Welding trajectory optimization method in dual-cavity laser welding process

Abstract

The present invention discloses a welding trajectory optimization method in a dual-cavity laser welding process, which belongs to the field of laser welding, and includes: obtaining real-time ultrasonic signals, performing stress discrimination analysis, and obtaining real-time stress in the weld area; judging whether it is at a stress threshold; if not, issuing a preliminary adjustment instruction, and adjusting the laser beam of the dual-cavity laser welding to obtain a preliminary adjusted stress; judging whether it is at a stress threshold; if not, issuing a trajectory optimization instruction, and adjusting the predetermined welding trajectory of the dual-cavity laser welding to obtain an optimized welding trajectory. The present application solves the technical problem that stress control in the weld area during dual-cavity laser welding in the prior art is not ideal, which easily leads to a decrease in welding quality. By real-time monitoring of ultrasonic signals in the weld area, and dynamically optimizing the welding trajectory according to stress discrimination analysis, the technical effect of effectively controlling stress in the weld area and improving welding quality is achieved.

Description

Technical Field

[0001] The invention relates to the field of laser welding, and in particular to a welding trajectory optimization method in a dual-cavity mode laser welding process. Background Art

[0002] Double-cavity laser welding has the advantages of fast welding speed and high weld quality, and is widely used in the manufacturing industry. However, during the double-cavity laser welding process, due to the influence of factors such as large welding heat input and fast cooling rate, large residual stress is easily generated in the weld area, resulting in quality problems such as reduced weld joint strength and weld deformation, which seriously affects the performance and reliability of the welded structure. Therefore, in the prior art, the stress control in the weld area during double-cavity laser welding is not ideal, which easily leads to reduced welding quality. Summary of the invention

[0003] The present application provides a welding trajectory optimization method in a dual-cavity laser welding process, aiming to solve the technical problem in the prior art that stress control in the weld area in a dual-cavity laser welding process is not ideal, which easily leads to a decrease in welding quality.

[0004] The present application discloses a welding trajectory optimization method in a dual-cavity laser welding process, comprising: obtaining a real-time ultrasonic signal, wherein the real-time ultrasonic signal refers to an ultrasonic signal obtained by real-time monitoring of a weld area of ​​dual-cavity laser welding by the ultrasonic device; calling a predetermined stress identification strategy to perform stress identification analysis on the real-time ultrasonic signal to obtain a real-time stress in the weld area; judging whether the real-time stress is within a stress threshold; if not, issuing a preliminary adjustment instruction, and adjusting the laser beam of the dual-cavity laser welding based on the preliminary adjustment instruction to obtain a preliminary adjusted stress; judging whether the preliminary adjusted stress is within the stress threshold; if not, issuing a trajectory optimization instruction, and adjusting a predetermined welding trajectory of the dual-cavity laser welding based on the trajectory optimization instruction to obtain an optimized welding trajectory.

[0005] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0006] Since the real-time ultrasonic signal is obtained, the real-time ultrasonic signal refers to the ultrasonic signal obtained by the ultrasonic equipment to monitor the weld area of ​​the dual-cavity laser welding in real time, reflecting the actual stress state of the weld area; the predetermined stress identification strategy is called to perform stress identification analysis on the real-time ultrasonic signal to obtain the real-time stress of the weld area, and the ultrasonic signal is converted into real-time stress data of the weld area through stress identification analysis; it is determined whether the real-time stress is within the stress threshold value, and by determining the relationship between the real-time stress and the stress threshold value, it is determined whether the stress state of the current weld area meets the requirements; if the real-time stress is not within the stress threshold value, a preliminary adjustment instruction is issued, and based on the preliminary adjustment instruction, the laser beam of the dual-cavity laser welding is adjusted to obtain the preliminary adjusted stress. The preliminary adjustment instruction is generated according to the deviation between the real-time stress and the stress threshold value, and by adjusting the laser beam parameters, such as the power The rate, spot size, etc. can change the heat input in the weld area, thereby affecting the stress state; determine whether the stress after preliminary adjustment is within the stress threshold, and evaluate the effect of the preliminary adjustment instruction by re-determining the adjusted stress state to determine whether further optimization is needed; if the stress after preliminary adjustment is still not within the stress threshold, a trajectory optimization instruction is issued, and the predetermined welding trajectory of the dual-cavity laser welding is adjusted based on the trajectory optimization instruction to obtain an optimized welding trajectory, so as to change the heating history and cooling conditions of the weld area and further optimize the stress state. This technical solution solves the technical problem that the stress control in the weld area during the dual-cavity laser welding process in the prior art is not ideal, which easily leads to a decrease in welding quality. By real-time monitoring of the ultrasonic signal in the weld area and dynamically optimizing the welding trajectory according to stress discrimination analysis, the technical effect of effectively controlling the stress in the weld area and improving the welding quality is achieved.

[0007] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0008] Figure 1 A flow chart of a welding trajectory optimization method in a dual-cavity laser welding process is provided for an embodiment of the present application;

[0009] Figure 2 A flow chart of generating an optimized welding trajectory in a welding trajectory optimization method in a dual-cavity mode laser welding process is provided for an embodiment of the present application. DETAILED DESCRIPTION

[0010] The overall idea of ​​the technical solution provided by this application is as follows:

[0011] The embodiment of the present application provides a welding trajectory optimization method in a dual-cavity laser welding process. First, the weld area is monitored in real time by an ultrasonic device to obtain a real-time ultrasonic signal reflecting the stress state of the weld. Then, a predetermined stress discrimination strategy is called to analyze the real-time ultrasonic signal to obtain the real-time stress of the weld area. Next, the real-time stress is compared with the preset stress threshold to determine whether the current stress state meets the requirements. If not, a preliminary adjustment instruction is issued to achieve dynamic optimization control of the weld stress by adjusting the laser beam parameters and process parameters such as the welding trajectory. If the requirements are still not met, a trajectory optimization instruction is further generated to achieve secondary optimization of the weld stress by adjusting process parameters such as the welding trajectory. By repeatedly iterating the optimization until the weld stress state meets the requirements, the stress in the weld area is effectively controlled and the welding quality is improved.

[0012] After introducing the basic principles of the present application, the non-limiting implementation methods of the present application will be specifically introduced below in conjunction with the drawings in the specification.

[0013] like Figure 1 As shown, an embodiment of the present application provides a welding trajectory optimization method in a dual-cavity mode laser welding process. The method is applied to a welding trajectory optimization system in a dual-cavity mode laser welding process. The welding trajectory optimization system in a dual-cavity mode laser welding process is communicatively connected with an ultrasonic device.

[0014] Specifically, dual-cavity laser welding uses two laser beams to weld the workpiece at the same time. Compared with single-cavity laser welding, dual-cavity laser welding can provide higher welding efficiency and weld quality. The welding trajectory optimization method in the dual-cavity laser welding process provided in the embodiment of the present application is applied to the welding trajectory optimization system in the dual-cavity laser welding process to realize real-time monitoring and dynamic optimization control of the dual-cavity laser welding process. Among them, the welding trajectory optimization system in the dual-cavity laser welding process is connected to the ultrasonic device through a communication interface to realize data interaction. During the dual-cavity laser welding process, the ultrasonic device monitors the weld area in real time, obtains ultrasonic signals reflecting the internal structure and defects of the weld, and transmits them to the welding trajectory optimization system in real time. The system quickly evaluates the weld quality based on the received ultrasonic signals, and determines whether it is necessary to optimize the laser beam parameters or adjust the welding trajectory, thereby realizing closed-loop optimization control of the dual-cavity laser welding process to ensure that the weld quality is stable within the set range.

[0015] The welding trajectory optimization method in the dual-cavity laser welding process includes:

[0016] S1: Acquire a real-time ultrasonic signal, where the real-time ultrasonic signal refers to an ultrasonic signal obtained by real-time monitoring of a weld area of ​​dual-cavity laser welding by the ultrasonic device.

[0017] Specifically, the ultrasonic equipment monitors the weld area in the dual-cavity laser welding process in real time, obtains real-time ultrasonic signals reflecting the internal structure and defects of the weld, and transmits them to the welding trajectory optimization system in real time through the communication interface. Among them, the ultrasonic equipment can adopt ultrasonic detection methods such as pulse-echo method and transmission method, and select ultrasonic probes, detection parameters and coupling methods according to specific application requirements. Among them, the ultrasonic probe can be integrated with the laser welding head to realize synchronous detection of the weld area; it can also track and detect the weld area at a certain distance behind the laser welding head. The real-time ultrasonic signal obtained by the ultrasonic equipment includes multiple parameters such as ultrasonic propagation time, echo amplitude, and spectrum characteristics. These parameters are closely related to defects such as weld penetration, weld width, pores, and cracks, and can be used to evaluate weld quality. In order to facilitate subsequent data processing and analysis, the ultrasonic signal is digitized into a series of discrete data points.

[0018] The real-time ultrasonic signal of the weld area is obtained by ultrasonic equipment, which provides a reliable basis for subsequent weld stress identification, laser parameter optimization, welding trajectory adjustment and other steps.

[0019] S2: Retrieving a predetermined stress identification strategy to perform stress identification analysis on the real-time ultrasonic signal to obtain the real-time stress of the weld area.

[0020] Specifically, the welding trajectory optimization system calls the predetermined stress identification strategy, performs stress identification analysis on the real-time ultrasonic signal, and obtains the real-time stress in the weld area. The so-called predetermined stress identification strategy refers to the method and rules for judging the stress size and distribution in the weld area based on the characteristic parameters of the ultrasonic signal. The predetermined stress identification strategy is based on a large amount of test data and theoretical analysis to establish the corresponding relationship between ultrasonic signal parameters and weld stress, including time domain analysis method, frequency domain analysis method, wavelet analysis method, etc.

[0021] For example, the predetermined stress identification strategy is based on the wavelet analysis method. First, the real-time ultrasonic signal is decomposed in time and frequency using wavelet transform to extract wavelet coefficients at different scales; then, the wavelet characteristic parameters reflecting the weld stress are calculated based on the energy distribution characteristics of the wavelet coefficients; then, the wavelet characteristic parameters are input into the pre-established stress prediction model to obtain the real-time stress values ​​at different positions in the weld area. Among them, the establishment of the stress prediction model requires a large amount of training data, which is obtained through finite element numerical simulation or physical testing. In the training data, each set of ultrasonic signal parameters corresponds to the stress distribution at different positions in the weld area. The training data is learned through machine learning algorithms (such as support vector machines, neural networks, etc.) to obtain a stress prediction model. Once the stress prediction model is established, it can be used for stress identification analysis of real-time ultrasonic signals.

[0022] S3: Determine whether the real-time stress is within a stress threshold.

[0023] Specifically, by performing stress discrimination analysis on the real-time ultrasonic signal, the real-time stress at different positions in the weld area is obtained, and it is compared with the preset stress threshold to determine whether the stress level of the weld is reasonable. Among them, the stress threshold is pre-set based on factors such as the mechanical properties of the welding material, the stress characteristics of the weld joint, and the requirements of the use environment, and represents the safe range of residual stress in the weld. The selection of the stress threshold can refer to relevant welding specifications and standards, or it can be obtained based on production experience and test data statistics.

[0024] When judging whether the real-time stress is within the stress threshold range, the characteristic stress value of the real-time stress in the weld area is compared with the stress threshold. The characteristic stress value includes the average stress, the maximum stress, the root mean square value of the stress, etc. When the characteristic stress value exceeds the stress threshold, it means that the weld stress is too large and the welding process parameters need to be adjusted.

[0025] S4: If not, issue a preliminary adjustment instruction, and adjust the laser beam of the dual-cavity laser welding based on the preliminary adjustment instruction to obtain the preliminary adjusted stress.

[0026] Specifically, when it is determined that the real-time stress in the weld area exceeds the preset stress threshold, the welding trajectory optimization system issues a preliminary adjustment instruction and dynamically adjusts the laser beam parameters of the dual-cavity laser welding according to the instruction to reduce the stress to an acceptable level, thereby obtaining the real-time stress distribution after preliminary adjustment, that is, the preliminary adjusted stress.

[0027] The preliminary adjustment instructions include a series of adjustment strategies for laser parameters, such as increasing or decreasing laser power, changing laser spot diameter, adjusting laser pulse frequency, etc. The preliminary adjustment instructions are obtained through theoretical analysis and experimental optimization based on factors such as the thermophysical properties of the welding material, the geometric dimensions of the weld, and the stress distribution characteristics. For example, the preliminary adjustment instructions are generated based on a pre-established laser parameter-weld stress mapping model. The model is trained with a large amount of experimental data and can achieve a quantitative correlation between laser parameters and weld stress. When the real-time stress of the weld is input, the model quickly outputs the corresponding optimal laser parameter combination, thereby guiding the dynamic adjustment of the laser beam. Among them, the laser beam of the dual-cavity laser welding needs to be adjusted by controlling the output of the two lasers at the same time. In order to coordinate the effects of the two laser beams, the master-slave control method is adopted, that is, the parameters of the master laser are adjusted first, and then the parameters of the slave laser are matched according to the output characteristics of the master laser to ensure the stability and repeatability of the dual-cavity laser welding process.

[0028] By preliminarily adjusting the laser parameters, rapid feedback control of the weld residual stress was achieved, fully utilizing the flexibility and adjustability of dual-cavity mode laser welding, providing favorable initial conditions for welding trajectory optimization.

[0029] S5: Determine whether the stress after the preliminary adjustment is within the stress threshold.

[0030] Specifically, by adjusting the laser parameters of dual-cavity laser welding, the stress is improved to a certain extent, and the real-time stress after preliminary adjustment is obtained. However, simply adjusting the laser parameters cannot guarantee that the weld stress will meet the requirements, and a secondary judgment of the adjusted stress distribution is required. The characteristic stress values ​​(such as average stress, maximum stress, etc.) of the preliminary adjusted stress in the weld area are compared with the preset stress threshold. If it is judged that the stress after preliminary adjustment meets the requirements of the stress threshold, it means that simply adjusting the laser parameters can effectively control the weld stress without further optimizing the welding trajectory. In this case, continue to perform welding operations according to the current laser parameter settings and welding trajectory, and monitor the weld stress in real time to ensure that it is always within a safe and controllable range. On the contrary, if the stress after preliminary adjustment still exceeds the stress threshold, it means that it is difficult to meet the requirements of residual stress control by simply adjusting the laser parameters, and further optimization is required from the perspective of the welding trajectory.

[0031] Through two stress judgments, the effect of laser parameter adjustment can be evaluated more accurately, avoiding blind adjustment of welding trajectory, thereby shortening the number of iterations and time cost of the entire control process.

[0032] S6: If not, issue a trajectory optimization instruction, and adjust the predetermined welding trajectory of the dual-cavity laser welding based on the trajectory optimization instruction to obtain an optimized welding trajectory.

[0033] Specifically, when it is judged that the stress after the initial adjustment still exceeds the preset stress threshold, the welding trajectory optimization system will issue a trajectory optimization instruction, and intelligently adjust the predetermined welding trajectory of the dual-cavity laser welding according to the instruction to obtain an optimized welding trajectory, thereby further reducing the weld stress and improving the welding quality. Among them, the trajectory optimization instruction contains a series of welding trajectory adjustment strategies, such as changing the welding speed, adjusting the geometry of the welding path, and optimizing the topological structure of the welding trajectory. The formulation of these adjustment strategies needs to fully consider the stress characteristics of the welding joint, the stress distribution law, the interaction mechanism between the laser and the material, and other factors, and use intelligent optimization algorithms for solution. For example, the trajectory optimization instruction is generated based on the genetic algorithm. The genetic algorithm is a heuristic search method that simulates the biological evolution process. By encoding, crossing, mutating and other operations on the welding trajectory parameters, it continuously iterates and optimizes to obtain the optimized welding trajectory that meets the constraints. In the optimization process, the minimization of weld stress is used as the optimization objective function, and the laser parameters, welding speed, etc. are used as constraints to obtain the welding trajectory solution with the best comprehensive performance. Optimizing the welding trajectory is the optimal path that can significantly improve the stress state of the weld. Compared with the predetermined welding trajectory, the optimized welding trajectory can effectively reduce welding stress and improve the comprehensive performance of the weld.

[0034] Through intelligent optimization of the predetermined welding trajectory, the process potential of dual-cavity laser welding is maximized, providing a guarantee for obtaining high-quality welds.

[0035] Furthermore, the embodiment of the present application also includes:

[0036] Collecting multi-dimensional features of a predetermined workpiece for dual-cavity laser welding to obtain workpiece material feature information;

[0037] According to the predetermined stress identification strategy, the workpiece material feature information is traversed and matched in the workpiece database to obtain the best matching workpiece data;

[0038] Comparing the real-time ultrasonic signal with the best-matching ultrasonic signal in the best-matching workpiece data to obtain a real-time signal comparison result;

[0039] The predetermined stress feedback coefficient is retrieved to analyze the real-time signal comparison result to obtain the real-time stress.

[0040] In a feasible implementation, first, determine the key material characteristic parameters that affect the stress distribution of the weld, including the material's thermophysical properties (such as thermal conductivity, specific heat capacity), mechanical properties (such as yield strength, Young's modulus), metallurgical properties (such as melting point, phase change temperature), etc. By consulting the material manual, standard database, or performing material performance testing, obtain the multi-dimensional material characteristic information of the predetermined workpiece, that is, the workpiece material characteristic information. Then, according to the predetermined stress identification strategy, the workpiece material characteristic information is traversed and matched in the workpiece database to find historical workpiece data similar to the predetermined workpiece material characteristics, and obtain the most matching workpiece data, which provides a reference for subsequent stress identification. The workpiece database stores a large amount of welding data of workpieces of different materials and specifications, including material characteristic information, ultrasonic signals, weld stress distribution, etc. When performing data matching, the priority and weight coefficient of material feature matching are determined according to the predetermined stress identification strategy. For example, characteristic parameters that have a greater impact on weld stress, such as thermal conductivity, thermal expansion coefficient, and yield strength, are given higher weights, while parameters such as melting point and phase change temperature are given lower weights. Then, commonly used similarity measurement methods such as weighted Euclidean distance and cosine similarity are used to calculate the similarity of material characteristics between the workpiece to be welded and each historical workpiece in the database, and the several groups of data with the highest similarity are found as the most matching workpiece data.

[0041] Subsequently, a group of ultrasonic signals that are closest to the current welding position and process parameters are extracted from the most matching workpiece data as the most matching ultrasonic signal. This signal represents the typical characteristics of the stress distribution of the weld under similar material properties and welding conditions. Then, the real-time ultrasonic signal and the most matching ultrasonic signal are compared and analyzed in multiple dimensions such as time domain, frequency domain, time-frequency domain, etc., to obtain a series of characteristic quantities reflecting the degree of signal difference, such as correlation coefficient, mean square error, spectrum deviation, etc., which constitute the real-time signal comparison result. After that, the signal difference characteristic quantity is converted into a quantitative representation of the weld stress by using the correlation between the ultrasonic signal and the stress. Among them, the predetermined stress feedback coefficient is a signal-stress mapping model. By machine learning training of a large number of ultrasonic signals under known stress states, the functional relationship between the signal characteristic quantity and the stress component is obtained. When applying this model, the obtained real-time signal comparison results are input into the model, and after calculations such as multivariate regression and neural network, the real-time stress at different positions and directions of the weld is obtained, which provides a direct basis for the subsequent optimization of the welding process.

[0042] Furthermore, the embodiment of the present application also includes:

[0043] Extracting any artifact data of any artifact in the artifact database;

[0044] Performing similarity analysis on any workpiece material feature information in the any workpiece data and the workpiece material feature information to obtain any similarity;

[0045] When the arbitrary similarity satisfies the similarity threshold, any ultrasonic signal in the arbitrary workpiece data is used as the best matching ultrasonic signal.

[0046] In a feasible implementation, first, before matching the workpiece material features, any workpiece data of any workpiece is read from the workpiece database. Any workpiece refers to each historical workpiece in the workpiece database, and any workpiece data includes data of multiple dimensions such as material feature information, ultrasonic signals, welding process parameters, etc. of the workpiece. For example, random access, sequential traversal, etc. are used to extract the workpiece data in the database one by one. In order to improve data processing efficiency, the workpiece database can be indexed and partitioned according to prior information such as material type and thickness specifications of the predetermined workpiece to narrow the search range of candidate workpieces. Then, any workpiece material feature information, including thermophysical parameters, mechanical performance parameters, etc., is extracted from the arbitrary workpiece data, and paired with the collected workpiece material feature information. Then, according to a predetermined similarity measurement method, such as Euclidean distance, Mahalanobis distance, KL divergence, etc., the similarity between the two sets of material feature information is calculated to obtain any similarity between any workpiece and the predetermined workpiece. After obtaining any similarity of any workpiece, it is compared with a preset similarity threshold value. When it is greater than or equal to the similarity threshold value, any ultrasonic signal in the arbitrary workpiece data is used as the most matching ultrasonic signal.

[0047] Furthermore, the embodiment of the present application also includes:

[0048] Obtain real-time laser feature information;

[0049] The real-time laser characteristic information is screened in combination with a predetermined laser index to obtain a real-time laser parameter set;

[0050] Using the real-time laser parameter set as input information of a laser welding heat prediction model to obtain real-time laser welding heat;

[0051] Randomly generate a first laser beam adjustment scheme, and obtain a first predicted laser welding heat through analysis of the laser welding heat prediction model;

[0052] When a first heat difference between the first predicted laser welding heat and the real-time laser welding heat meets a heat unit adjustment threshold, the laser beam is adjusted according to the first laser beam adjustment scheme.

[0053] In a feasible implementation, first, the real-time monitoring and collection of the laser beam state during the dual-cavity laser welding process is performed to obtain real-time laser characteristic information. Real-time laser characteristic information refers to a series of parameters that reflect the characteristics of the laser beam energy distribution, spot shape, polarization state, etc., which determine the interaction mode between the laser and the material and the welding effect. For example, special monitoring devices such as photoelectric sensors, grating spectrometers, and polarization cameras are installed in the optical path system of the dual-cavity laser welding equipment to measure and record the various characteristic parameters of the laser beam in real time. These devices are connected to the CNC system through a high-speed data acquisition card, which can realize online monitoring and feedback control of laser parameters. Then, the collected real-time laser characteristic information is analyzed and extracted through predetermined laser indicators to obtain a real-time laser parameter set. The predetermined laser indicators are a series of quantitative indicators for evaluating the quality and stability of the laser beam, which are predetermined according to the laser welding principle and empirical formula, including laser power, laser beam diameter, laser beam stability, incident angle, etc. When screening laser parameters, the real-time laser feature information is first preprocessed, such as denoising filtering, data compression, etc., to improve the quality and recognizability of the signal; then, according to the predetermined laser indicators, the processed feature information is quantitatively analyzed to extract various indicator parameters; then, the extracted indicator parameters are organized into a structured data set according to type and time series to form a real-time laser parameter set.

[0054] Subsequently, the real-time laser parameter set of the current laser beam is quickly evaluated and predicted using the pre-established laser welding heat prediction model. The laser welding heat prediction model is a mathematical model that quantitatively describes the evolution law of the laser welding temperature field based on the physical mechanism of the interaction between laser and material, combined with finite element numerical simulation and machine learning methods. The obtained real-time laser parameter set is input into the prediction model, and the real-time laser welding heat is obtained after a series of nonlinear mapping and iterative calculations. Then, under the current laser parameters, the laser beam is adjusted exploratoryly, and the welding heat input effect after adjustment is evaluated. The first laser beam adjustment scheme is a set of candidate laser parameter combinations. Specifically, a series of adjustment schemes are randomly generated within a predetermined laser parameter adjustment range using intelligent optimization methods such as random search and genetic algorithms. Each adjustment scheme corresponds to a new set of laser parameter sets, representing a potential laser beam optimization approach. Then, these new laser parameter sets are input one by one into the laser welding heat prediction model used, and a series of predicted welding heat values ​​are obtained after forward calculation to form the first predicted laser welding heat.

[0055] Afterwards, the candidate first laser beam adjustment schemes are evaluated and screened, and the actual laser beam parameters are adjusted and optimized online according to the evaluation results. The first heat difference refers to the difference between the first predicted laser welding heat and the obtained real-time laser welding heat, which reflects the degree of deviation between the current laser parameters and the expected optimization target. In order to adjust the effect of the laser parameters, the heat unit adjustment threshold is pre-set, which represents the improvement in welding heat that can be brought about by the laser parameter adjustment amount. When evaluating the laser beam adjustment scheme, the first heat difference of each candidate scheme is first calculated, and then compared with the heat unit adjustment threshold. If the first heat difference is greater than or equal to the threshold, it means that the adjustment scheme can significantly improve the welding heat input and is worthy of further optimization; conversely, if the first heat difference is less than the threshold, it means that the optimization effect of the adjustment scheme is limited and can be ignored. When the first laser beam adjustment scheme that meets the heat unit adjustment threshold is found, the laser parameter settings therein can be sent to the control unit of the laser to adjust the actual laser beam online.

[0056] Furthermore, the embodiment of the present application also includes:

[0057] Extracting a first historical welding record from the historical dual-cavity mode laser welding records, wherein the first historical welding record includes a first historical laser parameter set and a first historical molten pool parameter set;

[0058] Performing normalization analysis on first molten pool data obtained by traversing the first historical molten pool parameter set based on a predetermined molten pool index to obtain a first historical laser welding heat;

[0059] Machine learning is performed on a first training data group formed based on the first historical laser parameter set and the first historical laser welding heat to obtain the laser welding heat prediction model.

[0060] In a preferred embodiment, the historical dual-cavity laser welding record refers to a series of structured data obtained through process monitoring and quality inspection in past production practices, which reflects the process parameters and weld characteristics of dual-cavity laser welding under different conditions. The historical dual-cavity laser welding records are traversed, and one record is selected each time as the first historical welding record. The first historical welding record is composed of two major types of parameters, one is the first historical laser parameter set, including laser power, spot diameter, focal position, scanning speed and other process parameters that affect the welding process; the other is the first historical molten pool parameter set, including molten pool size, molten pool temperature, melting depth, solidification time and other characteristic parameters that reflect the weld formation state.

[0061] Then, based on the predetermined molten pool index, the first historical molten pool parameter set is traversed to obtain the first molten pool data for normalization analysis to obtain the first historical laser welding heat. The predetermined molten pool index refers to a series of physical quantities or dimensionless parameters that are predetermined according to the laser welding principle and the molten pool dynamics theory to evaluate the molten pool formation process and heat transfer characteristics. By screening and combining the first historical molten pool parameter set, a series of first molten pool data corresponding to the predetermined molten pool index are obtained. Since the dimensions and numerical ranges of different molten pool indicators vary greatly, in order to facilitate the comprehensive evaluation of the welding heat input level, the first molten pool data is normalized and analyzed, such as minimum-maximum normalization, and the data of different indicators are mapped to the same dimensionless interval through mathematical transformation to obtain the normalized first historical laser welding heat.

[0062] Subsequently, a first training data set is formed, which is a plurality of data pairs consisting of a first historical laser parameter set (as an input feature) and a first historical laser welding heat (as an output target) in one-to-one correspondence. The first training data set is input into a machine learning model, such as a support vector machine, a neural network, a decision tree, etc., and a nonlinear mapping relationship between laser parameters and welding heat is established through training and learning. By processing other historical welding records in the historical dual-cavity mode laser welding records and performing model training, a laser welding heat prediction model is obtained.

[0063] Furthermore, the predetermined molten pool index includes at least molten pool size and molten pool stability.

[0064] In a feasible implementation, the predetermined molten pool index includes molten pool size and molten pool stability. Among them, molten pool size is an important index for evaluating the heat input characteristics of laser welding. In the process of dual-cavity laser welding, the formation and development of the molten pool directly affect the geometric size and forming quality of the weld. The larger the molten pool, the higher the coupling efficiency of the laser energy. The more heat the weldment absorbs per unit time, and the better the fusion state of the weld joint. Molten pool stability is another important index for evaluating the reliability of the laser welding process and the consistency of weld quality. In actual dual-cavity laser welding, due to fluctuations in factors such as material properties, welding process, and equipment status, the molten pool often presents certain dynamic change characteristics, such as the fluctuation of the molten pool shape, the turbulence of the molten flow, etc. The worse the molten pool stability, the more likely it is to have welding defects, such as pores, cracks, and undercuts. By predetermining the molten pool index, the molten pool data of the historical molten pool parameter set can be normalized and analyzed to obtain the historical laser welding heat, which provides support for the establishment of a laser welding heat prediction model.

[0065] Further, such as Figure 2 As shown, the embodiment of the present application also includes:

[0066] Reading welding trajectory characteristics, and generating trajectory optimization space in combination with the predetermined welding trajectory;

[0067] Extracting a first optimization scheme in the trajectory optimization space, and simulating the first optimization scheme to obtain a first optimized stress;

[0068] When the first optimized stress is within the stress threshold, the optimized welding trajectory is generated according to the first optimization scheme.

[0069] In a preferred embodiment, first, the welding trajectory characteristics are read, and the trajectory optimization space is generated in combination with the predetermined welding trajectory. The welding trajectory characteristics refer to a series of geometric parameters and kinematic parameters that reflect the shape and movement law of the welding trajectory, such as the coordinates of the starting and ending points of the trajectory, the trajectory line type (straight line, arc, curve, etc.), the trajectory length, the welding speed, the acceleration, etc. The trajectory optimization space refers to a parameter space containing all possible trajectory optimization schemes, and each dimension of the space corresponds to an adjustable trajectory parameter, such as the position coordinates of the trajectory node, the curvature radius of the trajectory segment, the speed value, etc. The range of the trajectory optimization space is jointly determined by the value range of each parameter. In the trajectory optimization space, each point represents a potential optimization scheme. The first optimization scheme is extracted from the trajectory optimization space by random selection, traversal search, etc., and the first optimization scheme is simulated, that is, the temperature field and stress field distribution of the welding trajectory corresponding to the scheme in the actual welding process are calculated by numerical simulation and physical model, and a comprehensive evaluation is performed to obtain the first optimized stress reflecting the stress state of the weld. Subsequently, the first optimization scheme is judged and screened by the stress threshold value. When the first optimization stress falls within the range of the stress threshold value, the first optimization scheme is considered feasible, and an optimized welding trajectory is generated accordingly.

[0070] Furthermore, the embodiment of the present application also includes:

[0071] The welding trajectory characteristics include trajectory speed and trajectory angle;

[0072] Taking a predetermined trajectory speed in the predetermined welding trajectory as a speed optimization upper limit, obtaining a speed optimization interval of the trajectory speed;

[0073] Taking a predetermined trajectory angle in the predetermined welding trajectory as an angle optimization restriction, obtaining an angle optimization interval of the trajectory angle;

[0074] The trajectory optimization space is constructed according to the speed optimization interval and the angle optimization interval.

[0075] In a preferred embodiment, the welding trajectory characteristics include two parameters: trajectory speed and trajectory angle. The trajectory speed refers to the speed of the laser beam moving along the welding trajectory; the trajectory angle refers to the inclination angle of the welding trajectory in the spatial plane, which is expressed by the angle with the horizontal plane or the vertical plane. The predetermined welding trajectory refers to the welding trajectory initially input into the dual-cavity laser welding system, which includes a predetermined trajectory speed and a predetermined trajectory angle. Based on the predetermined trajectory speed, the optimization range of the trajectory speed is obtained.

[0076] The predetermined trajectory speed is taken as the upper limit of the optimization interval, and then the appropriate lower limit is set according to factors such as speed adjustment accuracy and equipment response characteristics to obtain the speed optimization interval. Within this interval, the trajectory speed can be adjusted and optimized to maximize the welding efficiency and weld quality while ensuring welding safety. For example, if the predetermined trajectory speed is 50 mm / s, the speed adjustment accuracy is 1 mm / s, and the minimum stable speed of the equipment is 10 mm / s, the speed optimization interval can be set to [10, 50] mm / s. In the subsequent construction of the trajectory optimization space, the value of the trajectory speed dimension is limited to this interval. Similar to the setting of the speed optimization interval, the optimization interval of the trajectory angle is also based on the predetermined trajectory angle in the predetermined welding trajectory, and then comprehensively considered factors such as angle adjustment accuracy and welding torch posture constraints. Among them, the predetermined trajectory angle is set according to the stress characteristics of the weld and the welding process requirements. For example, lap welds, butt welds, fillet welds, etc. have different requirements for the trajectory angle. During the optimization process, the adjustment range of the trajectory angle should not be too large, so as not to cause drastic changes in the weld geometry and internal structure, affecting the service performance of the joint. Therefore, the predetermined trajectory angle is used as the constraint condition for angle optimization, and the upper and lower floating range of the angle is appropriately relaxed to obtain the angle optimization range. For example, if the predetermined trajectory angle is 60°, the angle adjustment accuracy is 1°, and the posture constraint is ±15°, the angle optimization range can be set to [45°, 75°].

[0077] After obtaining the optimization intervals of the two parameters of trajectory speed and trajectory angle, the trajectory optimization space is obtained by performing a Cartesian product operation on the two intervals. The trajectory optimization space is a two-dimensional parameter space, in which one dimension represents the trajectory speed and the other dimension represents the trajectory angle. Specifically, within the speed optimization interval, a series of discrete speed values ​​are generated with the speed adjustment accuracy as the step size; similarly, within the angle optimization interval, a series of discrete angle values ​​are generated with the angle adjustment accuracy as the step size. Then, by combining the speed values ​​and the angle values ​​in pairs, all discrete points in the trajectory optimization space are obtained, and each point corresponds to a potential optimization solution, thereby obtaining the trajectory optimization space.

[0078] In summary, the welding trajectory optimization method in the dual-cavity laser welding process provided by the embodiment of the present application has the following technical effects:

[0079] Acquire real-time ultrasonic signals. Real-time ultrasonic signals refer to ultrasonic signals that are monitored in real time by ultrasonic equipment on the weld area of ​​dual-cavity laser welding, providing input information for subsequent stress discrimination analysis. Retrieve the predetermined stress discrimination strategy to perform stress discrimination analysis on the real-time ultrasonic signal, obtain the real-time stress in the weld area, and realize the quantitative characterization of the stress state of the weld. Determine whether the real-time stress is within the stress threshold, and determine whether the current weld stress state meets the quality requirements, so as to provide a decision basis for subsequent optimization control. If not, issue a preliminary adjustment instruction, and adjust the laser beam of the dual-cavity laser welding based on the preliminary adjustment instruction to obtain the preliminary adjusted stress. According to the degree to which the real-time stress deviates from the target threshold, generate the corresponding laser beam parameter adjustment instruction, and preliminarily optimize the weld stress by changing parameters such as laser power and focal position. Determine whether the stress after preliminary adjustment is within the stress threshold, evaluate the effect of preliminary optimization, and provide a basis for further optimization. If it is not, a trajectory optimization instruction is issued, and the predetermined welding trajectory of the dual-cavity mold laser welding is adjusted based on the trajectory optimization instruction to obtain an optimized welding trajectory. When the initial optimization effect is not ideal, the weld stress is optimized secondary by adjusting the welding trajectory to further improve the welding quality.

[0080] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0081] Furthermore, the first or second mentioned above may not only represent an order relationship, but may also represent a specific concept, and / or refer to multiple elements that can be selected individually or in full. Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application intends to include these changes and variations.

Claims

1. A welding trajectory optimization method in a dual-cavity laser welding process, characterized in that: The welding trajectory optimization method in the dual-cavity mode laser welding process is applied to the welding trajectory optimization system in the dual-cavity mode laser welding process, the welding trajectory optimization system in the dual-cavity mode laser welding process is communicatively connected with the ultrasonic device, and the welding trajectory optimization method in the dual-cavity mode laser welding process includes: Acquiring a real-time ultrasonic signal, wherein the real-time ultrasonic signal refers to an ultrasonic signal obtained by real-time monitoring of a weld area of ​​dual-cavity laser welding by the ultrasonic device; Retrieving a predetermined stress identification strategy to perform stress identification analysis on the real-time ultrasonic signal to obtain the real-time stress of the weld area; Determining whether the real-time stress is within a stress threshold; If not, issuing a preliminary adjustment instruction, and adjusting the laser beam of the dual-cavity laser welding based on the preliminary adjustment instruction to obtain a preliminary adjusted stress; Determining whether the initially adjusted stress is within the stress threshold; If not, issuing a trajectory optimization instruction, and adjusting the predetermined welding trajectory of the dual-cavity laser welding based on the trajectory optimization instruction to obtain an optimized welding trajectory; The laser beam of the dual-cavity laser welding is adjusted based on the preliminary adjustment instruction, which includes: Obtain real-time laser feature information; The real-time laser characteristic information is screened in combination with a predetermined laser index to obtain a real-time laser parameter set; Using the real-time laser parameter set as input information of a laser welding heat prediction model to obtain real-time laser welding heat; Randomly generate a first laser beam adjustment scheme, and obtain a first predicted laser welding heat through analysis of the laser welding heat prediction model; When a first heat difference between the first predicted laser welding heat and the real-time laser welding heat meets a heat unit adjustment threshold, adjusting the laser beam according to the first laser beam adjustment scheme; The method further comprises: Extracting a first historical welding record from the historical dual-cavity mode laser welding records, wherein the first historical welding record includes a first historical laser parameter set and a first historical molten pool parameter set; Performing normalization analysis on first molten pool data obtained by traversing the first historical molten pool parameter set based on a predetermined molten pool index to obtain a first historical laser welding heat; Machine learning is performed on a first training data group formed based on the first historical laser parameter set and the first historical laser welding heat to obtain the laser welding heat prediction model.

2. The welding trajectory optimization method in the dual-cavity laser welding process according to claim 1 is characterized in that: include: Collecting multi-dimensional features of a predetermined workpiece for dual-cavity laser welding to obtain workpiece material feature information; According to the predetermined stress identification strategy, the workpiece material feature information is traversed and matched in the workpiece database to obtain the best matching workpiece data; Comparing the real-time ultrasonic signal with the best-matching ultrasonic signal in the best-matching workpiece data to obtain a real-time signal comparison result; The predetermined stress feedback coefficient is retrieved to analyze the real-time signal comparison result to obtain the real-time stress.

3. The welding trajectory optimization method in the dual-cavity laser welding process according to claim 2 is characterized in that: include: Extracting any artifact data of any artifact in the artifact database; Performing similarity analysis on any workpiece material feature information in the any workpiece data and the workpiece material feature information to obtain any similarity; When the arbitrary similarity satisfies the similarity threshold, any ultrasonic signal in the arbitrary workpiece data is used as the best matching ultrasonic signal.

4. The welding trajectory optimization method in the dual-cavity laser welding process according to claim 1 is characterized in that: The predetermined molten pool indicators include at least molten pool size and molten pool stability.

5. The welding trajectory optimization method in the dual-cavity laser welding process according to claim 1 is characterized in that: include: Reading welding trajectory characteristics, and generating trajectory optimization space in combination with the predetermined welding trajectory; Extracting a first optimization scheme in the trajectory optimization space, and simulating the first optimization scheme to obtain a first optimized stress; When the first optimized stress is within the stress threshold, the optimized welding trajectory is generated according to the first optimization scheme.

6. The welding trajectory optimization method in the dual-cavity laser welding process according to claim 5 is characterized in that: include: The welding trajectory characteristics include trajectory speed and trajectory angle; Taking a predetermined trajectory speed in the predetermined welding trajectory as a speed optimization upper limit, obtaining a speed optimization interval of the trajectory speed; Taking a predetermined trajectory angle in the predetermined welding trajectory as an angle optimization restriction, obtaining an angle optimization interval of the trajectory angle; The trajectory optimization space is constructed according to the speed optimization interval and the angle optimization interval.

Images