based on a strategy that combines gamma correction with linear interval mapping, and combines the hot spot position distribution to adaptively set the local adjustment interval, so that the entire infrared image achieves the best balance between visual perception and quantitative analysis accuracy.

[0060] After completing the image enhancement and infrared response adjustment, the system enters the multimodal data registration stage. Since the visible light visual image and the infrared image come from different types of sensors, and the viewing angle, pixel density, and focal length parameters are not completely consistent, direct channel superposition will lead to problems such as image mismatch and structural dislocation. To this end, the system uses the SURF (Speeded Robust Features) key point extraction method to match the feature points of the enhanced image and the infrared image, and removes the matching external points based on the RANSAC algorithm. On this basis, the geometric registration is completed through affine transformation. The images of the two modalities are then aligned at the pixel level to unify their coordinate systems and image sizes to generate composite enhanced frame data. This data has spatial consistency and modal complementarity. It not only contains the texture, edge, and structural information in the visible light image, but also integrates thermal features such as temperature gradient and hot spot concentration in the infrared image.

[0061] While generating composite enhanced frame data, the system inherits the original acquired trajectory position labels and risk prediction level information of the frame data, and uniformly marks them as high-confidence feature fragments. The so-called "high confidence" means that: on the one hand, the data source comes from trajectory position points with high risk levels and high possibility of welding instability; on the other hand, after image enhancement and multimodal fusion processing, the information expression is significantly better than that of ordinary frame data. During the marking process, the system also adds the calculated texture complexity index and hot spot adjustment parameters as auxiliary factors in the subsequent modeling process. The texture complexity index can be calculated through the image grayscale co-occurrence matrix, which reflects the degree of change in the local entropy value of the image; the hot spot adjustment parameters include the slope of the response curve, the gamma value and the saturation interval position, all of which are used to characterize the enhanced processing characteristics of the thermal image in this frame.

[0062] In actual operation, the process of obtaining the texture complexity index depends on the analysis of the spatial structural characteristics of the image texture. Specifically, the statistical characteristics of the image texture can be extracted in multiple directions and at multiple pixel spacing scales by constructing a grayscale co-occurrence matrix of the visual image, including indicators such as energy, contrast, correlation and entropy. In particular, the distribution characteristics of the local entropy value are used as the main reference parameter to characterize the richness of the texture changes of the image at the microscopic scale. In the present invention, for each frame of visual image collected during the welding process, the system constructs a local grayscale co-occurrence matrix in its edge area and metal fusion detail area, and counts the maximum value, average value and change gradient of the local entropy, so as to comprehensively determine the texture complexity level of the image. If the grayscale value of the metal fusion zone in the image fluctuates violently within a small spatial range, it can be considered to have high complexity characteristics, and the system will assign a higher texture complexity index accordingly.

[0063] Hot spot adjustment parameters are acquired based on the parameter recording and adjustment process during the infrared thermal imaging signal enhancement process. At high-risk trajectory locations, the system analyzes the original grayscale histogram of the infrared image to determine whether there are concentrated hot spots, brightness saturation, or insufficient contrast in low-temperature areas. If these issues are present, the system activates the thermal image enhancement mechanism and dynamically adjusts the relevant parameters of the infrared response curve. The slope of the response curve is determined by the linear gain adjustment of the image brightness mapping function, the gamma value is derived from the setting of the nonlinear brightness compression or expansion processing, and the saturation interval position reflects whether the dynamic range of the high-temperature area of ​​the infrared image is compressed or widened before and after enhancement. The system records the response slope coefficient, the degree of nonlinear gamma adjustment, and the set highlight saturation start and end intervals for each infrared image frame in the image enhancement module. These enhancement process parameters collectively constitute the hot spot adjustment parameters for that frame and serve as a reference for subsequent thermal imaging channel quality evaluation and weighting adjustments.

[0064] Furthermore, in the process of obtaining the texture complexity index, the system first extracts the central 400×400 pixel area of ​​the weld area marked as high-risk in the visual image as an analysis window. This area generally covers the entire weld pool and the metal diffusion area at its edge. To measure texture complexity, the system uses the gray-level co-occurrence matrix (GLCM) method to extract the local entropy of the image as a representation of texture saliency.

[0065] Specifically, the system calculates the co-occurrence matrix for the region at pixel distances of 1 and 2 at 0°, 45°, 90°, and 135°, resulting in eight sets of matrices. For each set of co-occurrence matrices, the system calculates its entropy value. Entropy is an indicator of the uncertainty of the grayscale distribution of image pixels and can express the texture complexity of the image in that direction. For example, in this embodiment, the statistical results of the eight sets of entropy values ​​are:

[0066] Average entropy: 3.92

[0067] Maximum entropy value: 5.12

[0068] Minimum entropy value: 2.87

[0069] Standard deviation: 0.65

[0070] The system sets a normalization rule to map the maximum entropy value to 1 and the minimum entropy value to 0. The texture complexity index is calculated using the standard Z-score normalization method:

[0071] Substitute the average entropy value of the current region into the Z-score normalization expression:

[0072] Index = (current average entropy value - minimum entropy value) ÷ (maximum entropy value - minimum entropy value)

[0073] Substituting the values ​​into this equation, we get:

[0074] Index = (3.92-2.87) ÷ (5.12-2.87) ≈ 1.05 ÷ 2.25 ≈ 0.4667

[0075] To further consider the impact of local texture inhomogeneity, the system multiplies the standard deviation by an empirical weighting factor (such as 1.2) as an enhancement term and adds it to the index. The final calculation result is:

[0076] Final texture complexity index = 0.4667 + (0.65 × 1.2 ÷ 10) = 0.4667 + 0.078 ≈ 0.5447

[0077] In the above calculation, "10" is the normalization coefficient for the enhancement term, which is set based on the empirical distribution range of the product term of the standard deviation of local texture entropy values ​​and the enhancement factor in the training samples. Specifically, during the offline training phase, the system performed batch statistics on image data from multiple high-risk trajectory locations in typical welding scenarios. The measured entropy standard deviation generally ranged from 0.3 to 0.9, and the enhancement factor was set to an upper limit of 1.5. Therefore, the theoretical maximum value of the product term does not exceed 1.35. To avoid a shift in the order of magnitude between the enhancement term and the base entropy value, which could affect the numerical stability of the final texture complexity index, the system adopts a strategy of scaling the product term to the range of 0–0.1. After multiple rounds of experimental verification, it was ultimately determined that a fixed value of 10 would be used as the normalization denominator. This ensures that the product result, after being divided by 10, falls within an appropriate weighted superposition range. This enhances the sensitivity to complex texture structures while preventing false hotspot detections due to excessive perturbations.

[0078] Because the system uses a combined entropy normalization index and edge strength scoring model, Canny edge detection was used to extract edge point density for the same area (1783 edge points were detected in this area, with a total of 160,000 pixels). The edge density was 1783 / 160,000, which is ≈ 0.0111. The system normalized the edge density (maximum density 0.02, minimum density 0.002), resulting in an edge complexity score of (0.0111 - 0.002) ÷ (0.02 - 0.002) ≈ 0.505.

[0079] The system performs a weighted fusion of the texture entropy complexity and the edge complexity score, and the fusion ratio is as follows: Final texture complexity index = 0.5447×0.6+0.505×0.4=0.3268+0.202=0.5288

[0080] To further emphasize texture saliency in high-risk areas, the system boosts the above scores by a factor of 1.3 corresponding to risk level 3:

[0081] Final texture complexity index = 0.5288×1.3≈0.6874

[0082] For model consistency and empirical evaluation requirements, the system uses Sigmoid normalization (center 0.5, slope 10) on this value before final output, resulting in a standard index of:

[0083] Exponent = 1 / (1+exp(-10×(0.6874-0.5)))≈0.86

[0084] Therefore, the final recorded texture complexity index is 0.86, which is based on a clear image area, a clear image statistical feature extraction method, a dual fusion scoring and normalization process, and is processed by dynamic enhancement logic corresponding to the risk level.

[0085] For the hot spot adjustment parameters, the system uses a single-frame image captured by an infrared thermal image at a high-risk trajectory point, first extracts its grayscale histogram, and counts the proportion of pixels in the high-temperature area in the image. The high-temperature area is defined as pixels with grayscale values ​​in the range of 220 to 255 (for application temperatures of approximately 1400°C to the upper limit). In this embodiment, the total image pixels are 640×512, with a total of 327,680 pixels, of which 82,000 pixels are in the grayscale range of 220–255, accounting for approximately 25%. The system uses the proportion of high-temperature pixel concentration as the hot spot concentration index, normalized to 0.25.

[0086] The system sets the hot spot concentration adaptive saturation threshold to 0.20. When the actual hot spot concentration exceeds this value, the system triggers the response curve adjustment logic. In this example, because 0.25 > 0.20, the system enters adjustment mode.

[0087] To prevent oversaturation in high-temperature areas, which could lead to loss of temperature structure information, the system increases the linear slope of the response curve from 1.0 to 1.2 to moderately stretch the grayscale contrast. Furthermore, to preserve the details of the low-temperature areas in the image, the system adjusts the Gamma value to 0.8 (less than 1 represents highlight compression). Finally, because the proportion of pixels corresponding to the highest grayscale of 255 is too high (approximately 3%), the system adjusts the maximum grayscale threshold to 230 to ensure that the image dynamic range is not squeezed. In summary, the hot spot adjustment parameters for this frame are: a response curve slope of 1.2, a Gamma value of 0.8, and a saturation range set to [230–255].

[0088] These parameters are automatically calculated by the system based on the statistical characteristics of the image, without relying on subjective settings. They can maintain consistent calculation logic and technical implementation paths under different welding materials, welding paths or risk distribution conditions, thereby providing a stable, reliable and sensitively responsive key reference for the state modeling and control adjustment strategies in subsequent calculation modules.

[0089] Finally, the acquisition module uniformly outputs the high-confidence feature fragments to the calculation module as the core component for constructing physical feature information used to characterize the degree of metal mixing, interface diffusion behavior, and dynamic fluctuations in the molten pool. In order to ensure the comprehensiveness and accuracy of the modeling process, the acquisition module also inputs the conventional data results collected in the standard area (i.e., trajectory position points that have not reached the high-risk threshold) into the calculation module, and the two are fused in the calculation module. The fusion process not only considers the distribution consistency of the spatial position labels, but also assigns modeling weights according to the risk level and confidence coefficient. For example, when constructing a molten pool stability model, the boundary disturbance amplitude and edge continuity features in the high-confidence frame will obtain higher time series regression weights; when constructing an interface temperature difference fluctuation model, the temperature extremes and gradient direction changes provided by the high-confidence frame will be given priority as the dominant indicators of the global fluctuation trend.

[0090] In the present invention, after the calculation module receives the structured high-confidence feature fragment output by the acquisition module, it not only obtains the basic physical feature information contained therein (such as the edge position of the molten pool, metal mixed texture, temperature gradient field, etc.), but also uses the attached texture complexity index and hot spot adjustment parameters as auxiliary modeling factors to embed into the state indicator modeling process, thereby enhancing the ability to distinguish local welding abnormal behavior, improving modeling accuracy and response sensitivity to high-risk areas.

[0091] Specifically, when constructing the melt pool stability index, the calculation module first extracts the boundary dynamic change feature vector based on the boundary information, optical flow perturbation features and multi-frame temporal changes in the high-confidence feature segment. Subsequently, the system introduces the texture complexity index as a dynamic change weight correction factor to adjust the weight of the boundary perturbation contribution value of each frame in the temporal superposition process. For example, if the texture complexity index of a high-confidence image is 0.85 (value range 0–1), it means that the image contains rich metal fusion details and interface perturbation features. The system will assign a higher temporal superposition weight to the boundary change contribution of the frame, reflecting the high state discrimination value corresponding to its high structural information density. Conversely, if the texture complexity index of a frame image is only 0.3, its role in state modeling will be suppressed to avoid the low-quality image from offsetting the final state result.

[0092] For the calculation of the interface temperature difference fluctuation index, the temperature gradient map in the high-confidence thermal imaging frame is mainly used, and the hot spot adjustment parameter is introduced as a weighted reference for data reliability and regional sensitivity. The slope of the response curve in the hot spot adjustment parameter (such as 1.2, 1.5, 2.0) is used to determine whether the frame image has been enhanced. If the slope is high and the saturation interval is set reasonably, the system will increase the participation weight of the frame in the temperature difference extreme statistics. Furthermore, if the acquisition module records that the thermal imaging frame has undergone a dual adjustment mechanism of local Gamma compression + regional brightness stretching, the system can set the hot spot adjustment level to "level 2 processing" and input the parameter as an entropy sensitive factor into the global thermal difference fluctuation model, so that the model can perform appropriate response amplification processing on frame data with strong adjustment ability and prominent thermal distribution during the aggregation process.

[0093] Furthermore, during the modeling process, the Texture Complexity Index and Hot Spot Adjustment Parameters are used to assist in the identification of abnormal trajectory segments. For example, during the spatial distribution analysis of state indicators, if a trajectory segment contains multiple consecutive frames with a Texture Complexity Index greater than 0.9, and the Hot Spot Adjustment Parameters indicate that the image has reached level 3 saturation compensation, the system will comprehensively identify the region as potentially unstable due to the coincidence of structural texture anomalies and concentrated heat input. This segment will then be marked as a potentially high-risk segment and prioritized by subsequent control modules.

[0094] The calculation module 103 is used to perform dynamic feature modeling based on the physical feature information and the swing trajectory parameters, and use time series feature extraction and spatial energy distribution analysis methods to calculate multidimensional state indicators characterizing welding quality, including the molten pool stability index, the instantaneous metal fusion uniformity index, and the interface temperature difference fluctuation index.

[0095] The calculation module 103 is the core of the analysis and modeling process in the present invention. Its function is to jointly process the multi-source physical feature information provided by the acquisition module 102 and the beam swing parameters of the welding head 101 to construct a dynamic feature model. It then extracts key indicators that quantify the welding process state, which are used for real-time adjustments by the subsequent control module 104. This module's operation includes steps such as data reception, feature standardization, time series analysis, spatial energy distribution modeling, and multidimensional indicator calculation. This ensures that the extracted indicators have clear physical meaning, sufficient sensitivity, and good stability, and can accurately reflect the physical evolution of the metals during the welding process.

[0096] After receiving the visual image, spectral signature, and thermal imaging data from acquisition module 102, the computation module first synchronizes the timelines of the three data types, removing dropped frames and outliers. It then standardizes each type of feature data, such as by unifying units, eliminating background interference, and improving spatial resolution. Subsequently, based on the temporal progression of the welding process, the data is segmented and organized into feature sequences with equal time intervals to capture continuous dynamic trends.

[0097] During the time series analysis phase, the calculation module uses a sliding window technique to extract local fluctuations, average trends, and peak characteristics. This is combined with beam swing trajectory parameters (such as trajectory type, frequency, and amplitude) for paired modeling, enabling causal analysis of process states. For example, if visual images show dramatic fluctuations at the melt pool boundary over a certain period of time, and thermal images also exhibit sharp temperature jumps, accompanied by sudden changes in the intensity of an element in the spectrum, the system will flag this area as an abnormal state segment and initiate the indicator calculation process.

[0098] In terms of the definition and calculation of status indicators, the molten pool stability index is an indicator that measures the amplitude and frequency of changes in the shape of the molten pool boundary during the welding process. The specific calculation method is to extract the edge contours of the molten pool in consecutive frames in the visual image, and calculate the standard deviation and change rate of its boundary contour. The greater the fluctuation of the boundary contour, the higher the index value, indicating that the molten pool is unstable. For example, when the boundary position change in five consecutive frames exceeds a preset threshold (such as 0.3mm) and there is an oscillation frequency higher than 10Hz, the stability index will increase significantly, indicating the risk of welding spatter or local collapse.

[0099] The instantaneous metal fusion uniformity index is used to evaluate whether the mixing of dissimilar metal elements in the molten pool is uniform during the current time period. The calculation of this index relies on the emission intensity representing different metal elements in the spectral signal. After ensuring that the spectral signal has been baseline calibrated, the system analyzes the spectral intensity ratio of representative metal elements such as aluminum and iron, and calculates its degree of discreteness between different sampling points (or areas). If the difference in the aluminum / iron intensity ratio in multiple sampling areas at the same time point is small, it means that the mixing is more uniform and the index value is lower; on the contrary, if there is a local deviation or mutation, the index increases, indicating that there is an uneven metal fusion problem.

[0100] The interface temperature difference fluctuation index measures uniform heat input and smooth heat diffusion by analyzing the lateral temperature distribution along the weld interface. It is calculated by extracting the temperature values ​​of the weld centerline and several sampling points on both sides from the thermal image and calculating the rate of change and periodic oscillation amplitude of the temperature difference. If a region experiences frequent temperature changes or alternating high-temperature concentrations and low-temperature boundaries, the index value will increase, indicating a risk of heat input fluctuations in that region, which could lead to cracks or discontinuities in the interface structure.

[0101] The entire calculation process is performed once per laser trajectory cycle, ensuring the system can update state assessment results at a high frequency. The calculation module also organizes all indicators into a structured data format, binds them with timestamps and trajectory parameters, and forms a complete process state vector, which is output to the control module 104 for use.

[0102] To facilitate understanding of this module's workflow, the following example illustrates this. For example, in the actual scenario of aluminum-steel dissimilar metal lap welding, to enhance elemental mixing in the weld zone and suppress the formation of brittle intermetallic compounds, the system sets the laser beam output by the laser welding head to oscillate along an elliptical trajectory. The oscillation frequency is set to 100 Hz, the amplitude is 1.5 mm, and the trajectory type is a closed counterclockwise ellipse. The welding speed is set to 300 mm / minute, and the laser power is set to 2.2 kW.

[0103] During the welding process, the acquisition module simultaneously collects three types of data at a rate of 30 frames per second: a visible light image of the weld area, a plasma excitation spectrum, and an infrared thermal image. Each one-second processing cycle is used by the computing module 103 to combine these 30 frames of multi-source data with the trajectory parameters of the beam swing, completing dynamic feature modeling and extracting multidimensional state indicators.

[0104] To complete the dynamic feature modeling, the calculation module first calls the swing trajectory parameters recorded in the control system, including trajectory type, frequency, and amplitude. Based on the trajectory model in the system, it infers the precise position of the laser beam on the weld joint for each frame. For example, the first frame of the image corresponds to the laser beam at the lower left corner of the elliptical trajectory, and the 16th frame corresponds to the upper right corner. Each frame of data precisely corresponds to the position of its laser irradiation point in the weld coordinate system, marked as positions P0 to P29.

[0105] Based on this frame position information, the calculation module processes the collected data. First, for the visual image, a fixed threshold (for example, pixel grayscale values ​​greater than 180) is used to extract the melt pool area in each frame. Then, contour analysis algorithms (such as contour fitting and edge extension) are used to calculate the maximum lateral width of the melt pool in each frame. The resulting 30 width values ​​are as follows (in millimeters):

[0106] [2.1,2.2,2.3,2.0,2.2,2.1,2.5,2.4,2.2,2.3,...,2.0]

[0107] The calculation module performed a standard deviation calculation on the melt pool width sequence to obtain a melt pool stability index, which was 0.18 mm in this case. Furthermore, the calculation module grouped each image frame by trajectory position. Analysis revealed that the images corresponding to frames P7–P10 showed significant fluctuations, indicating that the melt pool was unstable when the laser beam was at the upper portion of the trajectory, which is related to possible energy concentration or reflection at this location.

[0108] Next, we process the spectral data corresponding to each frame. We extract the emission intensities of aluminum (396.15 nm) and iron (404.58 nm) from each frame of the spectral signal and calculate the aluminum / iron ratio, yielding the following sequence:

[0109] [1.10,1.12,1.09,1.08,1.15,1.14,1.07,1.05,1.20,...,1.06]

[0110] The calculation module calculates the standard deviation of this ratio sequence, yielding the instantaneous intermetallic fusion uniformity index, which is 0.17 in this case. Combined with analysis of the spot trajectory position, the aluminum content increases continuously from frames P14–P18, suggesting insufficient fusion and severe aluminum segregation in this region.

[0111] Next, the infrared thermal imaging data is processed. In each frame of the thermal image, the system selects five equally spaced pixels on each side of the weld centerline and calculates the average temperature difference between the left and right sides (left temperature - right temperature), forming a temperature difference time series as follows (in degrees Celsius):

[0112] [1.9,2.1,2.0,2.2,1.8,2.3,2.1,...,1.9]

[0113] The calculation module takes the difference between the maximum and minimum values ​​of the sequence to determine the interface temperature fluctuation index, which is 0.5°C in this case. Trajectory matching analysis reveals that the maximum temperature fluctuation occurs when the beam swings to the upper right corner of the ellipse (frames P20–P24). This suggests that the heat input in this region is slightly higher, suggesting possible heat accumulation.

[0114] Finally, the three status indicators of this processing cycle are output as follows:

[0115] Molten pool stability index: 0.18mm

[0116] Instantaneous metal fusion uniformity index: 0.17

[0117] Interface temperature fluctuation index: 0.5℃

[0118] In addition, each indicator is accompanied by a spatial position mark, indicating the trajectory segment where the potential poor welding area is located, which is used to guide the control module 104 to perform parameter adjustments, such as reducing laser power, reducing swing amplitude, or fine-tuning the swing trajectory to stabilize the welding behavior in the area.

[0119] Furthermore, the calculation module is specifically used to:

[0120] receiving the physical feature information including the risk prediction level and the trajectory position label output by the acquisition module, and dividing the physical feature information into a plurality of spatial partition data segments according to the trajectory position, wherein each data segment corresponds to a sampling area under a swing trajectory;

[0121] According to the risk prediction level, weighting coefficients are assigned to different spatial partition data segments, and the weighting coefficients are used to increase the contribution weight of the data corresponding to the high-risk area in the state indicator modeling;

[0122] When calculating the melt pool stability index, priority is given to the dynamic information fragments from the boundaries with higher risk levels, and the frequency of boundary contour changes is weighted averaged by time series superposition to form a stability state result that reflects the dominant behavior of the high-fluctuation area;

[0123] When calculating the instantaneous metal fusion uniformity index, the key frame data in the spectral feature information is selected, and the index is adjusted based on the standard deviation of the spectral ratio of different sampling points in the same swing trajectory area and the risk weight, so as to highlight the dominant contribution of abnormal fusion behavior;

[0124] When calculating the interface temperature difference fluctuation index, the temperature gradient information of each spatial partition is used to calculate the temperature difference extreme within the region, and all partition results are aggregated into a global temperature difference fluctuation index according to risk weighting to achieve a sensitive response to the imbalance of heat input at the dissimilar metal interface.

[0125] In this implementation, the computation module is primarily responsible for structured processing and dynamic modeling of the physical feature information output by the acquisition module to generate multidimensional state indicators representing welding quality, providing a quantitative basis for subsequent control and adjustment. This process is centered around three core indicators: the weld pool stability index, the instantaneous intermetallic fusion uniformity index, and the interface temperature difference fluctuation index. The entire workflow is centered around risk perception, integrating trajectory position labels to implement spatially differentiated feature modeling and a weighted response mechanism.

[0126] First, the computation module receives the output data from the acquisition module. This data not only includes the fusion results of the visual image, plasma excitation spectrum, and infrared thermal imaging signals, but also includes the trajectory location label and risk prediction level for each data segment. These labels indicate the specific location of the data segment within the welding trajectory and the potential probability of quality fluctuation in that area during the welding process. This information enables the computation module to spatially partition the data according to the actual laser beam trajectory.

[0127] The calculation module divides the data of the entire welding process according to the swing trajectory, discretizing the entire welding path into multiple spatial partitioned data segments. Each data segment corresponds to a data set in a certain time period or trajectory interval in the welding path. This division method allows the system to model and analyze in units of regions, enhancing the model's ability to perceive spatial position-related features. For example, when performing an "∞"-shaped swing trajectory welding task, the system can divide the trajectory into five partitions: the upper left circle, the upper right circle, the intersection center, the lower left circle, and the lower right circle. The data set in each partition includes the above-mentioned multi-source physical features.

[0128] After obtaining the partitions, the system will read the risk prediction level corresponding to each partition and assign a weighting coefficient based on this. The risk prediction level is usually a discrete value from 0 to 3, corresponding to low risk to high risk. The system assigns a higher weighting coefficient to high-risk areas based on the set risk-weight mapping function. This weighting coefficient plays a role in the subsequent calculation of status indicators. Its purpose is to increase the influence of high-risk area data on the overall status judgment. For example, if a partition is predicted to be level 3 high risk, its weighting coefficient may be set to 1.5; while the weighting coefficient of the level 0 area is 0.7. This mechanism effectively enhances the model's ability to respond to welding anomalies in key areas, which is conducive to foreseeing anomalies.

[0129] To calculate the melt pool stability index, the system focuses on analyzing the temporal trajectory of the melt pool boundary in visual image data. Specifically, the system extracts the melt pool boundary outline in each frame and calculates the frequency and amplitude of boundary changes between consecutive frames. For data segments with higher risk levels, the system prioritizes these boundary outline changes, forming a sample set that dominates the melt pool stability assessment. The system then performs a temporal overlay analysis on these samples, calculating the weighted average of their oscillation frequency and the dispersion of their amplitude changes, resulting in a quantitative index describing the fluctuation characteristics of the melt pool boundary during welding. Lower values ​​of this index indicate a more stable melt pool boundary, while larger fluctuations indicate uneven laser input energy or severe interface perturbations. For example, in the case of copper-steel dissimilar metal welding, due to the significant difference in the melting points of the two materials, the melt pool at the copper end exhibits significant fluctuations. During the initial heat input phase, the melt pool boundary amplitude reaches as high as 1.5 mm, with an oscillation frequency fluctuating around 20 Hz. In contrast, the steel end remains relatively stable. At this point, after extracting and weighting the dynamic features of the copper end using the high-frequency image sequence, the overall melt pool stability index is significantly lowered, prompting the control module to increase energy input balance adjustments in this area.

[0130] When calculating the instantaneous intermetallic fusion uniformity index, the system first identifies key frames within the multi-frame spectral signal—sampling moments where spectral line intensity changes significantly and contains distinct alloying element signatures. The system then selects several representative sampling points within each trajectory segment, extracts the spectral line intensities at specific wavelengths (e.g., Fe, Ni, Cu), and calculates the standard deviation of these line ratios. A larger standard deviation indicates an uneven distribution of elements within the fusion region, suggesting the presence of intermetallic diffusion barriers or uneven energy distribution at the weld instant. A risk-weighting mechanism is also applied here: standard deviations within high-risk regions are given a higher index adjustment influence, making them the dominant contributor to the output of the fusion uniformity index. For example, in the middle intersection of a Lissajous trajectory, due to reduced welding speed and high energy density, the Cu-Fe line intensity ratio in this region changes significantly (e.g., from 1.2 to 2.5), indicating a dramatic change in intermetallic mixing and poor fusion uniformity. Given this region's risk level of 3, the system increases its contribution weight, making the data from this region play a decisive role in the overall fusion uniformity index.

[0131] For the calculation of the interface temperature difference fluctuation index, the system focuses on analyzing the spatial variation of the temperature gradient in the infrared thermal imaging signal. Each trajectory partition corresponds to an image sampling area. The system first extracts the temperature values ​​of several sampling points in the area, calculates the extreme difference between the highest temperature and the lowest temperature within the area, and the gradient change rate of the temperature distribution. These indicators reflect whether the heat input is evenly distributed on the interface during the welding process. Afterwards, the system weightedly aggregates the temperature difference extreme indicators in all spatial partitions, where the weights are still set based on the risk level, and finally forms a global interface temperature difference fluctuation index. The more severe the temperature difference fluctuation, the more significant the uneven heat input at different metal interfaces during the welding process, which can easily lead to problems such as stress concentration, lack of fusion or cracks. Taking an actual welding as an example, in the lower right corner of the swing trajectory, due to the fast movement speed of the laser beam and insufficient heat accumulation, the temperature difference of the infrared image in this section is as high as 80°C, while the temperature difference in the upper left corner is only 40°C; after evaluating the risk level, the system gives the lower right corner section a higher weighting coefficient, resulting in a higher interface temperature difference fluctuation index in the final output, indicating that there is an energy distribution offset problem in this section, and the welding speed and swing frequency need to be adjusted through the control module.

[0132] For example, in a dissimilar metal laser welding process, assuming that the current welding task is the butt welding of copper and stainless steel components, the weld area uses an elliptical-straight line combination of laser oscillation trajectory for energy coverage. After the welding device is started, the acquisition module continuously collects the visual image, plasma excitation spectrum and infrared thermal imaging signal of the welding area at a high frame rate, while assigning a trajectory position label to each frame of data, and annotating the risk prediction level based on the initial stage stability prediction model. Assume that at this time, the C section (close to the copper substrate side) in the welding path is identified as risk level 3 due to the strong thermal conductivity of the material and low initial energy absorption. The visual image of the corresponding data segment shows that the molten pool boundary is discontinuous, the intensity of the Cu element spectral line in the spectral information changes dramatically, and the infrared image reflects that the temperature difference in this area is as high as 95°C.

[0133] After receiving the raw data, the computation module first divides the entire welding process into six trajectory zones, A to F, based on trajectory location labels. Each zone's spatial data set stores the image, spectral, and thermal imaging data within that interval, along with the risk level. The system then assigns a weight to each zone based on a pre-set risk-weight mapping function. For example, risk level 0 is assigned a weight of 0.8, level 1 is 1.0, level 2 is 1.2, and level 3 is 1.5. This ensures that data from high-risk areas has a greater impact in subsequent state modeling.

[0134] During the calculation of the melt pool stability index, the system focuses on processing the boundary contour features in the visual image data. For each frame, the system applies an image edge detection algorithm (such as Canny edge detection) to extract the outer contour of the melt pool and mark its boundary coordinates. By tracking the offset of the same boundary point in consecutive frames, the frequency and amplitude of boundary changes can be obtained. The system performs statistics for each partition separately and then combines the results according to the weight of the partition. Taking partition C as an example, its average boundary amplitude is 1.8mm and the change frequency is 22Hz, while most other partitions are less than 0.6mm and 12Hz. Therefore, partition C becomes the dominant indicator component area, and its weighted result significantly affects the overall index. The final melt pool stability index output value is 0.63 (normalized value, range 0–1), indicating that the system's current welding process has high-frequency unstable boundary fluctuations.

[0135] The calculation of the instantaneous intermetallic fusion uniformity index focuses on the excitation spectrum data. In each frame of the spectrum, the system extracts the spectral line intensities at key wavelengths such as Cu, Fe, and Cr, and calculates the standard deviation of the element ratios at different sampling points within the partition. In partition C, the Cu / Fe spectral line intensity ratios at multiple sampling points vary significantly. For example, point P1 is 1.1, P2 is 1.9, P3 is 2.3, and P4 is 1.3, with a standard deviation of 0.52. In contrast, low-risk partitions, such as partition A, have a standard deviation of only 0.08. Based on the high risk level and high ratio standard deviation of partition C, the system assigns it an index boost factor, allowing it to dominate the fusion uniformity results. The final index output value is 0.47 (normalized range 0–1), indicating significant uneven material mixing distribution and possible risks such as incomplete diffusion or interface inclusions.

[0136] The interface temperature difference fluctuation index is completed through the temperature data in the infrared thermal imaging image. In each trajectory partition, the system sets a number of fixed sampling points (for example, 5x5 grid distribution), extracts its temperature value, and calculates the maximum and minimum temperature difference (range) and the average gradient. Assume that the temperature range in partition C is 85°C and the average gradient is 22°C / mm, while the average values ​​of other partitions are only 35°C and 7°C / mm. The system aggregates the range and gradient data of all partitions according to the risk weight to form an overall temperature difference fluctuation index. The output result is 0.68, which is significantly higher than the empirical threshold of 0.4 for this type of welding process, further confirming that there is a problem of uneven heat input in this section.

[0137] By calculating the results of these three status indicators, the control module can further locate high-risk areas during the welding process and implement compensatory control strategies such as increasing laser power, refining oscillation frequency, and adjusting trajectory amplitude to achieve balanced welding quality and minimize defects. This complete set of status indicators forms the central data source of the closed-loop control chain. Its construction process comprehensively integrates spatial information, temporal dynamics, and risk assessment, and is a key foundation for intelligent perception and control of dissimilar metal welding.

[0138] The entire calculation process features three technological breakthroughs. First, the sampling area is explicitly divided by trajectory location and tied to the predicted risk level, providing a clear spatial reference and response basis for condition assessment. Second, a weighting mechanism is introduced. Instead of averaging all data, it enhances the data contribution of key areas through risk perception, strengthening the model's sensitivity to welding defects. Third, the condition indicators are designed to be quantifiable and calculated based on direct sensor data, rather than subjective evaluations, ensuring good repeatability and industrial application value.

[0139] In summary, the computation module described in this embodiment fully exploits the value of the physical feature information provided by the acquisition module through spatial partitioning, weight adjustment, time series analysis, and data fusion. This allows for precise modeling of the multidimensional state of welding quality, significantly enhancing the intelligent perception and control capabilities of the welding process. Through the module's outputs of the molten pool stability index, the instantaneous intermetallic fusion uniformity index, and the interface temperature difference fluctuation index, the control module can capture the multidimensional state changes of the welding process in real time, forming a feedback path that is highly coupled to the physical mechanism, ensuring a stable, efficient, and defect-free dissimilar metal welding process.

[0140] The control module 104 is used to receive multi-dimensional state indicators and perform joint optimization adjustment of laser parameters and trajectory parameters based on the control weight matrix, including real-time adjustment of laser power, welding speed, oscillation frequency and trajectory shape.

[0141] The control module 104 is the core component of the present invention for realizing the closed-loop adaptive control function. Its main function is to dynamically adjust key parameters of the laser welding head, such as laser power, welding speed, oscillation frequency and oscillation trajectory shape, based on the multi-dimensional state indicators output in real time by the calculation module 103, so as to effectively deal with the quality fluctuations caused by material differences, uneven heat diffusion or energy input offset during the welding of dissimilar metals, and ensure the stability of the welding process and the reliability of the final joint.

[0142] The control module is directly connected to the computing module in the system and receives three main status indicators from the computing module: the melt pool stability index, the instantaneous metal fusion uniformity index, and the interface temperature difference fluctuation index. Among them, the melt pool stability index is used to measure the degree of shape fluctuation of the melt pool boundary over a period of time. It is usually represented by the standard deviation of the maximum lateral width of the melt pool in multiple consecutive frames of images. The larger the value, the more unstable the melt pool. The instantaneous metal fusion uniformity index reflects the uniformity of the mixing of dissimilar metals in the welding pool. It is generally quantified by the distribution standard deviation of the emission intensity ratio of representative metal elements in the spectrum. The smaller the value, the more uniform the mixing. The interface temperature difference fluctuation index represents the fluctuation range of the temperature difference on both sides of the weld centerline per unit time. The difference between the maximum and minimum temperature differences is calculated using infrared thermal imaging data. The larger the value, the greater the heat input unevenness, and the potential for thermal stress accumulation or cracks.

[0143] A set of control weight matrices are preset inside the control module to determine the adjustment priority based on the importance of these three indicators and the degree of abnormality of the current values. For example, when the molten pool stability index is close to the set threshold (such as 0.3mm) and the other two indicators are still within the normal range, the system will give priority to reducing the laser power or slightly slowing down the welding speed to reduce the molten pool disturbance. When the metal fusion uniformity index exceeds the threshold (such as 0.2), the system will give priority to changing the shape of the swing trajectory, for example, from an ellipse to an "∞" shape, to extend the difference in the laser's residence time on the two metals at the interface and improve the lateral mixing efficiency.

[0144] At the control strategy execution level, the module quickly determines the status indicators within each second cycle and uses incremental adjustment to gradually optimize the welding parameters to prevent drastic changes from causing new instabilities. For example, if the interface temperature difference fluctuation index rises to 0.8°C in a certain cycle, the control module can reduce the oscillation frequency from 100Hz to 95Hz and slightly increase the oscillation amplitude to ensure that the laser has more time to transfer heat in the boundary area, thereby balancing the interface temperature distribution. After each parameter adjustment, the control module feeds the adjustment results back to the system and waits for data input and indicator updates in the next cycle to determine whether further adjustment is needed.

[0145] To prevent parameter oscillation or excessive control, the control module is equipped with an inhibition threshold zone. When the status indicators fluctuate within an acceptable range (such as a change of less than 5%), no adjustment will be performed temporarily to maintain the current welding parameters unchanged. When multiple indicators deviate from the set values ​​at the same time, the control module will adjust them in sequence according to the preset priority to prevent mutual interference caused by adjusting multiple parameters at the same time.

[0146] The entire control module is implemented using a programmable logic controller (PLC) or embedded control board and features a real-time scheduling mechanism for executing laser control commands, motion control commands (such as swing motor frequency), and responding to communication interruptions. By combining status indicators with a control weight matrix, the module completes a complete calculation-judgment-execution cycle within a single second, ensuring timely system response and stable regulation.

[0147] In order to more clearly illustrate the actual operation mechanism of the control module 104, a set of specific welding tasks is taken as an example to illustrate the entire process from receiving indicators, referring to the weight matrix, judging the status, executing adjustments to forming a feedback loop.

[0148] During an aluminum-steel dissimilar metal lap weld, the system periodically receives three status indicator inputs from the calculation module 103 every second. During a certain period, the molten pool stability index is 0.25 mm, the instantaneous metal fusion uniformity index is 0.22, and the interface temperature difference fluctuation index is 0.6 degrees Celsius. The control weight matrix preset in the system is as follows:

[0149] Status indicators Laser power Welding speed Oscillation frequency Swing trajectory shape Melt pool stability index 0.6 0.3 0.1 0 Metal fusion uniformity index 0.2 0.1 0.4 0.3 Interface temperature fluctuation index 0.4 0.2 0.3 0.1

[0150] This matrix was developed by engineers through multiple rounds of experiments. Specifically, during welding experiments, they individually adjusted a parameter and observed the changing trends of the three status indicators. Principal component analysis and regression analysis were used to deduce the relative influence of each parameter on each status indicator. This was then normalized to form the aforementioned proportional matrix. Each number in the matrix represents the priority weight of that control parameter for the corresponding status indicator. A higher weight indicates a more sensitive parameter to improving that indicator.

[0151] During this cycle, the molten pool stability index reached 0.25, close to the system-set threshold of 0.30, and was judged to be slightly unstable. The control module performs a weighted combination adjustment of the laser power (0.6), welding speed (0.3), and oscillation frequency (0.1) based on the corresponding weight of this index in the matrix. The system adopts a joint optimization strategy, that is, the adjustment range distribution is determined according to the proportion of the weighted coefficients. Taking the maximum acceptable total adjustment range set for this cycle as 10% as an example, the laser power will be adjusted by 6%, the welding speed by 3%, and the oscillation frequency by 1%. During execution, the laser power dropped from 2.2kW to approximately 2.07kW, the welding speed was reduced from 300mm / min to 291mm / min, and the oscillation frequency was slightly reduced from 100Hz to 99Hz.

[0152] At the same time, the fusion uniformity index was 0.22, exceeding the threshold of 0.2. The system identified this indicator as abnormal and triggered further joint adjustments. In the weight matrix corresponding to this indicator, the swing frequency (0.4) and the swing trajectory shape (0.3) accounted for a large proportion. Therefore, the system lowered the swing frequency by another 3%, from 99Hz to approximately 96Hz, and simultaneously switched the trajectory from the original elliptical shape to a "∞" shape (lemniscate) to increase the lateral scanning span of the laser beam and improve the lateral metal mixing uniformity.

[0153] For an interface temperature difference fluctuation index of 0.6, while not exceeding the set threshold (0.8), it is still critical. Based on the temperature difference index control matrix, the system only slightly increases the swing amplitude (corresponding to adjusting the trajectory energy coverage) from 1.5mm to 1.6mm to enhance the heat diffusion coverage area, but does not perform large-scale movements.

[0154] Once all parameters are adjusted, the control module writes the new parameter set to the laser head and motion control unit, generates an adjustment execution report, and returns it to the system's main control node. It then continues to collect and verify the adjustment results in the next cycle. If the adjusted indicators show a clear downward trend, the system will maintain the parameters and enter the observation state. If the indicators continue to deteriorate or other indicators show abnormalities, the system will restart the joint adjustment process based on the new status.

[0155] It can be seen from the above specific process that the role of the control module 104 in the present invention is not a passive response, but is based on a set of clear, adjustable, and quantitative weight systems to achieve real-time and refined linkage control of multiple indicators and multiple parameters, thereby providing stable and efficient quality control support for the laser welding process of dissimilar metals.

[0156] Furthermore, the control module is specifically configured to:

[0157] receiving the multi-dimensional status indicator including the trajectory position label and the risk prediction level output by the calculation module, and dividing the swing trajectory into a plurality of trajectory segments, each trajectory segment corresponding to a status indicator subset;

[0158] According to the risk prediction level of the trajectory segment, a trajectory segment-level control weight sub-matrix is ​​constructed to dynamically allocate the adjustment priority of the laser power, welding speed, oscillation frequency and trajectory shape among the trajectory segments;

[0159] Based on the spatial distribution differences of the state indicators in each trajectory segment, the trajectory segments with local significant abnormal characteristics are identified and given adjustment priority weights so that the control response first acts on the areas with significant state fluctuations;

[0160] According to the risk prediction level corresponding to each trajectory segment, the upper limit of the control adjustment amplitude is set, so that the parameters corresponding to the high-risk segments allow a wider range of laser power changes, swing amplitude adjustments or trajectory shape switching;

[0161] The parameter adjustment strategy within the entire swing cycle is jointly scheduled within and between trajectory segments, achieving enhanced control compensation for local quality fluctuation areas while maintaining the overall welding process continuity;

[0162] The adjusted laser parameters and trajectory parameters are synchronously sent to the laser welding head for implementing segmented response control to optimize the welding quality of the dissimilar metal joining area.

[0163] During the laser welding of dissimilar metals, the welding quality is affected by a variety of complex factors. Especially in the high-temperature, high-energy-density welding area, the dynamic changes of the molten pool, the fusion uniformity between the metals, and the thermal diffusion characteristics of the interface may fluctuate to varying degrees as the laser beam swing trajectory changes. In order to achieve precise control of this highly dynamic process, the control module must have the ability to make fine-grained response adjustments based on the state perception results. The control module proposed in this invention is based on this core idea and is established on the understanding of the coupling relationship between the swing trajectory and the physical state indicators, thereby achieving dynamic and segmented optimization of the laser parameters and trajectory parameters.

[0164] Specifically, the control module first receives the multi-dimensional state indicators output from the calculation module. These state indicators are derived from the fusion, modeling and analysis of multi-source data collected by the acquisition module during the welding process. They contain information closely related to welding quality, such as the molten pool stability index, the instantaneous metal fusion uniformity index and the interface temperature difference fluctuation index. At the same time, the state indicators are also accompanied by the specific position label of each sampling point on the swing trajectory, as well as the risk prediction level corresponding to the position, thereby providing an accurate mapping of the state in both spatial and risk dimensions. Based on this, the control module divides the entire swing trajectory into several trajectory segments in time and space order. Each segment corresponds to a continuous trajectory path and corresponds to an independent subset of state indicators.

[0165] For example, imagine a laser beam oscillating along a sinusoidal or Lissajous curve, with multiple rotations and intersections within a single oscillation cycle. In actual welding, due to the laser beam's prolonged dwell time in specific trajectory segments or concentrated local energy density, these areas are very likely to experience increased molten pool disturbance, uneven metal fusion, or significant temperature fluctuations. By analyzing a subset of status indicators from these segments, the control module can identify trajectory segments with significant quality fluctuations. For example, within a particular segment, if the molten pool stability index fluctuates dramatically, fusion uniformity decreases, and the temperature difference widens, then that segment can be identified as a high-risk segment.

[0166] After identifying high-risk sections, the control module constructs a set of trajectory-segment-level control weight sub-matrices for different trajectory sections. These weight sub-matrices are used to prioritize laser and trajectory parameter adjustments across different trajectory sections. Each control parameter, such as laser power, welding speed, oscillation frequency, and trajectory shape, has its adjustment priority adjusted based on the risk level of that section. The control module prioritizes these sections by assigning higher weights to them within the sub-matrix. For example, in a typical sub-matrix, a section with a "high" risk level would have a corresponding laser power adjustment weight of 0.8 and an oscillation frequency adjustment weight of 0.7, while a section with a "low" risk level would have a laser power adjustment weight of only 0.2 and an oscillation frequency adjustment weight of 0.1. This dynamic weighting mechanism allows control system resources to be focused on addressing areas most likely to cause quality defects.

[0167] Furthermore, the control module must not only identify the risk level, but also analyze the spatial distribution characteristics of the status indicators in the trajectory section. The key to this step is to extract areas with significant local abnormalities and assign them additional adjustment priority. For example, within a trajectory section, although the state of most areas is stable, there are violent vibrations at the edge of the molten pool at several locations on the trajectory boundary. Such local anomalies will be identified through spatial clustering analysis or boundary dynamic detection methods and will be assigned a higher local response weight in the control matrix. The resulting adjustment strategy will no longer be a unified adjustment of the entire section, but will achieve more refined spatial control, so that the control response is concentrated on the area where the problem is most prominent.

[0168] While assigning adjustment priorities, the control module also sets an upper limit for the control adjustment amplitude for each trajectory segment. Specifically, for trajectory segments judged to be high-risk, the allowable laser power variation range can be expanded, for example, from ±5% to ±15%; the swing amplitude adjustment range can also be increased, for example, from ±0.3mm to ±1mm; and the trajectory shape can even be switched from the original sinusoidal swing to an elliptical or Lissajous pattern to achieve a more uniform energy distribution. This strategy can provide greater adjustment space for high-risk areas after the control weights are determined, thereby improving the system's response flexibility and correction capabilities.

[0169] In order to ensure the continuity and stability of the overall welding process, the control module also needs to coordinate the control strategies between different trajectory segments. To this end, on the basis of the control within the trajectory segment, the module further performs joint scheduling between trajectory segments. This process must not only take into account the priority adjustment of high-risk segments, but also avoid abrupt parameter jumps between different segments, which will lead to new molten pool disturbances. For example, in the process of transitioning from a low-risk segment to a high-risk segment, the laser power should not be instantly increased from 80% to 100%, but should be smoothly transitioned with the support of the control algorithm to achieve seamless connection of continuous control trajectories. The joint scheduling mechanism ensures that the entire welding process can maintain the stability of the overall process parameters in the face of local disturbances, thereby improving the quality consistency of the entire weld.

[0170] After the control strategy is determined, the control module will synchronously send the calculated final laser parameters and trajectory parameters to the laser welding head, which will execute the adjustment commands at the physical level. It is worth noting that there is a time delay between the parameter issuance and the laser head response. Therefore, the control module needs to introduce a certain prediction mechanism to pre-schedule the control instructions based on the response characteristics of the welding head, the swing cycle and the welding speed to ensure timely and accurate response. In addition, in order to prevent frequent adjustments of the laser welding head from causing mechanical fatigue or control system instability, the control module will set the minimum adjustment interval and the minimum stable period according to the rate of change of state between sections, thereby balancing the contradiction between the control response rate and the life of the equipment.

[0171] Let's use a complete example to illustrate this. When welding aluminum-titanium dissimilar metal joints, local overheating or poor fusion at the interface is highly likely due to the significant differences in thermal conductivity and melting points between the two metals. The initial oscillation trajectory is set to an elliptical shape, with an oscillation frequency of 50 Hz and a power of 1.2 kW. The acquisition module identifies that at the two ends of the trajectory's long axis, the melt pool stability index rapidly decreases and the interface temperature difference fluctuation index increases due to a slightly longer laser beam dwell time. The calculation module identifies this area as a high-risk trajectory segment. Accordingly, the control module sets the power control weight to 0.9 and the oscillation frequency adjustment weight to 0.8 in the corresponding segment, allowing a maximum power increase of ±20%. Furthermore, to avoid new disturbances caused by sudden power changes, the control module introduces a time-series smoothing mechanism for the power change, linearly adjusting it to the new set value within five oscillation cycles. The oscillation trajectory then smoothly switches from an elliptical shape to a figure-eight trajectory within seven cycles, expanding the lateral coverage and improving the uniformity of heat energy distribution. After this section, the control module gradually calls back the parameters and restores to the standard trajectory configuration, thereby achieving dynamic compensation for local poor quality trends.

[0172] In summary, the control module in this invention provides a comprehensive laser welding control strategy with spatial positioning and risk response capabilities. It can provide segmented enhanced control of localized abnormal areas while ensuring overall weld stability. Its core lies in constructing a multidimensional control weight matrix based on state perception data, and implementing spatial segmentation of parameter adjustment, risk priority allocation, and time-series joint scheduling, thereby significantly improving the quality consistency, stability, and process robustness of the dissimilar metal joint welding process.

[0173] Although the present application is disclosed as above with the preferred embodiments, it is not intended to limit the present application. Any person skilled in the art may make possible changes and modifications without departing from the spirit and scope of the present application. Therefore, the scope of protection of the present application shall be based on the scope defined by the claims of the present application.

Claims

1. A dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring, characterized in that: include: A laser welding head for outputting a laser beam and applying welding energy to the dissimilar metal joint, the laser welding head having a beam swinging function for scanning the laser beam energy along a nonlinear trajectory within the welding area based on preset swing trajectory parameters, the swing trajectory parameters including trajectory type, swing frequency, and swing amplitude; The acquisition module is used to synchronously collect multi-source data of the welding area at a high frame rate during the laser welding process, including visual images of the weld pool, plasma excitation spectra, and infrared thermal imaging signals. It also extracts physical characteristic information characterizing the degree of metal mixing, interface diffusion behavior, and dynamic fluctuations of the weld pool based on multimodal fusion methods; a calculation module for performing dynamic feature modeling based on the physical feature information and the swing trajectory parameters, and calculating multidimensional state indicators characterizing welding quality using a time series feature extraction and spatial energy distribution analysis method, including a molten pool stability index, an instantaneous metal fusion uniformity index, and an interface temperature difference fluctuation index; The control module is used to receive multi-dimensional status indicators and perform joint optimization adjustment of laser parameters and trajectory parameters based on the control weight matrix, including real-time adjustment of laser power, welding speed, oscillation frequency and trajectory shape.

2. The dissimilar metal laser welding device based on swinging beam and molten pool state monitoring according to claim 1 is characterized in that: The acquisition module is specifically used for: In the initial stage of welding, a weld area stability prediction model based on the oscillation trajectory parameters of the laser welding head and the preset metal combination characteristics is established to predict the potential welding quality fluctuation risk level at each trajectory position point throughout the oscillation cycle; Based on the output of the stability prediction model, the swing trajectory is divided into multiple dynamic acquisition areas, and enhanced acquisition parameters are automatically configured for areas with high predicted risk levels, including improved image resolution, increased thermal image refresh rate, or extended spectral integration time; During the actual welding process, the acquisition module performs local enhanced sampling on high-risk areas according to the dynamic acquisition area configuration, and the spatial resolution or time density of the acquired visual images, plasma excitation spectra, and infrared thermal imaging signals is higher than that of the standard area; The data obtained by enhanced sampling is marked as high-confidence feature fragments, and its corresponding trajectory position label and risk prediction level are attached, which are used as key frame priority input in subsequent modeling; The physical feature information extracted from the standard area and the high-risk area is uniformly structured and output as physical feature information for characterizing the degree of metal mixing, interface diffusion behavior and dynamic fluctuations of the molten pool, and output to the calculation module for dynamic feature modeling based on the physical feature information.

3. The dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring according to claim 2 is characterized in that: The calculation module is specifically used for: receiving the physical feature information including the risk prediction level and the trajectory position label output by the acquisition module, and dividing the physical feature information into a plurality of spatial partition data segments according to the trajectory position, wherein each data segment corresponds to a sampling area under a swing trajectory; According to the risk prediction level, weighting coefficients are assigned to different spatial partition data segments, and the weighting coefficients are used to increase the contribution weight of the data corresponding to the high-risk area in the state indicator modeling; When calculating the melt pool stability index, priority is given to the dynamic information fragments from the boundaries with higher risk levels, and the frequency of boundary contour changes is weighted averaged by time series superposition to form a stability state result that reflects the dominant behavior of the high-fluctuation area; When calculating the instantaneous metal fusion uniformity index, the key frame data in the spectral feature information is selected, and the index is adjusted based on the standard deviation of the spectral ratio of different sampling points in the same swing trajectory area and the risk weight, so as to highlight the dominant contribution of abnormal fusion behavior; When calculating the interface temperature difference fluctuation index, the temperature gradient information of each spatial partition is used to calculate the temperature difference extreme within the region, and all partition results are aggregated into a global temperature difference fluctuation index according to risk weighting to achieve a sensitive response to the imbalance of heat input at the dissimilar metal interface.

4. The dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring according to claim 3 is characterized in that: The control module is specifically used for: receiving the multi-dimensional status indicator including the trajectory position label and the risk prediction level output by the calculation module, and dividing the swing trajectory into a plurality of trajectory segments, each trajectory segment corresponding to a status indicator subset; According to the risk prediction level of the trajectory segment, a trajectory segment-level control weight sub-matrix is ​​constructed to dynamically allocate the adjustment priority of the laser power, welding speed, oscillation frequency and trajectory shape among the trajectory segments; Based on the spatial distribution differences of the state indicators in each trajectory segment, the trajectory segments with local significant abnormal characteristics are identified and given adjustment priority weights so that the control response first acts on the areas with significant state fluctuations; According to the risk prediction level corresponding to each trajectory segment, the upper limit of the control adjustment amplitude is set, so that the parameters corresponding to the high-risk segments allow a wider range of laser power changes, swing amplitude adjustments or trajectory shape switching; The parameter adjustment strategy within the entire swing cycle is jointly scheduled within and between trajectory segments, achieving enhanced control compensation for local quality fluctuation areas while maintaining the overall welding process continuity; The adjusted laser parameters and trajectory parameters are synchronously sent to the laser welding head for implementing segmented response control to optimize the welding quality of the dissimilar metal joining area.

5. The dissimilar metal laser welding device based on oscillating beam and molten pool state monitoring according to claim 2 is characterized in that: The acquisition module is also used for: When the risk prediction level at a certain swing trajectory position reaches a set high-risk threshold, the visual image enhancement acquisition process is automatically initiated. The visual image enhancement acquisition process includes: performing multi-scale wavelet reconstruction processing on the visual image collected at the high-risk trajectory position, and combining edge-guided filtering with a local contrast enhancement algorithm to enhance the texture saliency of the weld pool edge, interface disturbance, and metal-to-metal detail structure contained in the visual image; Based on the grayscale histogram distribution information of the infrared thermal imaging signal collected at the high-risk trajectory location, the hot spot concentration index of the thermal imaging channel is dynamically calculated. When the hot spot concentration exceeds the adaptive saturation threshold, the slope of the response curve and the saturation upper limit parameters of the infrared thermal imaging signal are adjusted in real time, so that the infrared thermal imaging signal maintains the details of the high-temperature area while improving the overall contrast of the low-temperature area. Perform pixel-level alignment and inter-channel geometric registration on the enhanced visual image and the infrared thermal imaging signal after response curve adjustment to form composite enhanced frame data with spatial consistency and modal complementarity. The composite enhanced frame data synchronously inherits its corresponding trajectory position label and risk prediction level information. Structurally marking the composite enhanced frame data as high-confidence feature segments, and adding a texture complexity index and a hot spot adjustment parameter as auxiliary modeling factors during the structuring process to provide a basis for regional refinement weights when subsequently output to the calculation module; The high-confidence feature fragments are uniformly output as part of the physical feature information used to characterize the degree of metal mixing, interface diffusion behavior and dynamic fluctuations of the molten pool, and are fused with the standard area acquisition results and input into the calculation module to enhance the accuracy and response sensitivity of the modeling results at the high-risk trajectory position points.

Images