Narrow-gap oscillating laser welding pore prediction and regulation method and device based on cfd-real welding mapping and electronic equipment

Abstract

The present application relates to a kind of narrow gap swing laser welding technical field, it is a kind of narrow gap swing laser welding pore prediction and regulation method, device and electronic equipment based on CFD-real welding mapping, comprising: setting process parameter constraint condition, with pore rate mapping model output value minimum as objective function, execute process parameter reverse solution, determine optimal process parameter and pore rate prediction value;Optimal process parameter is converted into corresponding pore inhibition regulation strategy in combination with process parameter conversion rule, obtain the pore inhibition regulation strategy is applied to actual welding.This application establishes the quantitative mapping relationship between narrow gap swing laser welding CFD simulation and real welding porosity, designs the reverse optimization method of laser welding process parameter based on pore rate mapping model, controls pore inhibition regulation strategy, solves the existing simulation and real welding serious disjunction, pore defect is difficult to quantitatively predict, process regulation relies on experience trial and error, lacks closed-loop verification mechanism.

Description

Technical Field

[0001] This invention relates to the field of narrow-gap oscillating laser welding technology, specifically a method, apparatus, and electronic device for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping. Background Technology

[0002] Narrow-gap laser oscillation welding has become a core technology for the efficient connection of thick plate components in high-end equipment due to its concentrated heat input, small welding deformation, and excellent joint performance. It is widely used, especially in the root pass welding process in pipelines, pressure vessels, ships, and energy equipment. In droplet-free self-fusion welding mode, laser oscillation can reduce the incidence of porosity defects to some extent by forcibly disturbing the molten pool, improving sidewall wetting, and promoting bubble escape. However, due to the strong geometric constraints of the narrow gap, the unsteady heat input of the oscillation heat source, the complex flow of the molten pool, and the intense dynamic fluctuations of the small hole, porosity defects are still difficult to control stably, becoming a key issue restricting the long-term reliability of welded joints.

[0003] In recent years, methods for defect prediction and closed-loop control have been gradually established in conventional laser welding, for example: Chinese patent document CN121254753A discloses an adaptive closed-loop control method for laser welding process based on a defect prediction model. The method includes: synchronously collecting multimodal time-series data of the welding process through multiple physical sensors; inputting the data into a pre-trained defect prediction model to predict defect risks in advance and generate early warning signals; responding to the early warning, activating a digital twin model to perform online optimization simulation of process parameters; and finally converting the optimized parameters into control commands to drive the actuator to adjust the welding process in real time, forming a closed-loop control of "perception-prediction-decision-execution".

[0004] Chinese patent document CN117483956A discloses a method for optimizing process parameters of oscillating scanning laser-arc hybrid welding. The method includes: acquiring a set of process parameters for oscillating scanning laser-arc hybrid welding; conducting oscillating scanning laser-arc hybrid welding experiments on the set of process parameters using an orthogonal experimental control variable method to obtain welding test results; performing post-processing on the welding test results to output a training dataset; constructing a backpropagation BP neural network structure; optimizing the BP neural network structure using a differential evolution (DE) algorithm; training the optimized BP neural network structure using the training dataset to obtain a laser welding quality prediction model; and using the porosity and welding defects of the welded sample as optimization objectives, performing reinforcement learning on the output results of the laser welding quality prediction model using a dual-delay depth deterministic gradient (TD3) algorithm to output the optimal process parameters.

[0005] The above methods are only applicable to conventional laser welding and are not suitable for solving problems in the field of narrow-gap oscillating laser welding, such as a serious disconnect between simulation and actual welding, difficulty in quantitatively predicting porosity defects, reliance on experience-based trial and error for process control, and lack of closed-loop verification mechanisms.

[0006] In recent years, computational fluid dynamics (CFD) has gradually become an important tool for studying the behavior and defect mechanisms of laser welding molten pools. Existing CFD-based welding simulation and defect prediction technologies mostly focus on analyzing the temperature and flow fields of the molten pool, using numerical simulation to predict defects such as poor fusion and incomplete penetration. However, there are still significant limitations in truly applying CFD simulations to guide actual production and achieve precise porosity suppression.

[0007] Currently, most CFD-based simulation studies of narrow-gap laser welding remain at the theoretical analysis level, focusing only on calculations of physical quantities such as molten pool flow, temperature distribution, and sidewall wettability. They fail to establish an effective correlation with the actual welding process and real defect characteristics. Simulation results cannot accurately reflect the porosity, pore size, and spatial distribution patterns of actual welds, making it difficult to apply simulation conclusions to guide on-site process debugging and resulting in insufficient engineering practicality. Furthermore, existing technologies generally lack a quantitative mapping mechanism between simulation output characteristics and actual porosity defects. Simulation parameters such as orifice stability, molten pool velocity gradient, gas-liquid interface fluctuations, and temperature gradients cannot be converted into quantifiable porosity risk indicators, making it difficult for process engineers to determine the probability of actual porosity occurrence and conduct precise parameter optimization based on simulation results. In addition, existing defect prediction methods are mostly one-way passive modes of "simulation-defect judgment," lacking the ability to base predictions on simulation results. While the ability to automatically generate complete set of process parameters for porosity suppression is lacking, key parameters such as laser oscillation frequency, oscillation amplitude, laser power, welding speed, and defocusing amount still rely on experience-based debugging and extensive trial-and-error experiments. This results in problems such as long cycles, high costs, and poor consistency, failing to meet the needs of stable industrial production. Moreover, most CFD models use fixed heat source coefficients, boundary conditions, and material parameters, lacking a closed-loop verification system of "simulation-experiment-detection-correction." They cannot iteratively optimize based on actual welding results, and the simulation accuracy rapidly declines when materials, bevels, shielding gas, and equipment conditions change, limiting their applicability. Most critically, the closest existing technology only focuses on the wettability of the weld pool and fusion defects, without addressing the mechanisms of porosity nucleation, growth, and escape, failing to achieve quantitative prediction of porosity, and not disclosing the simulation-actual welding mapping, control strategy generation, and closed-loop verification. Summary of the Invention

[0008] This invention provides a method, device, and electronic device for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping. It overcomes the shortcomings of the prior art and can effectively solve the problems of serious disconnect between simulation and actual welding, difficulty in quantitatively predicting porosity defects, and reliance on experience-based trial and error in process control in the current field of narrow-gap oscillating laser welding.

[0009] One of the technical solutions of this invention is achieved through the following measures: a method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping, comprising: For each base material, a porosity mapping model was constructed using multiple samples. Each sample included porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features included the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The process parameters included laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. Select the base material, set process parameter constraints, take the minimum output value of the porosity mapping model corresponding to the base material as the objective function, perform reverse process parameter solution, and determine the optimal process parameters and the corresponding porosity prediction value. By combining the process parameter conversion rules, the optimal process parameters are converted into the corresponding porosity suppression and control strategy. The measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding is obtained. Based on the prediction error between the measured porosity value and the predicted porosity value, the relevant parameters of the narrow gap oscillating laser welding CFD simulation are reversed, and the porosity mapping model of the base material is updated.

[0010] The following are further optimizations and / or improvements to the above-mentioned technical solution: The above-mentioned process parameter constraints are set, with the objective function being to minimize the output value of the porosity mapping model. Inverse process parameter solving is then performed to determine the optimal process parameters and the corresponding predicted porosity values, including: The objective function and constraints are defined based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. An optimization algorithm is selected, and within the constraint space formed by the constraints, the optimization algorithm is used to perform multi-parameter collaborative optimization. The process parameters are used as decision variables, and the objective function is used as the basis for fitness calculation. The optimal solution that satisfies the constraints is taken as the optimal process parameters, and the corresponding porosity prediction value is output. The optimal process parameters are verified for process rationality. If the process rationality verification rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated.

[0011] The aforementioned optimization algorithms include grid search, genetic algorithm, particle swarm optimization, gradient descent, etc., and the appropriate algorithm should be selected based on the specific needs.

[0012] The aforementioned stomatal suppression and control strategy includes a basic process parameter table, control levels and contents, quality quantification results, and operating condition adaptation correction results, as well as process parameter conversion rules, including: The basic process parameters include: laser power, welding speed, oscillation frequency, oscillation amplitude, oscillation trajectory, defocusing amount, shielding gas type, shielding gas flow rate, and shielding gas pressure; The control level is determined based on the predicted porosity value and the stomatal classification and adjustment rules, which include: If the predicted porosity exceeds the set threshold but does not exceed n times the set threshold, first-level control is executed. The control includes controlling the oscillation frequency to increase to the set range or controlling the oscillation amplitude to increase to the set range. If the predicted porosity exceeds n times the set threshold, secondary control is executed. The control includes controlling the laser power to decrease to the set range, controlling the welding speed to decrease to the set range, and controlling the defocusing amount to be adjusted in the positive direction to the set range. Based on the predicted porosity value, the quality quantification result is determined according to the quality quantification standard. The quality quantification standard includes whether the porosity is greater than the allowable threshold, whether the maximum diameter of a single pore is greater than k, whether there is a dense pore area, the pore distribution pattern, whether there are depressions on the weld surface, whether there is spatter, and whether there is obvious undercut. Select a correction method to perform operating condition adaptation correction.

[0013] The porosity mapping model for each of the above-mentioned parent materials has the same construction structure. The process of determining a parent material and constructing a porosity mapping model using multiple samples includes: Select the base material, set multiple different combinations of process parameters, and perform CFD simulation and actual welding detection of narrow gap oscillating laser welding with the same parameters for each combination of process parameters to obtain the corresponding porosity tendency characteristics and measured values ​​of porosity, forming multiple samples. The porosity tendency characteristics include the fluctuation amplitude of the pinhole depth, the velocity gradient inside the molten pool, the fluctuation variance of the volume fraction of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. All samples are divided into training set, validation set and test set according to the proportions; A learning algorithm is selected, and the porosity mapping model is obtained by learning from the training set. The porosity mapping model is validated and tested using validation and test sets. The model parameters of the porosity mapping model are optimized, and a porosity mapping model that meets the test evaluation requirements is output.

[0014] The learning algorithms mentioned above include multiple linear regression, backpropagation neural network, support vector regression (SVR), and random forest algorithm.

[0015] The above-mentioned method obtains the measured porosity value after applying the porosity suppression and control strategy to actual welding. Based on the prediction error between the measured porosity value and the predicted porosity value, the relevant parameters of the narrow-gap oscillating laser welding CFD simulation are corrected in reverse, and the porosity mapping model is updated, including: Obtain the measured porosity value after applying the porosity suppression and control strategy to actual welding; The difference between the measured porosity and the predicted porosity is taken as the prediction error, and it is determined whether the absolute value of the prediction error is greater than the error threshold. If the error exceeds the error threshold, the prediction error will be fed back to the CFD simulation platform to correct the relevant parameters, including the heat source coefficient, boundary conditions, source term parameters and material thermal property parameters. Expand the sample size and update the porosity mapping model.

[0016] The second technical solution of the present invention is achieved through the following measures: a narrow-gap oscillating laser welding porosity prediction and control device based on CFD-actual welding mapping, comprising: The model building unit constructs a porosity mapping model using multiple samples for each base material. Each sample includes porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features include the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The process parameters include laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. The optimal process parameter solution unit selects the base material, sets process parameter constraints, takes the minimum output value of the porosity mapping model corresponding to the base material as the objective function, performs reverse process parameter solution, and determines the optimal process parameters and the corresponding porosity prediction value. The reverse optimization unit combines the process parameter conversion rules to convert the optimal process parameters into the corresponding porosity suppression and control strategy, obtains the measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding, and corrects the relevant parameters of the narrow gap oscillating laser welding CFD simulation based on the prediction error between the measured porosity value and the predicted porosity value, and updates the porosity mapping model of the base material.

[0017] The following are further optimizations and / or improvements to the above-mentioned technical solution: The above-mentioned optimal process parameter solution unit includes: The optimization basic condition setting module defines the objective function and constraints based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. The multi-parameter collaborative optimization module selects an optimization algorithm and performs multi-parameter collaborative optimization within the constraint space formed by the constraint conditions. It uses process parameters as decision variables and the objective function as the basis for fitness calculation to perform iterative search. The optimal solution that satisfies the constraint conditions is taken as the optimal process parameters, and the corresponding porosity prediction value is output. The process rationality verification module verifies the rationality of the optimal process parameters. If the process rationality verification rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated.

[0018] The aforementioned model building units include: The sample construction module selects the base material and sets multiple different combinations of process parameters. For each combination of process parameters, CFD simulation and actual welding detection of narrow gap oscillating laser welding with the same parameters are performed to obtain the corresponding porosity tendency characteristics and measured values ​​of porosity, forming multiple samples. The porosity tendency characteristics include the fluctuation amplitude of the pinhole depth, the internal velocity gradient of the molten pool, the fluctuation variance of the volume fraction of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The sample partitioning module divides all samples into training set, validation set and test set according to the proportions. The model learning module selects a learning algorithm and uses the training set to learn the porosity mapping model. The model testing module uses validation and test sets to validate and test the porosity mapping model, optimize the model parameters, and output a porosity mapping model that meets the test evaluation requirements.

[0019] The third technical solution of the present invention is achieved through the following measures: an electronic device, characterized in that it includes a processor and a memory, wherein the memory stores a computer program, and the computer program is loaded and executed by the processor to realize the steps in the prediction and control of porosity in narrow gap oscillating laser welding based on CFD-soldering mapping.

[0020] This invention addresses the problems in CFD simulation and actual welding of narrow-gap oscillating laser welding by establishing a closed-loop control method of "simulation-experiment-detection-control-optimization". This method solves the current problems in the field of narrow-gap oscillating laser welding, such as severe disconnect between simulation and actual welding, difficulty in quantitatively predicting porosity defects, reliance on empirical trial and error for process control, and lack of a closed-loop verification mechanism. Specifically: A reverse optimization method for laser welding process parameters based on a porosity mapping model was designed. An optimization chain was established consisting of a quantitative mapping model, an optimization algorithm, and process rationality verification. Specifically, for various base materials, a quantitative mapping relationship (i.e., a porosity mapping model) was established between the porosity of narrow-gap oscillating laser welding CFD simulations and actual welds. This transforms simulation from purely theoretical analysis into a reliable tool that accurately reflects actual defects and directly guides on-site production. An optimization algorithm was introduced to solve for the optimal process parameters and corresponding predicted porosity values, outputting a porosity suppression and control strategy. This strategy was further applied to verify and iteratively optimize the simulation process and the porosity mapping model. This approach combines practicality and efficiency, linking simulation with actual welding, accurately quantifying and predicting porosity defects, and determining the final process parameters that best reflect actual production. It is highly suitable for complex manufacturing processes like narrow-gap oscillating laser welding, which involve strong coupling of multiple parameters. Furthermore, the porosity mapping model is continuously corrected and learned from actual welding data, maintaining high accuracy over the long term, adapting to changes in materials, equipment, and operating conditions, and exhibiting strong industrial robustness. Attached Figure Description

[0021] Appendix Figure 1 This is a schematic diagram of the process for predicting and controlling porosity in narrow-gap oscillating laser welding, as disclosed in an embodiment of the present invention.

[0022] Appendix Figure 2 This is a schematic diagram of the optimal process parameter solution method disclosed in an embodiment of the present invention.

[0023] Appendix Figure 3 This is a schematic diagram of the reverse correction method for CFD simulation of narrow-gap oscillating laser welding disclosed in an embodiment of the present invention.

[0024] Appendix Figure 4 This is a schematic diagram of the porosity mapping model construction method disclosed in an embodiment of the present invention.

[0025] Appendix Figure 5 This is a schematic diagram of the narrow-gap oscillating laser welding porosity prediction and control device disclosed in an embodiment of the present invention. Detailed Implementation

[0026] The present invention is not limited to the following embodiments, and the specific implementation can be determined according to the technical solution of the present invention and the actual situation.

[0027] Those skilled in the art will understand that, unless specifically stated otherwise, in the embodiments of the present invention, a "module" or "unit" refers to a computer program or part of a computer program that has a predetermined function and works together with other related parts to achieve a predetermined goal, and can be implemented wholly or partially using software, hardware (such as processing circuitry or memory), or a combination thereof. Similarly, a processor (or multiple processors or memory) can be used to implement one or more modules or units. Furthermore, each module or unit can be part of an overall module or unit that includes the functionality of that module or unit.

[0028] In addition, in the embodiments of the present invention, "multiple" refers to two or more, and "first" and "second" are used to distinguish descriptions and should not be construed as implying relative importance.

[0029] This invention provides a method, apparatus, and electronic device for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping. This CFD-soldering mapping-based narrow-gap oscillating laser welding porosity prediction and control apparatus can be integrated into a computer device, which can be a server, a terminal, or other similar device; it can also be executed jointly by a terminal and a server. The above examples should not be construed as limiting the invention.

[0030] The aforementioned terminals may include mobile phones, wearable smart devices, tablet computers, laptops, personal computers (PCs), and in-vehicle computers, etc., and this invention does not limit the number of such terminals.

[0031] The aforementioned server can be an independent physical server, a server cluster or distributed system composed of multiple physical servers, or a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communication, middleware services, domain name services, security services, CDN, and big data and artificial intelligence platforms. This invention does not limit these features.

[0032] For example, for each base material, the computer equipment constructs a porosity mapping model using multiple samples. Each sample includes porosity tendency characteristics obtained from a narrow-gap oscillating laser welding CFD simulation based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency characteristics include the amplitude of pinhole depth fluctuation, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation at the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. Process parameters include laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. A base material is selected, and process parameter constraints are set. The objective function is to minimize the output value of the porosity mapping model corresponding to that base material. The process parameters are then solved in reverse to determine the optimal process parameters and the corresponding predicted porosity value. Combining process parameter conversion rules, the optimal process parameters are converted into a corresponding porosity suppression and control strategy. The measured porosity value obtained after applying this porosity suppression and control strategy to actual welding is obtained. Based on the prediction error between the measured porosity value and the predicted porosity value, the relevant parameters of the narrow-gap oscillating laser welding CFD simulation are corrected in reverse, and the porosity mapping model for that base material is updated.

[0033] The following is a definition of terms related to this invention: Computational fluid dynamics (CFD) is the process of solving the equations for conservation of mass, momentum, and energy. It is used to calculate the flow velocity, temperature field distribution, and solid-liquid phase transition (melting and solidification) of liquid metal in a molten pool.

[0034] CFD-Weld Mapping is an abbreviation that refers to the quantitative mapping relationship between the CFD simulation of narrow-gap oscillating laser welding and the porosity of actual welds (i.e., the porosity mapping model).

[0035] Based on this, the technical solution of the present invention will be described and explained below with reference to several examples.

[0036] Example 1: As shown in the attached document Figure 1 As shown, this invention discloses a method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping, comprising: Step S110: For each type of base material, a porosity mapping model is constructed using multiple samples. Each sample includes porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features include the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. Process parameters include laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. Step S120: Select the base material, set process parameter constraints, take the minimum output value of the porosity mapping model corresponding to the base material as the objective function, perform reverse process parameter solution, and determine the optimal process parameters and the corresponding predicted porosity value. Step S130: Combine the process parameter conversion rules to convert the optimal process parameters into the corresponding porosity suppression and control strategy, obtain the measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding, and correct the relevant parameters of the narrow gap oscillating laser welding CFD simulation based on the prediction error between the measured porosity value and the predicted porosity value, and update the porosity mapping model of the base material.

[0037] It should be noted that, in this embodiment, through step S110, a porosity mapping model is established for each type of base material, which reflects the quantitative mapping relationship between the CFD simulation and the actual weld porosity of narrow gap oscillating laser welding. This porosity mapping model can be continuously updated to enable accurate prediction.

[0038] This invention discloses a method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping. In the field of narrow-gap oscillating laser welding, a quantitative mapping relationship (i.e., porosity mapping model) is established between the CFD simulation and actual welding porosity of narrow-gap oscillating laser welding. A reverse optimization method for laser welding process parameters based on the porosity mapping model is designed. The optimal process parameters and corresponding predicted porosity values ​​are determined by reverse solving, and a porosity suppression and control strategy is output. This porosity control method has the ability to implement strategy verification and iterative optimization, forming a closed-loop control of "simulation-experiment-detection-control-optimization". This method not only achieves stable control of porosity in narrow-gap laser welding of thick plates, but also continuously corrects itself, learns autonomously, maintains high precision over a long period of time, adapts to changes in materials, equipment, and working conditions, and has strong industrial robustness.

[0039] Example 2: As shown in the attached document Figure 2 As shown, this embodiment of the invention is a further optimization of the above embodiment. It includes setting process parameter constraints, using the minimum output value of the porosity mapping model as the objective function, performing inverse process parameter solving, and determining the optimal process parameters and the corresponding predicted porosity value. Step S210: Define the objective function and constraints based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. The minimum and maximum values ​​for laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount were set to conform to the actual equipment and process window limitations.

[0040] The above M(X1,X2,X3,X4,X5)≤[φ] is a porosity threshold constraint, which is used to pre-limit the porosity within an acceptable range, avoiding the pursuit of extremely low porosity that exceeds the equipment's capabilities or causes other defects.

[0041] Step S220: Select an optimization algorithm. Within the constraint space formed by the constraints, use the optimization algorithm to perform multi-parameter collaborative optimization. Use process parameters as decision variables and the objective function as the fitness calculation basis for iterative search. Use the optimal solution that satisfies the constraints as the optimal process parameters and output the corresponding porosity prediction value. The aforementioned optimization algorithms include grid search, genetic algorithm, particle swarm optimization, gradient descent, etc., and the appropriate algorithm should be selected based on the specific needs. This embodiment uses the genetic algorithm as an example, and its optimization process includes: Initialization: Randomly generate N combinations of process parameters (individuals), each individual including laser power P, oscillation frequency f, oscillation amplitude A, welding speed v, and defocusing amount D; Fitness assessment: For each individual, determine the corresponding X1, X2, X3, X4, X5, and call the objective function (stomatal mapping model) to obtain the stomatal rate prediction value; calculate the fitness of the objective function (if a penalty function is included, add a penalty term); Selection: Select the best individuals based on fitness (the lower the better), such as in tournament selection or roulette. Crossover: Exchange some variables of the selected parent according to probability (simulated binary crossover for continuous variables, and uniform crossover for discrete trajectories). Variation: Perturb continuous variables with low probability (by adding Gaussian noise) or randomly switch trajectory types; Generate a new population: Repeat this process multiple times until the population size recovers to NN; Termination criteria: The predicted stomatal percentage of the optimal individual is ≤ [φ] when the maximum number of generations has been reached or the optimal fitness has not improved for multiple consecutive generations. Output: Laser power P, oscillation frequency f, oscillation amplitude A, welding speed v, defocusing amount D, and predicted porosity for the globally optimal individual.

[0042] Step S230: Perform a process rationality check on the optimal process parameters. If the process rationality check rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated. The process rationality check rules include: (1) Calculate the heat input based on the optimal process parameters of laser power and welding speed. The heat input is within a reasonable range of the base material. The formula for calculating heat input is as follows: q=η·P / v Where q is the linear energy, i.e., the heat input; P is the laser power; v is the welding speed; and η is the laser thermal efficiency coefficient. (2) The swing parameter of the optimal process parameter is less than the set threshold.

[0043] This embodiment designs a reverse optimization method for laser welding process parameters based on a porosity mapping model. It establishes an optimization chain of quantitative mapping model + optimization algorithm + process rationality verification, which is both practical and efficient. It also links simulation with actual welding, accurately quantifies and predicts porosity defects and the final process parameters that are closest to actual production. It is very suitable for complex manufacturing processes with strong coupling of multiple parameters, such as narrow gap oscillating laser welding.

[0044] Example 3: This embodiment of the invention is a further optimization of the above embodiments. The stomatal suppression control strategy includes a basic process parameter table, control levels and contents, quality quantification results, and operating condition adaptation correction results. The process parameter conversion rules include: The basic process parameters include: laser power (W), welding speed (mm / s), oscillation frequency (Hz), oscillation amplitude (mm), oscillation trajectory, defocusing amount (mm), shielding gas type, shielding gas flow rate (L / min), and shielding gas pressure (MPa).

[0045] The control level is determined based on the predicted porosity value and the stomatal classification and adjustment rules, which include: If the predicted porosity exceeds the set threshold but does not exceed n times (which can be 1.5 times) the set threshold, first-level control is executed. The control includes controlling the oscillation frequency to increase to the set range (the oscillation frequency can be increased by 5% to 15%) or controlling the oscillation amplitude to increase to the set range (the oscillation amplitude can be increased by 5% to 10%). If the predicted porosity exceeds n times (which can be 1.5 times) the set threshold, secondary control is executed. The control includes controlling the laser power to decrease to the set range (the laser power can be decreased by 5% to 10%), controlling the welding speed to decrease to the set range (the welding speed can be decreased by 5% to 15%), and controlling the defocusing amount to be adjusted in the positive direction to the set range (the defocusing amount can be adjusted in the positive direction by 0.5 mm to 1.0 mm).

[0046] The quality quantification result is determined based on the predicted porosity value and the quality quantification standard. The quality quantification standard includes whether the porosity is greater than the allowable threshold, whether the maximum diameter of a single pore is greater than k (which can be 0.5 mm), whether there is a dense pore area, the pore distribution pattern, whether there are depressions on the weld surface, whether there is spatter, and whether there is obvious undercut.

[0047] Select a correction method to perform working condition adaptation correction. For example, use linear ratio for working condition adaptation correction. When the plate thickness increases, the laser power and welding speed are increased synchronously according to the heat input formula. When the groove gap increases, the swing amplitude is increased to the set range (the swing amplitude can be increased by 0.2 mm to 0.5 mm), and the defocus amount is positively adjusted to the set value. When the material is changed, the thermophysical parameters are re-imported and the heat source absorption coefficient is corrected.

[0048] Example 4: As shown in the appendix Figure 3 As shown, this embodiment of the invention is a further optimization of the above embodiment. It involves obtaining the measured porosity value obtained after applying the porosity suppression and control strategy to actual welding, and using the prediction error between the measured porosity value and the predicted porosity value to inversely correct the relevant parameters of the narrow-gap oscillating laser welding CFD simulation, thereby updating the porosity mapping model. This includes: Step S310: Obtain the measured value of porosity obtained after applying the porosity suppression and control strategy to the actual welding. After applying the porosity suppression and control strategy to actual welding, industrial CT was performed again to obtain the measured value of porosity.

[0049] Step S320: The difference between the measured porosity and the predicted porosity is taken as the prediction error, and it is determined whether the absolute value of the prediction error is greater than the error threshold. The above error threshold can be set as needed; in this embodiment, it can be set to 10%.

[0050] Step S330: If the error exceeds the error threshold, the prediction error is fed back to the CFD simulation platform to correct the relevant parameters, including the heat source coefficient, boundary conditions, source term parameters, and material thermal property parameters.

[0051] Step S340: Expand the sample size and update the porosity mapping model.

[0052] This embodiment designs a welding verification and model closed-loop iterative optimization mechanism, which can continuously improve the accuracy of porosity prediction and the stability of process parameter optimization.

[0053] Example 5: As shown in the attached document Figure 4 As shown, the embodiments of the present invention are further optimizations of the above embodiments, wherein the porosity mapping model construction structure is the same for each type of parent material. The process of determining a parent material and constructing a porosity mapping model using multiple samples includes: Step S410: Select the base material, set multiple different combinations of process parameters, and perform CFD simulation and actual welding inspection of narrow-gap oscillating laser welding with the same parameters for each combination of process parameters to obtain the corresponding porosity characteristics and measured values ​​of porosity, forming multiple samples. Specifically: (1) For each combination of process parameters, a narrow-gap oscillating laser welding CFD simulation with the same parameters was performed to obtain the corresponding porosity characteristics, including: A three-dimensional geometric model for narrow-gap oscillating laser welding was established based on the Fluent platform. The groove size and material thermal properties (such as density, specific heat, thermal conductivity, viscosity, surface tension coefficient, etc.) were imported to match the actual base material. The process parameters were set to be exactly the same as those for actual welding, including laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. A strong coupling transient model of multiphase flow-heat transfer is established by using UDF, which includes vapor recoil pressure source term, buoyancy source term, Marangoni surface tension source term, and gas-liquid two-phase VOF interface tracking. Based on the principle of momentum-energy conservation, a three-dimensional transient control equation set is established, including momentum equation, energy transport equation and fluid volume function equation, to describe the flow of molten pool, heat and mass transfer and gas-liquid interface evolution process. Based on the characteristics of actual welding conditions, a three-dimensional geometric model was constructed using the ICEM module, generating a structured computational network dominated by hexahedral elements. After naming the key walls of the three-dimensional geometric model, the mesh data was imported into the Fluent platform, and corresponding convection-radiation composite boundary conditions were set. Select a pressure-corrected basis solver, define the material properties of the two-phase flow, load the moving laser heat source model, set the momentum and energy source terms, and activate the VOF phase transition module; The three-dimensional transient control equations are spatially discretized using the finite volume method, and the pressure-velocity coupled solution is achieved using the PISO algorithm. The time discretization is performed using a second-order implicit scheme. The double-precision solver is enabled for iterative calculations, and the fluctuations of the molten pool interface and the temperature gradient changes are monitored in real time. The flow field vector, temperature cloud map and keyhole morphology data are output every 0.0005s. An adaptive time step is set, with a range of 10⁻⁶ to 10⁻⁵ s. CFD-POST was used to perform post-processing of the computational data, extract the temperature field and flow field characteristics of the molten pool, quantitatively analyze the influence of process parameters on the oscillation behavior of the molten pool, and extract the porosity tendency characteristic value.

[0054] (2) For each combination of process parameters, actual welding was carried out using the same base material, plate thickness, bevel type, and process parameters as the simulation, to obtain the corresponding porosity characteristics, including: Actual welding was carried out using base material, plate thickness, bevel type, and process parameters that were completely consistent with the simulation. Standard weld samples were prepared, and the standard weld samples were tested by industrial CT to obtain the measured value of porosity.

[0055] Step S420: Divide all samples into training set, validation set and test set according to the proportions; Step S430: Select a learning algorithm, use the training set to learn, and obtain a porosity mapping model; The learning algorithms mentioned above may include multiple linear regression, backpropagation neural network, support vector regression (SVR), random forest algorithm, etc.

[0056] Step S440: Validate and test the porosity mapping model using the validation set and test set, optimize the model parameters of the porosity mapping model, and output a porosity mapping model that meets the test evaluation requirements.

[0057] The above-mentioned test evaluation indicators may include, but are not limited to, the coefficient of determination R. 2 Mean square error (MSE).

[0058] Example 6: As shown in the appendix Figure 5 As shown, this invention discloses a narrow-gap oscillating laser welding porosity prediction and control device based on CFD-soldering mapping, comprising: The model building unit constructs a porosity mapping model using multiple samples for each base material. Each sample includes porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features include the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The process parameters include laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. The optimal process parameter solution unit selects the base material, sets process parameter constraints, takes the minimum output value of the porosity mapping model corresponding to the base material as the objective function, performs reverse process parameter solution, and determines the optimal process parameters and the corresponding porosity prediction value. The reverse optimization unit combines the process parameter conversion rules to convert the optimal process parameters into the corresponding porosity suppression and control strategy, obtains the measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding, and corrects the relevant parameters of the narrow gap oscillating laser welding CFD simulation based on the prediction error between the measured porosity value and the predicted porosity value, and updates the porosity mapping model of the base material.

[0059] The optimal process parameter solution unit includes: The optimization basic condition setting module defines the objective function and constraints based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. The multi-parameter collaborative optimization module selects an optimization algorithm and performs multi-parameter collaborative optimization within the constraint space formed by the constraint conditions. It uses process parameters as decision variables and the objective function as the basis for fitness calculation to perform iterative search. The optimal solution that satisfies the constraint conditions is taken as the optimal process parameters, and the corresponding porosity prediction value is output. The process rationality verification module verifies the rationality of the optimal process parameters. If the process rationality verification rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated.

[0060] The model building unit includes: The sample construction module selects the base material and sets multiple different combinations of process parameters. For each combination of process parameters, CFD simulation and actual welding detection of narrow gap oscillating laser welding with the same parameters are performed to obtain the corresponding porosity tendency characteristics and measured values ​​of porosity, forming multiple samples. The porosity tendency characteristics include the fluctuation amplitude of the pinhole depth, the internal velocity gradient of the molten pool, the fluctuation variance of the volume fraction of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The sample partitioning module divides all samples into training set, validation set and test set according to the proportions. The model learning module selects a learning algorithm and uses the training set to learn the porosity mapping model. The model testing module uses validation and test sets to validate and test the porosity mapping model, optimize the model parameters, and output a porosity mapping model that meets the test evaluation requirements.

[0061] Example 7: This embodiment of the invention discloses an electronic device, including a processor and a memory. The memory stores a computer program, which is loaded and executed by the processor to implement the steps in the method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping.

[0062] The processor described above can be a central processing unit (CPU), a general-purpose processor, a digital signal processor (DSP), an ASIC, an FPGA, or other programmable logic devices, transistor logic devices, hardware components, or any combination thereof. It can implement or execute the various exemplary logic blocks, modules, and circuits described in conjunction with the disclosure of this invention. It can also be a combination that implements computational functions, such as a combination of one or more microprocessors, a combination of a DSP and a microprocessor, etc. The memory can include, but is not limited to, various media capable of storing computer programs, such as USB flash drives, read-only memory, portable hard drives, magnetic disks, or optical disks.

[0063] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code. The solutions in the embodiments of the present invention can be implemented using various computer languages, such as the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0064] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0065] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0066] The above content is only a specific embodiment of the present invention, which has strong adaptability and implementation effect. However, the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily conceived by those skilled in the art within the technical scope disclosed in the present invention should be covered within the protection scope of the present invention. Therefore, equivalent changes made in accordance with the claims of the present invention are still within the scope of the present invention.

Claims

1. A method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping, characterized in that, include: For each base material, a porosity mapping model was constructed using multiple samples. Each sample included porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features included the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The process parameters included laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. Select the base material, set process parameter constraints, take the minimum output value of the porosity mapping model corresponding to the base material as the objective function, perform reverse process parameter solution, and determine the optimal process parameters and the corresponding porosity prediction value. By combining the process parameter conversion rules, the optimal process parameters are converted into the corresponding porosity suppression and control strategy. The measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding is obtained. Based on the prediction error between the measured porosity value and the predicted porosity value, the relevant parameters of the narrow gap oscillating laser welding CFD simulation are reversed, and the porosity mapping model of the base material is updated.

2. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping according to claim 1, characterized in that, Set process parameter constraints, take minimizing the output value of the porosity mapping model as the objective function, perform inverse process parameter solving, and determine the optimal process parameters and the corresponding predicted porosity values, including: The objective function and constraints are defined based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. An optimization algorithm is selected, and within the constraint space formed by the constraints, the optimization algorithm is used to perform multi-parameter collaborative optimization. The process parameters are used as decision variables, and the objective function is used as the basis for fitness calculation. The optimal solution that satisfies the constraints is taken as the optimal process parameters, and the corresponding porosity prediction value is output. The optimal process parameters are verified for process rationality. If the process rationality verification rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated.

3. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping according to claim 2, characterized in that, Optimization algorithms include grid search, genetic algorithm, particle swarm optimization, and gradient descent.

4. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping according to claim 1, 2, or 3, characterized in that, Stomatal suppression control strategies include a basic process parameter table, control levels and contents, quality quantification results, and operating condition adaptation correction results, as well as process parameter conversion rules, including: The basic process parameters include: laser power, welding speed, oscillation frequency, oscillation amplitude, oscillation trajectory, defocusing amount, shielding gas type, shielding gas flow rate, and shielding gas pressure; The control level is determined based on the predicted porosity value and the stomatal classification and adjustment rules, which include: If the predicted porosity exceeds the set threshold but does not exceed n times the set threshold, first-level control is executed. The control includes controlling the oscillation frequency to increase to the set range or controlling the oscillation amplitude to increase to the set range. If the predicted porosity exceeds n times the set threshold, secondary control is executed. The control includes controlling the laser power to decrease to the set range, controlling the welding speed to decrease to the set range, and controlling the defocusing amount to be adjusted in the positive direction to the set range. Based on the predicted porosity value, the quality quantification result is determined according to the quality quantification standard. The quality quantification standard includes whether the porosity is greater than the allowable threshold, whether the maximum diameter of a single pore is greater than k, whether there is a dense pore area, the pore distribution pattern, whether there are depressions on the weld surface, whether there is spatter, and whether there is obvious undercut. Select a correction method to perform operating condition adaptation correction.

5. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping according to claim 1, 2, or 3, characterized in that, The porosity mapping model is constructed using the same structure for each type of parent material. The process of determining a parent material and constructing a porosity mapping model using multiple samples includes: Select the base material, set multiple different combinations of process parameters, and perform CFD simulation and actual welding detection of narrow gap oscillating laser welding with the same parameters for each combination of process parameters to obtain the corresponding porosity tendency characteristics and measured values ​​of porosity, forming multiple samples. The porosity tendency characteristics include the fluctuation amplitude of the pinhole depth, the flow velocity gradient inside the molten pool, the fluctuation variance of the volume fraction of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. All samples are divided into training set, validation set and test set according to the proportions; A learning algorithm is selected, and the porosity mapping model is obtained by learning from the training set. The porosity mapping model is validated and tested using validation and test sets. The model parameters of the porosity mapping model are optimized, and a porosity mapping model that meets the test evaluation requirements is output.

6. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-soldering mapping according to claim 5, characterized in that, The learning algorithms include multiple linear regression, backpropagation neural network, support vector regression (SVR), and random forest algorithm.

7. The method for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping according to claim 1, 2, or 3, characterized in that, Obtain the measured porosity values ​​obtained after applying the porosity suppression and control strategy to actual welding. Based on the prediction error between the measured and predicted porosity values, reverse-correct the relevant parameters of the narrow-gap oscillating laser welding CFD simulation, and update the porosity mapping model, including: Obtain the measured porosity value after applying the porosity suppression and control strategy to actual welding; The difference between the measured porosity and the predicted porosity is taken as the prediction error, and it is determined whether the absolute value of the prediction error is greater than the error threshold. If the error exceeds the error threshold, the prediction error will be fed back to the CFD simulation platform to correct the relevant parameters, including the heat source coefficient, boundary conditions, source term parameters and material thermal property parameters. Expand the sample size and update the porosity mapping model.

8. A device for predicting and controlling porosity in narrow-gap oscillating laser welding based on CFD-actual welding mapping, using the method described in any one of claims 1 to 7, characterized in that, include: The model building unit constructs a porosity mapping model using multiple samples for each base material. Each sample includes porosity tendency features obtained from CFD simulation of narrow-gap oscillating laser welding based on a combination of process parameters, as well as the measured porosity value corresponding to that combination of process parameters. Porosity tendency features include the fluctuation amplitude of pinhole depth, the velocity gradient inside the molten pool, the variance of the volume fraction fluctuation of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The process parameters include laser power, oscillation frequency, oscillation amplitude, welding speed, and defocusing amount. The optimal process parameter solution unit selects the base material, sets process parameter constraints, takes the minimum output value of the porosity mapping model corresponding to the base material as the objective function, performs reverse process parameter solution, and determines the optimal process parameters and the corresponding porosity prediction value. The reverse optimization unit combines the process parameter conversion rules to convert the optimal process parameters into the corresponding porosity suppression and control strategy, obtains the measured porosity value obtained after applying the porosity suppression and control strategy to the actual welding, and corrects the relevant parameters of the narrow gap oscillating laser welding CFD simulation based on the prediction error between the measured porosity value and the predicted porosity value, and updates the porosity mapping model of the base material.

9. The narrow-gap oscillating laser welding porosity prediction and control device based on CFD-real welding mapping according to claim 8, characterized in that, The optimal process parameter solution unit includes: The optimization basic condition setting module defines the objective function and constraints based on the porosity mapping model, as shown below: Objective function: Min M(X1, X2, X3, X4, X5); Constraints include: P_min≤P≤P_max f_min≤f≤f_max A_min≤A≤A_max v_min≤v≤v_max D_min≤D≤D_max M(X1,X2,X3,X4,X5)≤[φ] Where M is the porosity mapping model; X1 is the aperture depth fluctuation amplitude, X2 is the flow velocity gradient inside the molten pool, X3 is the volume fraction fluctuation variance of the gas-liquid interface, X4 is the longitudinal temperature gradient of the molten pool, and X5 is the transverse temperature gradient of the molten pool; P is the laser power, and P_min and P_max are the minimum and maximum values ​​of the laser power, respectively; f is the oscillation frequency, and f_min and f_max are the minimum and maximum values ​​of the oscillation frequency, respectively; A is the oscillation amplitude, and A_min and A_max are the minimum and maximum values ​​of the oscillation amplitude, respectively; v is the welding speed, and v_min and v_max are the minimum and maximum values ​​of the welding speed, respectively; D is the defocusing amount, and D_min and D_max are the minimum and maximum values ​​of the defocusing amount, respectively; [φ] is the production-allowed porosity threshold. The multi-parameter collaborative optimization module selects an optimization algorithm and performs multi-parameter collaborative optimization within the constraint space formed by the constraint conditions. It uses process parameters as decision variables and the objective function as the basis for fitness calculation to perform iterative search. The optimal solution that satisfies the constraint conditions is taken as the optimal process parameters, and the corresponding porosity prediction value is output. The process rationality verification module verifies the rationality of the optimal process parameters. If the process rationality verification rules are met, the optimal process parameters are retained; otherwise, the optimal process parameters are recalculated. or / and, Model building units include: The sample construction module selects the base material and sets multiple different combinations of process parameters. For each combination of process parameters, CFD simulation and actual welding detection of narrow gap oscillating laser welding with the same parameters are performed to obtain the corresponding porosity tendency characteristics and measured values ​​of porosity, forming multiple samples. The porosity tendency characteristics include the fluctuation amplitude of the pinhole depth, the internal velocity gradient of the molten pool, the fluctuation variance of the volume fraction of the gas-liquid interface, the longitudinal temperature gradient of the molten pool, and the transverse temperature gradient of the molten pool. The sample partitioning module divides all samples into training set, validation set and test set according to the proportions. The model learning module selects a learning algorithm and uses the training set to learn the porosity mapping model. The model testing module uses validation and test sets to validate and test the porosity mapping model, optimize the model parameters, and output a porosity mapping model that meets the test evaluation requirements.

10. An electronic device, characterized in that, It includes a processor and a memory, wherein the memory stores a computer program, which is loaded and executed by the processor to implement the steps in the method as described in any one of claims 1 to 7.

Images