Welding procedure calling system and method with weld profile matching function

Abstract

The present invention belongs to the field of welding technology and discloses a welding procedure calling system and method with a weld profile matching function; the method comprises: collecting the material grade of a base material, and indexing it in a preset material database according to the material grade, obtaining the thermal expansion coefficient and thermal conductivity coefficient corresponding to the base material, classifying the welding material according to the thermal expansion coefficient and thermal conductivity coefficient, and obtaining the classification result; matching the base material with the welding material according to the classification result to obtain the material grade of the welding material; establishing a 3D finite element model based on the selected base material and welding material, collecting and inputting the thermal expansion coefficient and thermal conductivity coefficient of the material, setting the optimization variable range, and optimizing the 3D finite element model; the optimization variables include current, voltage and welding speed; outputting the heat affected zone width and residual stress based on the optimized 3D finite element model; the present invention can comprehensively improve welding quality, efficiency and automation level, and promote the standardization and normalization of welding processes.

Description

Technical Field

[0001] The present invention relates to the field of welding technology, and more particularly to a welding procedure calling system and method with a weld profile matching function. Background Art

[0002] In modern manufacturing, welding, as a key joining process, is widely used in a wide range of fields, including aerospace, automotive manufacturing, and shipbuilding. With the continuous advancement of industrial technology, the requirements for welding quality, efficiency, and automation are increasing, and traditional welding processes are facing many challenges.

[0003] Patent application publication number CN117300466A discloses a welding process control method, apparatus, computer equipment, and storage medium. The method involves obtaining welding parameters of a welding plate and a cross-sectional image of a weld groove; selecting a welding process that matches the welding parameters of the welding plate from a preset welding process library; calculating the area of a filled droplet based on the welding process and the cross-sectional image of the weld groove; and calculating the operating current during welding of the weld groove cross-sectional image based on the filled droplet area. This technical solution adjusts the initial welding current based on the calculated current offset. This current can be adjusted within a preset buffer length, ensuring that the welding plate operates in a welding process consistent with the welding parameters during welding. This allows for monitoring the overall condition of the weld plate weld, enabling controlled current regulation during the welding process and improving welding quality.

[0004] Although the above methods can meet most scenarios, research and practical application of the above methods and existing technologies have revealed that the above methods and existing technologies have at least the following defects:

[0005] Relying on preset paths or manual teaching cannot adapt to the dynamic deformation of complex geometric welds; parameter selection relies on experience or fixed rules and cannot dynamically respond to material fluctuations;

[0006] In view of this, the present invention proposes a welding procedure calling system and method with a weld profile matching function to solve the above problems. Summary of the Invention

[0007] In order to overcome the above-mentioned defects of the prior art and achieve the above-mentioned purpose, the present invention provides the following technical solution: a welding procedure specification calling method with a weld profile matching function, comprising the following steps:

[0008] Collect the material grade of the parent material and index it in the preset material database according to the material grade to obtain the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material. Classify the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient to obtain the classification results; match the welding materials with the parent material according to the classification results to obtain the material grade of the welding material;

[0009] Based on the selected base material and welding material, a 3D finite element model is established. The thermal expansion coefficient and thermal conductivity of the material are collected and input. The optimization variable range is set and the 3D finite element model is optimized. The optimization variables include current, voltage and welding speed. Based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output.

[0010] Obtain the Cr and Ni content in the molten pool, and establish a Cr / Ni element calibration equation by fitting. When the Cr content is detected to be lower than the preset content threshold, a feedback mechanism for adjusting the optimization variables is triggered.

[0011] Collect 3D point cloud data of the weld, extract weld contour features, match the weld contour features with the contours in the preset contour library to obtain the weld contour matching results; perform trajectory correction on the weld contour based on the weld contour matching results;

[0012] Construct a multi-level rule library based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching. Define the process procedures according to the multi-level rule library for dynamic calling of the process procedures in the later stage.

[0013] Furthermore, the method for optimizing the 3D finite element model includes:

[0014] Collect the shape and size of the welded structure and input them into the finite element pre-processing software to create a three-dimensional geometric model;

[0015] Obtain the thermophysical parameters of the base material, including density, specific heat capacity, thermal expansion coefficient and thermal conductivity;

[0016] Input the thermophysical property parameters into the corresponding material area and build the corresponding heat conduction equation;

[0017] Preset the value range of current, voltage and welding speed, and set current, voltage and welding speed as optimization variables;

[0018] The optimization variables are associated with the heat source model of the welding process, and the welding heat source is divided into two ellipsoid parts to build a double ellipsoid heat source model;

[0019] Preset heat-affected zone width constraint intervals, residual stress constraint intervals, and energy efficiency constraint intervals; use the NSGA-II algorithm to simultaneously optimize the heat-affected zone width, residual stress, and energy efficiency during welding, generate an optimal solution set, and select the parameter combination consisting of the optimization variables with the highest comprehensive score in the optimal solution set;

[0020] The total heat power is obtained by integrating the double ellipsoid heat source model , and then solve the heat conduction equation to obtain the temperature during welding;

[0021] The node displacement is obtained by establishing and solving a set of nonlinear equations; the residual stress is then calculated based on the geometric equations and constitutive equations.

[0022] Furthermore, the method of selecting the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set includes:

[0023] Step 1: define the objective function of the 3D finite element model, which includes a first objective function, a second objective function, and a third objective function;

[0024] Step 2: Randomly generate H individuals within the range of current, voltage, and welding speed, each of which represents a parameter combination obtained by splicing current, voltage, and welding speed;

[0025] Step 3: Calculate the three objective function values corresponding to each individual;

[0026] Step 4: For any two individuals p and q, if the three objective function values of individual p are less than or equal to the three objective function values corresponding to individual q, and at least one objective function value of individual p is less than the objective function value corresponding to individual q, then individual p is judged to dominate individual q;

[0027] Step 5: Repeat step 4 to obtain the first set of individuals consisting of all individuals that are not dominated by other individuals, and divide the first set of individuals into the first level;

[0028] Step 6: Repeat steps 4-5 for the remaining individuals that have not been classified into different levels, obtain new sets of individuals that are not dominated by the remaining individuals, and divide them into the next level until all individuals are divided into the corresponding level;

[0029] Step 7: Calculate the crowding degree of individuals in each level;

[0030] Step 8: Use the tournament selection method to randomly select K individuals from the population, and select the individual with the highest rank and a crowding degree higher than the preset crowding threshold as the parent;

[0031] Step 9: Select the parent individuals and Perform crossover operation to generate offspring individuals and ;

[0032] Step 10: performing mutation operations on the offspring individuals to change the parameter values of the offspring individuals;

[0033] Step 11: Merge the parent population of size N and the mutated offspring population of size N to obtain a merged population of size 2N.

[0034] Step 12: Perform non-dominated sorting on the merged population, select the top N individuals according to rank and crowding degree, and form a new population;

[0035] Step 13: When the preset update requirements are met, stop updating the new population; obtain the corresponding optimal solution set;

[0036] Step 14: Calculate the comprehensive score of each individual in the optimal solution set based on the comprehensive scoring function, and select the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set.

[0037] Furthermore, the weld profile features include geometric features and topological features; and the method for obtaining the geometric features includes:

[0038] For edge points in three-dimensional point cloud data, k neighboring points are found through neighborhood search and a covariance matrix is constructed. The covariance matrix is decomposed into eigenvalues. The eigenvector corresponding to the minimum eigenvalue is taken as the normal vector of the point. The curvature of the edge point is calculated based on the normal vector and the neighborhood points. The curvature of all edge points is counted to obtain a curvature set, which is used as a geometric feature.

[0039] Furthermore, the method for obtaining topological features includes:

[0040] Step A: Select dimension R and sort the 3D point cloud data according to the value of the selected dimension.

[0041] Step B: Take the point corresponding to the middle value of the sorting result as the root node, and divide the 3D point cloud data into two parts, the left part and the right part, according to the root node;

[0042] Step C: Repeat step B for the left and right parts respectively to build a subtree;

[0043] Step D: Repeat steps B to C until both the left and right parts contain only one node.

[0044] Step E: From the seed point Initially, region growing is performed based on the normal vector similarity and distance threshold between points;

[0045] Step F: Get the boundary points of the growth area , that is, the weld contour point , calculate weld contour points With neighboring points The normal vector angle ;like is greater than the set normal vector similarity threshold, and the point with dot The distance is less than the distance threshold; then the point Add the growth area and continue to do the above operation for the newly added points until there are no more points that meet the conditions to be added;

[0046] Step G: Count all the boundary points in the growing area to obtain a topological set, and use the topological set as a topological feature.

[0047] Furthermore, the method for obtaining the weld profile matching result includes:

[0048] Calculate the Euclidean distance between the geometric features of the weld contour and the geometric features of the contours in the contour library, and calculate the similarity between the topological features of the weld contour and the topological features of the contours in the contour library; compare and judge based on the preset distance matching threshold and similarity matching threshold:

[0049] If the Euclidean distance between the weld contour and the contour in the contour library is less than the distance matching threshold, and the similarity is higher than the similarity matching threshold, then it is determined that the weld contour is the same as the matched contour in the contour library;

[0050] If the following situations occur:

[0051] There is more than one matching result;

[0052] The Euclidean distance is greater than the distance matching threshold and the similarity is greater than the similarity matching threshold;

[0053] The Euclidean distance is greater than the distance matching threshold and the similarity is lower than the similarity matching threshold;

[0054] The Euclidean distance is less than the distance matching threshold and the similarity is less than the similarity matching threshold;

[0055] The contour in the contour library corresponding to the highest similarity is selected as the final matching result;

[0056] Methods for correcting the weld contour trajectory based on the weld contour matching results include:

[0057] The weld offset during the welding process is obtained. When the weld offset is greater than the preset offset threshold, the weld point trajectory is corrected.

[0058] Furthermore, the method of constructing a multi-level rule base includes:

[0059] Step a, obtaining the material grade of the base material, the material grade of the welding material, the thermal expansion coefficient and thermal conductivity of the material, and outputting the heat-affected zone width and residual stress, the content of Cr and Ni elements, the profile characteristics, and the welding offset based on the optimized 3D finite element model;

[0060] Step b: designing and storing rule logic for welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching during the welding process, wherein the rule logic includes basic rules, dynamic rules, and composite rules;

[0061] Step c: Use the rule engine to implement condition matching and action triggering.

[0062] Further, methods for defining process specifications include:

[0063] Real-time collection of base material grade, thermal expansion coefficient and thermal conductivity, Cr and Ni content, and three-dimensional point cloud data of welds;

[0064] Compare the collected data with the condition fields in the rule logic defined in the multi-level rule library to trigger. If multiple rules are triggered at the same time, they will be executed in the order of preset priority.

[0065] Convert the corresponding action field in the triggered rule logic into a device instruction, and send the device instruction to the corresponding device, so that the device executes the device instruction;

[0066] If the device instruction is a feedback mechanism for adjusting optimization variables and a trajectory correction rule, the Cr and Ni content is collected again to determine whether it has recovered to the preset content threshold. If not, the feedback mechanism is triggered until the Cr and Ni content recovers to the preset content threshold.

[0067] Rescan the adjusted weld profile and perform a second match with the preset profile library; if the match fails E times in a row, an alarm is triggered and the machine is shut down, prompting manual intervention; and the adjustment data, rule trigger records, and quality inspection results are stored in the preset call database.

[0068] Furthermore, the method for obtaining the classification result includes:

[0069] Preset G thermal expansion coefficient thresholds, divide according to the preset G thermal expansion coefficient thresholds, obtain G+1 thermal expansion coefficient intervals, and number each thermal expansion coefficient interval in ascending order; preset M thermal conductivity coefficient thresholds, divide according to the preset M thermal conductivity coefficient thresholds, obtain M+1 thermal conductivity coefficient intervals, and number each thermal conductivity coefficient interval in ascending order; traverse and compare the thermal expansion coefficient and thermal conductivity coefficient of the welding material with the G+1 thermal expansion coefficient intervals and the M+1 thermal conductivity coefficient intervals, respectively, to obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number to which the welding material belongs as the classification result.

[0070] Furthermore, the method for obtaining the material grade of the welding material includes:

[0071] Obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number corresponding to the base material, and select the material corresponding to the material brand whose thermal expansion coefficient interval number and thermal conductivity coefficient interval number are the same as or adjacent to the thermal expansion coefficient interval number and thermal conductivity coefficient interval number of the base material as the material brand of the welding material.

[0072] A welding procedure specification calling system with a weld profile matching function is used to implement a welding procedure specification calling method with a weld profile matching function, comprising:

[0073] Material matching module: collects the material grade of the parent material, and indexes it in the preset material database according to the material grade, obtains the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material, classifies the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient, and obtains the classification results; matches the parent material with the welding material according to the classification results to obtain the material grade of the welding material;

[0074] Model prediction module: Based on the selected base material and welding material, a 3D finite element model is established, the thermal expansion coefficient and thermal conductivity of the material are collected and input, the optimization variable range is set, and the 3D finite element model is optimized; the optimization variables include current, voltage, and welding speed; based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output;

[0075] Feedback adjustment module: obtains the Cr and Ni content in the molten pool, fits and establishes a Cr / Ni element calibration equation, and triggers a feedback mechanism to adjust the optimization variables when the Cr content is detected to be lower than the preset content threshold;

[0076] Contour matching module: collects 3D point cloud data of welds, extracts weld contour features, matches weld contour features with contours in a preset contour library, and obtains weld contour matching results; performs trajectory correction on weld contours based on the weld contour matching results;

[0077] Procedure calling module: Build a multi-level rule library based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld contour matching. Define process procedures based on the multi-level rule library for dynamic calling of process procedures in the later stage.

[0078] The technical effects and advantages of the welding procedure specification calling system and method with weld profile matching function of the present invention are as follows:

[0079] The present invention collects the brand of the parent material and classifies and matches the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient, thus overcoming the limitation of traditional parameter selection relying on experience or fixed rules, and can dynamically respond to the fluctuation of the parent material characteristics, ensuring the synergy of the performance of the parent material and the welding material during the welding process, thus guaranteeing the welding quality from the source and improving the reliability and stability of the welded joint; the invention uses the 3D finite element model to optimize the variables, provides a scientific basis for the welding process, accurately predicts the physical phenomena in the welding process, optimizes the process parameters in advance, reduces welding defects, and improves the welding quality and production efficiency; the invention can monitor the welding process in real time and feedback the adjustment based on the element content in the molten pool. The system can timely adjust welding parameters to avoid quality problems caused by fluctuations in material composition, ensure the consistency and stability of welding quality, and realize adaptive weld tracking by collecting three-dimensional point cloud data of welds. It breaks through the bottleneck of existing technologies that rely on preset paths or manual teaching and cannot adapt to the dynamic deformation of complex geometric welds. It can track the actual position and shape of welds in real time, automatically adjust the welding trajectory, realize high-precision welding of complex welds, improve the automation and intelligence level of welding, and finally build a multi-level rule library to realize intelligent process calling, comprehensively improve welding quality, efficiency and automation level, and promote the standardization and normalization of welding processes. BRIEF DESCRIPTION OF THE DRAWINGS

[0080] Figure 1 A schematic flow chart of a method for calling a welding procedure specification with a weld profile matching function according to the present invention;

[0081] Figure 2 A schematic flow chart of a method for constructing a multi-level rule base according to the present invention;

[0082] Figure 3 Schematic diagram of parameters of the double ellipsoid heat source model of the present invention;

[0083] Figure 4 It is the temperature field distribution diagram of the 3D finite element model of the present invention;

[0084] Figure 5 It is the temperature field distribution and the spatial distribution diagram of welding residual stress of the present invention;

[0085] Figure 6 Schematic diagram of temperature field distribution, spatial distribution of welding residual stress and deformation parameters of the present invention;

[0086] Figure 7 This is a schematic diagram of the structure of a welding procedure calling system with a weld profile matching function according to the present invention;

[0087] Figure 8 This is a schematic diagram of the welding procedure calling system interface with weld profile matching function of the present invention. DETAILED DESCRIPTION

[0088] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative efforts are within the scope of protection of the present invention.

[0089] Example 1

[0090] The implementation conditions for this example are as follows: a FANUC ARC Mate 120iD welding robot with a maximum payload of 20 kg, a repeatability of ±0.08 mm, a working range of 1443 mm, support for six-axis linkage and offline programming (OLP), and a Fronius TPS 400i CMT cold metal transfer welder. The system is equipped with an Ocean Optics LIBS 2500+ laser-induced breakdown spectroscopy (LIBS) system, featuring a 1064 nm laser wavelength, adjustable pulse energy from 50 to 100 mJ, and a spectral range of 200 to 980 nm. The system is equipped with a 2048-pixel high-resolution CCD detector, a sampling frequency of 1 to 10 Hz, and the SpectraSuite spectral analysis platform to support Cr / Ni element calibration curve fitting. A GOM ATOS Q 12M 3D point cloud scanner with a scanning accuracy of ±0.01mm, a resolution of 12 megapixels, and a full-frame scanning speed of 1.5 seconds per frame was used. Point cloud data preprocessing and weld contour matching were performed in conjunction with GOM Inspect Pro software. Finite element analysis was performed using ANSYS Mechanical APDL 2022 R1 software. A 3D model of the welded structure was created using DesignModeler. A hexahedral mesh was used, with local meshing to 0.5mm in the heat-affected zone. The thermophysical properties of the base material (12Cr1MoV) and the weld consumable (ER70S-6) were defined. The ANSYS TransientThermal solver was used, utilizing the NSGA-II multi-objective genetic algorithm integrated into ANSYS optiSLang for optimization. A dual-ellipsoid heat source model was used to simulate the weld heat source, ensuring optimal coordination between the equipment and software during implementation.

[0091] See also Figure 1 As shown, the method for calling a welding procedure with a weld profile matching function described in this embodiment includes the following steps:

[0092] The material grade of the parent material is collected and indexed in the preset material database according to the material grade to obtain the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material, and the welding materials are classified according to the thermal expansion coefficient range and thermal conductivity coefficient range respectively; the parent material is matched with the welding material according to the classification results to obtain the welding material; the material grade of the parent material can be collected to index the preset database to obtain thermophysical parameters such as thermal expansion coefficient and thermal conductivity coefficient, assist in matching welding materials, and ensure the synergy of material performance during the welding process; it can also help to establish an accurate finite element model and input the corresponding material parameters to make the simulation more realistic; in addition, it also provides basic information for subsequent quality control and process analysis, helping to evaluate welding quality and formulate reasonable process procedures. Obtaining the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material helps to accurately classify and match welding materials according to their characteristics to ensure the adaptability of welding; it provides key parameters for establishing a 3D finite element model, making the simulation closer to the actual welding process, thereby optimizing welding parameters; it can help predict problems such as heat affected zone width and residual stress that may occur during welding, providing a basis for controlling and improving welding quality.

[0093] Methods for obtaining classification results include:

[0094] Preset G thermal expansion coefficient thresholds, divide according to the preset G thermal expansion coefficient thresholds, obtain G+1 thermal expansion coefficient intervals, and number each thermal expansion coefficient interval in ascending order; preset M thermal conductivity coefficient thresholds, divide according to the preset M thermal conductivity coefficient thresholds, obtain M+1 thermal conductivity coefficient intervals, and number each thermal conductivity coefficient interval in ascending order; traverse and compare the thermal expansion coefficient and thermal conductivity coefficient of the welding material with the G+1 thermal expansion coefficient intervals and the M+1 thermal conductivity coefficient intervals, respectively, to obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number to which the welding material belongs as the classification result.

[0095] Methods for obtaining the material grade of welding materials include:

[0096] Obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number corresponding to the base material, and select the material corresponding to the material brand whose thermal expansion coefficient interval number and thermal conductivity coefficient interval number are the same as or adjacent to the thermal expansion coefficient interval number and thermal conductivity coefficient interval number of the base material as the material brand of the welding material.

[0097] Based on the selected base material and welding material, a 3D finite element model is established. The thermal expansion coefficient and thermal conductivity of the material are collected and input. The optimization variable range is set and the 3D finite element model is optimized. The optimization variables include current, voltage and welding speed. Based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output.

[0098] Methods for optimizing 3D finite element models include:

[0099] Collect the shape and size of the welded structure and input them into the finite element pre-processing software to create a three-dimensional geometric model;

[0100] Obtain the thermophysical parameters of the base material, including density, specific heat capacity, thermal expansion coefficient and thermal conductivity;

[0101] Input the thermal physical property parameters into the corresponding material area and build the corresponding heat conduction equation ;in, is the density of welding material; is the specific heat capacity; is temperature; For time; is the thermal conductivity; is the total thermal power, is the distance the heat source moves in the horizontal direction; is the distance the heat source moves in the longitudinal direction; is the distance the heat source moves in the vertical direction;

[0102] Preset current ,Voltage and welding speed The value range of the current setting ,Voltage and welding speed is the optimization variable;

[0103] Associate the optimization variables with the heat source model in the welding process, build a double ellipsoid heat source model, and divide the welding heat source into two ellipsoid parts to build a double ellipsoid heat source model. Figure 3 , which presents the geometric structure of the double ellipsoid heat source model and intuitively displays its three-dimensional shape. The heat flux density of the first half of the double ellipsoid heat source model ; Heat flux density in the second half ; Among them, the total thermal power Q is 3000W; is the arc thermal efficiency, such as Figure 3 0.85 in and are the semi-axis lengths of the front and rear ellipsoids in the horizontal direction, such as Figure 3 4mm and 8mm; is the semi-axis length of the front and rear ellipsoids in the longitudinal direction, such as Figure 3 3mm in the figure, where the semi-axis lengths of the front and rear ellipsoids in the longitudinal direction are equal, that is, the extension ranges of the two in the longitudinal direction are the same, which can achieve the purpose of simplifying the model; is the semi-axis length of the front and rear ellipsoids in the vertical direction, such as Figure 3Here, the semi-axis lengths of the front and rear ellipsoids in the vertical direction are equal, that is, the extension ranges of the two in the vertical direction are the same, which can also achieve the purpose of simplifying the model; and is the energy distribution coefficient of the front and rear ellipsoids, such as Figure 3 0.6 and 0.4 in;

[0104] Preset heat-affected zone width constraints (e.g., ≤5mm), residual stress constraints (e.g., ≤300MPa), and energy efficiency constraints (e.g., ≥80%). Use the NSGA-II algorithm to simultaneously optimize the heat-affected zone width, residual stress, and energy efficiency during welding, generate an optimal solution set, and select the parameter combination consisting of the optimization variables with the highest comprehensive score in the optimal solution set.

[0105] The total heat power is obtained by integrating the double ellipsoid heat source model , and then solve the heat conduction equation to obtain the temperature during welding;

[0106] Formulate and solve nonlinear equations Get the node displacement; where, is the stiffness matrix; is the node displacement; is the load vector; then according to the geometric equation and constitutive equation The residual stress is calculated; where is the strain vector; is a geometric matrix; is the stress vector, is the constitutive matrix; the stress vector is obtained by stress analysis based on the mechanical equilibrium equation, which is as follows:

[0107] ;

[0108] ;

[0109] ;

[0110] in, is the stress component; is the body force component; is the shear stress component, which indicates the magnitude of the shear force on the material in different planes; , representing the x-axis direction, y-axis direction and z-axis direction.

[0111] Methods for selecting the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set include:

[0112] Step 1: Define the objective function of the 3D finite element model. The objective function includes the first objective function, the second objective function and the third objective function. The first objective function ;in, , , and is the coefficient or index obtained by fitting the experimental data; the second objective function ;in, , , , and is the coefficient obtained by fitting the simulated data; the third objective function ;in, is the heat transfer efficiency, determined by fitting experimental data; and are the heat dissipation coefficient and the splashing energy loss coefficient, respectively, which are obtained by fitting the simulation data; and is a coefficient related to electrode loss and weld surface condition, determined by fitting experimental data; L is the weld length;

[0113] Step 2: Within the range of current, voltage and welding speed, randomly generate H individuals, each of which represents a set of parameter combinations obtained by splicing current, voltage and welding speed, that is, ;

[0114] Step 3: Calculate the three objective function values corresponding to each individual;

[0115] Step 4: For any two individuals p and q, if the three objective function values of individual p are less than or equal to the three objective function values corresponding to individual q, and at least one objective function value of individual p is less than the objective function value corresponding to individual q, then it is judged that individual p dominates individual q, denoted as p q.

[0116] Step 5: Repeat step 4 to obtain the first set of individuals consisting of all individuals that are not dominated by other individuals, and divide the first set of individuals into the first level;

[0117] Step 6: Repeat steps 4-5 for the remaining individuals that have not been classified into different levels, obtain new sets of individuals that are not dominated by the remaining individuals, and divide them into the next level until all individuals are divided into the corresponding level;

[0118] Step 7: Calculate the crowding degree of individuals in each level ;in, is the number of individuals, As the objective function, It represents the first objective function when It represents the second objective function when When represents the third objective function; and are the first and second order of adjacent individuals after sorting. objective function values; and The maximum and minimum values of the objective function in the population are set respectively; the crowding degree of the boundary individuals (individuals with the maximum and minimum objective function values) is set to infinity;

[0119] Step 8: Use the tournament selection method to randomly select K individuals from the population, and select the individual with the highest rank and a crowding degree higher than the preset crowding threshold as the parent;

[0120] Step 9: Select the parent individuals and Perform crossover operation to generate offspring individuals and :

[0121] ;

[0122] ;

[0123] in, is the crossover probability random number;

[0124] Step 10: Perform mutation operations on the offspring individuals to change the parameter values of the offspring individuals ;in, is the parameter value of the offspring individual after mutation; and For parameters The value range of is the random number of mutation probability; is the parameter value of the offspring individual;

[0125] Step 11: Merge the parent population of size N and the mutated offspring population of size N to obtain a merged population of size 2N.

[0126] Step 12: Perform non-dominated sorting on the merged population, select the top N individuals according to rank and crowding degree, and form a new population.

[0127] Step 13: When the preset update requirement is met (e.g., the preset maximum number of iterations is reached or the change of the optimal solution of consecutive M1 generations is less than the preset change threshold), stop updating the new population; and obtain the corresponding optimal solution set;

[0128] Step 14: Based on comprehensive scoring function Calculate the comprehensive score of each individual in the optimal solution set, and select the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set; , and The weights of the first objective function, the second objective function, and the third objective function are determined according to actual needs and can be further optimized through a nature-inspired optimization algorithm; and are the maximum and minimum values of the first objective function in the Pareto optimal solution set respectively; and are the maximum and minimum values of the second objective function in the Pareto optimal solution set respectively; and are the maximum and minimum values of the third objective function in the Pareto optimal solution set respectively;

[0129] Based on the optimized 3D finite element model, the heat affected zone width and residual stress are output. The specific results are as follows:

[0130] After the double ellipsoid heat source model is optimized, the temperature field distribution during welding is as follows: Figure 4 As shown, the maximum temperature reaches 1650°C, and the molten pool range (800-1500°C) clearly shows a concentrated heat source area. The ambient temperature is 25°C. The heat-affected zone width calculated through simulation is 2.8 mm (measured value: 2.8 ± 0.2 mm, error ≤ 3.7%), the weld depth is 3.5 mm, and the molten pool length is 8.6 mm. All parameters meet the preset constraints (e.g., HAZ ≤ 5 mm). The temperature field distribution results verify the dual-ellipsoid heat source model's ability to accurately simulate heat input.

[0131] The residual stress distribution during welding is as follows: Figure 5 As shown, the maximum stress is 380 MPa, concentrated primarily in the center of the weld; moderate stress is distributed in the heat-affected zone (180-280 MPa); and the minimum stress is ≤100 MPa (located at the distal end of the base metal). The residual stress value (245 MPa) obtained by solving the nonlinear equations agrees with the X-ray diffraction measured value by ≤5%, meeting the requirements of the GB / T7704-2017 standard. The stress distribution contour map intuitively reflects the welding deformation trend, demonstrating the engineering feasibility of the residual stress constraint range.

[0132] When simulating and analyzing the welding process using finite element software, Figure 6 As shown, Figure 6 The left and middle correspond to the temperature field and stress distribution respectively. Figure 6The deformation parameters are obtained on the far right. The deformation parameters show that the maximum deformation is 2.5mm, the medium deformation is 1.0-2.0mm, and the minimum deformation is ≤0.5mm. These data help to evaluate the impact of welding deformation on weld formation and quality, and provide a key reference for the dynamic adjustment and optimization of subsequent welding process regulations, further reflecting the technical advantages of the present invention in multi-dimensional control of welding quality. Figure 6 The comprehensive analysis of these parameters realizes the comprehensive simulation and evaluation of the welding process, and strengthens the scientificity and practicality of this embodiment.

[0133] The Cr and Ni content in the molten pool is obtained, and a Cr / Ni element calibration equation is established through fitting. When the Cr content is detected to be lower than the preset content threshold, a feedback mechanism for adjusting the optimization variables is triggered. An integrated LIBS system (laser-induced breakdown spectroscopy technology system) collects molten pool spectral data in real time, and obtains the Cr and Ni content in the molten pool through spectral analysis. The Cr and Ni content in the molten pool can be obtained and a calibration curve can be established through fitting to determine whether the composition is qualified. When the Cr content is abnormal, a feedback mechanism is triggered to adjust the welding process in time. The content of these elements will affect the performance and forming of the weld. Understanding their content helps predict and control the quality of the weld. It can also provide data support for subsequent quality analysis and process optimization to ensure the reliability and stability of the welded structure.

[0134] The method for fitting and establishing the Cr / Ni element calibration equation includes:

[0135] Standard samples with different Cr and Ni content are obtained, and spectra of the standard samples are collected; the characteristic peak positions of Cr and Ni in the spectra are respectively identified and obtained, and the integrated intensity of the characteristic peaks of Cr and Ni in each standard sample is obtained by integrating the spectral intensity within the wavelength range where the characteristic peaks are located. The known contents of Cr and Ni in the standard samples are used as the abscissa, and the corresponding characteristic peak integrated intensity is used as the ordinate, and a calibration equation is obtained by least squares fitting.

[0136] Methods for triggering feedback mechanisms that adjust optimization variables include:

[0137] The Cr and Ni element contents are detected in real time and compared with the preset content thresholds. When the Cr or Ni element content is lower than the preset content threshold, a feedback mechanism is triggered. The feedback mechanism includes increasing the wire feeding speed to the preset wire feeding speed threshold, reducing the welding speed to the preset welding speed threshold, and adjusting the shielding gas flow to the preset coverage threshold.

[0138] Collect 3D point cloud data of welds, extract weld contour features, match weld contour features with contours in a preset contour library to obtain matching results; weld contour features include geometric features and topological features; collect 3D point cloud data of welds, which can be used to extract weld contour features, match them with a preset contour library, determine whether the weld formation meets the standards, and then perform trajectory correction to ensure weld quality; it can provide a data basis for the subsequent establishment of a multi-level rule library based on weld contour matching, facilitating the definition and calling of process procedures; it can also intuitively present the 3D shape of the weld, which is used to analyze the causes of welding defects and help optimize welding process parameters.

[0139] Methods for obtaining geometric features include:

[0140] For edge points in 3D point cloud data , found by neighborhood search Neighborhood points , construct the covariance matrix ;

[0141] in, Neighborhood points The center of mass; is the edge point in the 3D point cloud data No. Neighborhood points; is the transpose of the vector;

[0142] Pair covariance matrix Perform eigenvalue decomposition ;in, is the eigenvalue diagonal matrix; is the eigenvector matrix; take the eigenvector corresponding to the minimum eigenvalue as the point Normal vector ;

[0143] Calculate edge points based on normal vectors and neighborhood points Curvature ; Count the curvatures of all edge points, obtain the curvature set, and use the curvature set as a geometric feature.

[0144] Methods for obtaining topological features include:

[0145] Step A: Select dimension R and sort the 3D point cloud data according to the value of the selected dimension.

[0146] Step B: Take the point corresponding to the middle value of the sorting result as the root node, and divide the 3D point cloud data into two parts, the left part and the right part, according to the root node;

[0147] Step C: Repeat step B for the left and right parts respectively to build a subtree;

[0148] Step D: Repeat steps B to C until both the left and right parts contain only one node.

[0149] Step E: From the seed point Initially, region growing is performed based on the normal vector similarity and distance threshold between points;

[0150] Step F: Obtain the boundary points of the growth area , that is, the weld contour point , calculate weld contour points With neighboring points The normal vector angle ;in, Weld contour points The normal vector of Neighborhood points The normal vector of is greater than the set normal vector similarity threshold, and the point with dot The distance is less than the distance threshold; then the point Add the growth area and continue to do the above operation for the newly added points until there are no more points that meet the conditions to be added;

[0151] Step G: Count all the boundary points in the growing area to obtain a topological set, and use the topological set as a topological feature.

[0152] Methods for obtaining weld profile matching results include:

[0153] Calculate the Euclidean distance between the geometric features of the weld profile and the geometric features of the profiles in the profile library (e.g., 50 standard profiles including V-shaped, U-shaped, and J-shaped grooves). Calculate the similarity between the topological features of the weld profile and the topological features of the profiles in the profile library. Compare and judge based on the preset distance matching threshold and similarity matching threshold:

[0154] If the Euclidean distance between the weld contour and the contour in the contour library is less than the distance matching threshold, and the similarity is higher than the similarity matching threshold, then it is determined that the weld contour is the same as the matched contour in the contour library;

[0155] If the following situations occur:

[0156] There is more than one matching result;

[0157] The Euclidean distance is greater than the distance matching threshold and the similarity is greater than the similarity matching threshold;

[0158] The Euclidean distance is greater than the distance matching threshold and the similarity is lower than the similarity matching threshold;

[0159] The Euclidean distance is less than the distance matching threshold and the similarity is less than the similarity matching threshold;

[0160] The contour in the contour library corresponding to the highest similarity is selected as the final matching result.

[0161] Methods for correcting the weld contour trajectory based on the weld contour matching results include:

[0162] Get the weld offset during welding. When the weld offset is greater than the preset offset threshold, the weld point correction formula is used. Perform trajectory correction on welding points; is the welding point coordinate correction value; It is an empirical coefficient, generally taken as 0.8, which can be calculated based on the response time of the welding robot; is the actual weld offset; is the preset offset threshold;

[0163] Construct a multi-level rule library based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching. Define the process procedures according to the multi-level rule library for dynamic calling of the process procedures in the later stage.

[0164] Reference Figure 2 ,The methods for building a multi-level rule base include:

[0165] Step a, obtaining the material grade of the base material, the material grade of the welding material, the thermal expansion coefficient and thermal conductivity of the material, and outputting the heat-affected zone width and residual stress, the content of Cr and Ni elements, the profile characteristics, and the welding offset based on the optimized 3D finite element model;

[0166] Step b: designing and storing rule logic for welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching during the welding process, wherein the rule logic includes basic rules, dynamic rules, and composite rules;

[0167] Step c: Use the rule engine to implement condition matching and action triggering;

[0168] Methods for defining process procedures include:

[0169] Real-time collection of base material grade, thermal expansion coefficient and thermal conductivity, Cr and Ni content, and three-dimensional point cloud data of welds;

[0170] Compare the collected data with the condition fields in the rule logic defined in the multi-level rule library to trigger. If multiple rules are triggered at the same time, they will be executed in the order of preset priority.

[0171] The corresponding action field in the triggered rule logic is converted into a device instruction, and the device instruction is sent to the corresponding device, and the device executes the device instruction.

[0172] If the device instruction is a feedback mechanism for adjusting optimization variables and a trajectory correction rule, the Cr and Ni content is collected again to determine whether it has recovered to the preset content threshold. If not, the feedback mechanism is triggered until the Cr and Ni content recovers to the preset content threshold.

[0173] Rescan the adjusted weld profile and perform a second match with the preset profile library; if the match fails E times in a row, an alarm is triggered and the machine is shut down, prompting manual intervention; and the adjustment data, rule trigger records, and quality inspection results are stored in the preset call database.

[0174] Example 2

[0175] See also Figure 7 As shown, the welding procedure specification calling system with weld profile matching function described in this embodiment includes:

[0176] Material matching module: collects the material grade of the parent material, and indexes it in the preset material database according to the material grade, obtains the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material, classifies the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient, and obtains the classification results; matches the parent material with the welding material according to the classification results to obtain the material grade of the welding material;

[0177] Model prediction module: Based on the selected base material and welding material, a 3D finite element model is established, the thermal expansion coefficient and thermal conductivity of the material are collected and input, the optimization variable range is set, and the 3D finite element model is optimized; the optimization variables include current, voltage, and welding speed; based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output;

[0178] Feedback adjustment module: obtains the Cr and Ni content in the molten pool, fits and establishes a Cr / Ni element calibration equation, and triggers a feedback mechanism to adjust the optimization variables when the Cr content is detected to be lower than the preset content threshold;

[0179] Contour matching module: collects 3D point cloud data of welds, extracts weld contour features, matches weld contour features with contours in a preset contour library, and obtains weld contour matching results; performs trajectory correction on weld contours based on the weld contour matching results;

[0180] Procedure calling module: Build a multi-level rule base based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching. Define process procedures according to the multi-level rule base for dynamic calling of process procedures in the later stage. In the later stage, the process procedures stored in the multi-level rule base can be directly called through the procedure calling module for weld profile matching.

[0181] Reference Figure 8 In the material matching module, the base material grade 12Cr1MoV is displayed, which can be used to index its key parameters such as thermal expansion coefficient and thermal conductivity in the preset material database, providing a basis for welding material classification and matching. As shown in Table 1, Table 2 and Table 3:

[0182] Table 1 Physical properties of base material (12Cr1MoV)

[0183]

[0184] Table 1 provides the core physical parameters of the base material 12Cr1MoV. The density, specific heat capacity, and thermal conductivity are directly used to solve the heat conduction equation in the finite element model, while the thermal expansion coefficient and elastic modulus are used for residual stress and deformation analysis. The parameter testing standards cover internationally accepted methods (such as ASTM and ISO), ensuring authoritative data. This provides the basic input for establishing the 3D finite element model and supports the output of technical characteristics such as heat-affected zone width and residual stress.

[0185] Table 2 Measured data of base material (12Cr1MoV)

[0186]

[0187] Test environment: room temperature (25℃±1℃), humidity 50%±5%.

[0188] Sample numbers: MT-12Cr1MoV-001 to MT-12Cr1MoV-005 (5 sets of parallel samples).

[0189] Table 2 provides measured data for the base material 12Cr1MoV. The accuracy of the material parameters was verified using high-precision testing methods such as the Archimedean drainage method and laser flash method. The measured data remained within reasonable tolerances compared to the theoretical values (Table 1), such as ±0.02 g / cm³ for density and ±1.5% for thermal conductivity. This demonstrates the reliability of the experimental data and provides reliable input parameters for the finite element model, ensuring the accuracy of the heat conduction equation solution and residual stress calculations.

[0190] Table 3 Parallel sample measured data

[0191]

[0192] Table 3 shows the measured data of five parallel samples. The test environment was room temperature (25°C ± 1°C) and humidity was 50% ± 5%. The Archimedes drainage method, laser flash method (LFA467), and dilatometer method (Netzsch DIL402) were used to ensure test accuracy. The density was ± 0.02 g / cm³, the thermal conductivity was ± 1.5%, and the thermal expansion coefficient was ± 0.3 × 10 -6 / K, all within the allowable error range, demonstrating the stability of the parent material properties. The excellent repeatability of the parallel sample data ensures the reliability of the finite element model input parameters (such as density and thermal conductivity), supporting the accuracy of the 3D finite element model. The measured values differ from the theoretical values (Table 1) by less than 3%, providing experimental validation for the simulated results of the heat-affected zone width and residual stress.

[0193] The material performance comparison radar chart intuitively presents the welding material matching results. The recommended welding material is ER70S-6, with a matching rate of 85%, which can ensure the synergistic effect of material performance during welding. See Table 4, Table 5 and Table 6:

[0194] Table 4 Physical properties of welding material (ER70S-6)

[0195]

[0196] Table 4 provides the core physical parameters of the welding material ER70S-6, and its thermal expansion coefficient (13.2×10 -6 / K) and base material 12Cr1MoV (12.5×10 -6 / K, Table 1), the thermal conductivity of 45W / (m・K) matches that of the base material at 42W / (m・K), satisfying the matching rule that the thermal expansion coefficient interval and the thermal conductivity coefficient interval are the same or adjacent, ensuring that the thermal stress at the welding interface is controllable.

[0197] Test standard description: Thermal conductivity is calculated through thermal diffusivity (ASTM E1461). Other parameters are tested using international standards.

[0198] Table 5 Measured data of welding material (ER70S-6)

[0199]

[0200] Test environment: room temperature (25℃±1℃), humidity 50%±5%.

[0201] Sample numbers: MT-ER70S6-001 to MT-ER70S6-005 (5 sets of parallel samples).

[0202] Table 5 presents the measured data of the welding material (ER70S-6), covering three key parameters: density, thermal conductivity, and thermal expansion coefficient. The test was conducted in a strictly controlled environment with a room temperature of 25°C ± 1°C and a humidity of 50% ± 5%. Five sets of parallel samples (MT-ER70S6-001 to MT-ER70S6-005) were used to ensure data reliability and repeatability. The density was measured by the Archimedean drainage method and the test value was 7.86 g / cm 3 The allowable error is ±0.02g / cm 3The thermal conductivity was measured by the laser flash method (LFA467) with an error of ±1.5%; the thermal expansion coefficient (20-300°C) was measured by the dilatometer method (Netzsch DIL402) and the result was 13.1×10 -6 / K, allowable error ±0.3×10 -6 The precise measurement of these parameters provides a basis for evaluating the thermal compatibility of the welding consumables and the base material (Table 1), ensuring that the width of the heat-affected zone and residual stresses during welding are controllable. It also provides key input for heat source calculation and thermal-mechanical analysis of finite element models, strengthening the technical logic of matching the properties of the welding consumables and base material to improve welding quality.

[0203] Further field measurements on five parallel samples (Table 6) verified the stability of the ER70S-6 welding consumable. The sample density, thermal conductivity, and thermal expansion coefficient showed minimal fluctuations, demonstrating the stable quality of the welding consumable and providing a reliable material parameter basis for the subsequent development of welding procedure specifications. These data, in conjunction with Table 5, ensure the accuracy of the thermophysical property parameters in the finite element model, and thus the reliability of the simulation results for the heat-affected zone width and residual stress. This demonstrates the rigorous and scientific nature of the present invention in the acquisition and application of material parameters.

[0204] Table 6 Parallel sample measured data

[0205]

[0206] Table 6 shows measured data for five parallel samples of the welding consumable (ER70S-6), covering density, thermal conductivity, and thermal expansion coefficient. Sample numbers are MT-ER70S6-001 to MT-ER70S6-005. The testing environment is consistent with that in Table 5 (room temperature 25°C ± 1°C, humidity 50% ± 5%). The parallel sample testing verifies the stability of the welding consumable performance: the density test values are between 7.85 and 7.87 g / cm 3 The thermal conductivity is between 44.8-45.0W / (m·K), and the thermal expansion coefficient is between 13.0−13.2×10 −6 / K, and the fluctuations of various parameters are all within a reasonable range, indicating that the welding material performance is stable. These data further support the measured results in Table 5, provide a more sufficient basis for evaluating the matching of welding materials and base materials, and ensure the reliability of the finite element model input parameters, thereby accurately simulating the heat conduction, deformation, and stress distribution during the welding process.

[0207] In the optimization parameter area, three optimization variables, current (100-300A), voltage (15-35V) and welding speed (2-10mm / s), and their corresponding variable intervals are set. The optimization variables are optimized in combination with the 3D finite element model. Adjusting the optimization variables can optimize the welding process. The experimental data can be referred to Table 7:

[0208] Table 7 Experimental data comparison table

[0209]

[0210] Table 7 is a comparison of experimental data, presenting simulated and measured data for heat-affected zone width, residual stress, and energy efficiency under eight different welding parameters (current, voltage, and welding speed), as well as spatter rate and uniformity score.

[0211] Table 7 shows that the error between the simulated and measured HAZ widths for the eight experimental groups ranged from 1.9% to 5.9%, with an average error of 3.2% and a standard deviation of 0.15 mm, meeting the requirements of GB / T 19869.1. For the second experiment (current 150 A, voltage 22 V, welding speed 4.5 mm / s), the simulated HAZ width was 2.8 mm, while the measured value was 2.9 mm, representing a mere 3.4% error. The error for the eighth experiment (current 300 A, voltage 35 V, welding speed 10.0 mm / s) was 5.9%, the largest error, but still below the 10% error standard permitted by industrial testing. The high agreement between the simulation results and the measured data demonstrates that the 3D finite element model can accurately predict the HAZ width and provide a reliable basis for optimizing process parameters.

[0212] The error between the simulated and measured residual stress values ranged from 6.4% to 8.6%, with an average error of 7.3%. For example, in experiment group 5 (current 230A, voltage 32V, welding speed 8.0mm / s), the error between the simulated value of 315MPa and the measured value of 330MPa was 4.5%, meeting the ±10% error requirement specified in GB / T7704-2017. The model's residual stress simulation accuracy met engineering application requirements, validating the effectiveness of the dual-ellipsoid heat source model and the nonlinear stress solution method.

[0213] When the current increases from 120A to 300A, the width of the heat-affected zone increases from 2.5mm to 4.0mm, showing a nearly linear increase (slope of 0.0078mm / A). This is because the increased current increases heat input and expands the high-temperature area in the weld pool. In scenarios where thermal deformation must be controlled (such as welding thin-walled structures), the heat-affected zone width can be controlled to within 3.0mm by limiting the upper current limit (e.g., ≤200A).

[0214] When the welding speed increases from 3.0 mm / s to 10.0 mm / s, energy efficiency (simulation) increases from 82% to 92%, indicating that increasing the welding speed can reduce heat input per unit weld length and lower energy loss. However, excessively fast welding speeds (e.g., ≥8 mm / s) can significantly increase the spatter rate (from 3.0% to 4.2%), affecting weld quality (the uniformity score decreases from 3.5 to 3.0). Considering both energy efficiency and weld quality, the recommended welding speed range is 4.5-6.0 mm / s (corresponding to an energy efficiency of 85%-88% and a spatter rate of ≤2.5%).

[0215] By comparing the simulation results with the measured results, the accuracy of the finite element model in simulating the welding process was verified. The influence of various parameters on welding quality (heat-affected zone width, residual stress), energy utilization (energy efficiency) and forming effect (spatter rate, uniformity score) was analyzed. This provided an experimental basis for optimizing welding process parameters based on the NSGA-Ⅱ algorithm and ensured that the optimized parameters were effective in actual welding. See Table 8:

[0216] Table 8 Partial NSGA-Ⅱ algorithm optimization process data

[0217]

[0218] Table 8 shows some of the optimization process data of the NSGA-II algorithm, including 10 sets of welding parameters (current, voltage, welding speed) and the corresponding heat-affected zone width, residual stress, energy efficiency and comprehensive score.

[0219] In Table 8, solution number 2 (current 200A, voltage 24V, welding speed 5.0mm / s) achieved the highest comprehensive score (S=0.91), with excellent heat-affected zone width (2.8mm), residual stress (245MPa), and energy efficiency (85%). In contrast, solution number 7 (current 220A, voltage 26.5V, welding speed 4.0mm / s), while having the smallest heat-affected zone width (2.5mm), exhibited only an energy efficiency of 80.5%, and a high residual stress of 290MPa, resulting in poor overall performance. The algorithm effectively balanced multi-objective conflicts through non-dominated sorting and congestion calculation, avoiding performance imbalances caused by single-metric optimization. During the optimization process, the average comprehensive score for the first 50 generations was 0.78, increasing to 0.85 by the 100th generation and stabilizing between 0.88 and 0.91 by the 200th generation. This indicates that the algorithm converged to the optimal solution within 200 iterations, achieving computational efficiency that met the requirements of real-time process optimization.

[0220] This data allows us to observe the algorithm's search process for multi-objective optimization (heat-affected zone width, residual stress, and energy efficiency). The comprehensive score S reflects the quality of each solution. For example, solution number 2 has a high comprehensive score of 0.91. Its parameters (current 200A, voltage 24.0V, welding speed 5.0mm / s) result in a heat-affected zone width of 2.80mm, residual stress of 245MPa, and energy efficiency of 85.0%, achieving a good balance among multiple objectives. This provides a direct basis for selecting the optimal process parameters and fully demonstrates the algorithm's ability to balance and optimize multiple objectives in welding process optimization.

[0221] The temperature distribution curve, heat-affected zone width of 2.8 mm, and residual stress of 245 MPa obtained based on the output of the optimized 3D finite element model can be used to monitor and evaluate welding quality, and also provide a reference for the dynamic adjustment and call of subsequent welding process specifications.

[0222] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications or substitutions that can be easily conceived by a person skilled in the art within the technical scope disclosed in the present invention should be included within the scope of protection of the present invention. Therefore, the scope of protection of the present invention should be based on the scope of protection of the claims.

[0223] Finally: The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principles of the present invention should be included in the scope of protection of the present invention.

Claims

1. A welding procedure specification calling method with a weld profile matching function, characterized in that: The steps include: Collect the material grade of the parent material, and index it in the preset material database according to the material grade, obtain the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material, classify the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient, and obtain the classification results; Match the welding materials to the parent materials according to the classification results to obtain the material grades of the welding materials; Based on the selected base material and welding material, a 3D finite element model is established. The thermal expansion coefficient and thermal conductivity of the material are collected and input. The optimization variable range is set and the 3D finite element model is optimized. The optimization variables include current, voltage and welding speed. Based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output. Obtain the Cr and Ni content in the molten pool, and establish a Cr / Ni element calibration equation by fitting. When the Cr content is detected to be lower than the preset content threshold, a feedback mechanism for adjusting the optimization variables is triggered. Collect 3D point cloud data of the weld, extract weld contour features, match the weld contour features with the contours in the preset contour library to obtain the weld contour matching results; perform trajectory correction on the weld contour based on the weld contour matching results; Build a multi-level rule base based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching. Define process procedures based on the multi-level rule base for dynamic calling of process procedures in the later stage. Methods for building a multi-level rule base include: Step a, obtaining the material grade of the base material, the material grade of the welding material, the thermal expansion coefficient and thermal conductivity of the material, and outputting the heat-affected zone width and residual stress, the content of Cr and Ni elements, the profile characteristics, and the welding offset based on the optimized 3D finite element model; Step b: designing and storing rule logic for welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld profile matching during the welding process, wherein the rule logic includes basic rules, dynamic rules, and composite rules; Step c: Use the rule engine to implement condition matching and action triggering; Methods for defining process procedures include: Real-time collection of base material grade, thermal expansion coefficient and thermal conductivity, Cr and Ni content, and three-dimensional point cloud data of welds; Compare the collected data with the condition fields in the rule logic defined in the multi-level rule library to trigger. If multiple rules are triggered at the same time, they will be executed in the order of preset priority. Convert the corresponding action field in the triggered rule logic into a device instruction, and send the device instruction to the corresponding device, so that the device executes the device instruction; If the device instruction is a feedback mechanism for adjusting optimization variables and a trajectory correction rule, the Cr and Ni content is collected again to determine whether it has recovered to the preset content threshold. If not, the feedback mechanism is triggered until the Cr and Ni content recovers to the preset content threshold. Rescan the adjusted weld profile and perform a second match with the preset profile library; if the match fails E times in a row, an alarm is triggered and the machine is shut down, prompting manual intervention; and the adjustment data, rule trigger records, and quality inspection results are stored in the preset call database.

2. The method for calling welding procedure with weld profile matching function according to claim 1, characterized in that: Methods for optimizing 3D finite element models include: Collect the shape and size of the welded structure and input them into the finite element pre-processing software to create a three-dimensional geometric model; Obtain the thermophysical parameters of the base material, including density, specific heat capacity, thermal expansion coefficient and thermal conductivity; Input the thermophysical property parameters into the corresponding material area and build the corresponding heat conduction equation; Preset the value range of current, voltage and welding speed, and set current, voltage and welding speed as optimization variables; The optimization variables are associated with the heat source model of the welding process, and the welding heat source is divided into two ellipsoid parts to build a double ellipsoid heat source model; Preset heat-affected zone width constraint intervals, residual stress constraint intervals, and energy efficiency constraint intervals; use the NSGA-II algorithm to simultaneously optimize the heat-affected zone width, residual stress, and energy efficiency during welding, generate an optimal solution set, and select the parameter combination consisting of the optimization variables with the highest comprehensive score in the optimal solution set; The total heat power is obtained by integrating the double ellipsoid heat source model , and then solve the heat conduction equation to obtain the temperature during welding; The node displacement is obtained by establishing and solving a set of nonlinear equations; the residual stress is then calculated based on the geometric equations and constitutive equations.

3. The method for calling welding procedure with weld profile matching function according to claim 2, characterized in that: Methods for selecting the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set include: Step 1: define the objective function of the 3D finite element model, which includes a first objective function, a second objective function, and a third objective function; Step 2: Randomly generate H individuals within the range of current, voltage, and welding speed, each of which represents a parameter combination obtained by splicing current, voltage, and welding speed; Step 3: Calculate the three objective function values corresponding to each individual; Step 4: For any two individuals p and q, if the three objective function values of individual p are less than or equal to the three objective function values corresponding to individual q, and at least one objective function value of individual p is less than the objective function value corresponding to individual q, then individual p is judged to dominate individual q; Step 5: Repeat step 4 to obtain the first set of individuals consisting of all individuals that are not dominated by other individuals, and divide the first set of individuals into the first level; Step 6: Repeat steps 4-5 for the remaining individuals that have not been classified into different levels, obtain new sets of individuals that are not dominated by the remaining individuals, and divide them into the next level until all individuals are divided into the corresponding level; Step 7: Calculate the crowding degree of individuals in each level; Step 8: Use the tournament selection method to randomly select K individuals from the population, and select the individual with the highest rank and a crowding degree higher than the preset crowding threshold as the parent; Step 9: Select the parent individuals and Perform crossover operation to generate offspring individuals and ; Step 10: performing mutation operations on the offspring individuals to change the parameter values of the offspring individuals; Step 11: Merge the parent population of size N and the mutated offspring population of size N to obtain a merged population of size 2N. Step 12: Perform non-dominated sorting on the merged population, select the top N individuals according to rank and crowding degree, and form a new population; Step 13: When the preset update requirements are met, stop updating the new population; obtain the corresponding optimal solution set; Step 14: Calculate the comprehensive score of each individual in the optimal solution set based on the comprehensive scoring function, and select the parameter combination composed of the optimization variables with the highest comprehensive score in the optimal solution set.

4. The method for calling welding procedure with weld profile matching function according to claim 1, characterized in that: The weld profile features include geometric features and topological features; Methods for obtaining geometric features include: For edge points in three-dimensional point cloud data, k neighboring points are found through neighborhood search and a covariance matrix is constructed. The covariance matrix is decomposed into eigenvalues. The eigenvector corresponding to the minimum eigenvalue is taken as the normal vector of the point. The curvature of the edge point is calculated based on the normal vector and the neighborhood points. The curvature of all edge points is counted to obtain a curvature set, which is used as a geometric feature.

5. The method for calling welding procedure with weld profile matching function according to claim 4, characterized in that: Methods for obtaining topological features include: Step A: Select dimension R and sort the 3D point cloud data according to the value of the selected dimension. Step B: Take the point corresponding to the middle value of the sorting result as the root node, and divide the 3D point cloud data into two parts, the left part and the right part, according to the root node; Step C: Repeat step B for the left and right parts respectively to build a subtree; Step D: Repeat steps B to C until both the left and right parts contain only one node. Step E: From the seed point Initially, region growing is performed based on the normal vector similarity and distance threshold between points; Step F: Get the boundary points of the growth area , that is, the weld contour point , calculate weld contour points With neighboring points The normal vector angle ;like is greater than the set normal vector similarity threshold, and the point with dot The distance is less than the distance threshold; then the point Add the growth area and continue to do the above operation for the newly added points until there are no more points that meet the conditions to be added; Step G: Count all the boundary points in the growing area to obtain a topological set, and use the topological set as a topological feature.

6. The method for calling welding procedure specifications with weld profile matching function according to claim 5, characterized in that: Methods for obtaining weld profile matching results include: Calculate the Euclidean distance between the geometric features of the weld contour and the geometric features of the contours in the contour library, and calculate the similarity between the topological features of the weld contour and the topological features of the contours in the contour library; compare and judge based on the preset distance matching threshold and similarity matching threshold: If the Euclidean distance between the weld contour and the contour in the contour library is less than the distance matching threshold, and the similarity is higher than the similarity matching threshold, then it is determined that the weld contour is the same as the matched contour in the contour library; If the following situations occur: There is more than one matching result; The Euclidean distance is greater than the distance matching threshold and the similarity is greater than the similarity matching threshold; The Euclidean distance is greater than the distance matching threshold and the similarity is lower than the similarity matching threshold; The Euclidean distance is less than the distance matching threshold and the similarity is less than the similarity matching threshold; The contour in the contour library corresponding to the highest similarity is selected as the final matching result; Methods for correcting the weld contour trajectory based on the weld contour matching results include: The weld offset during the welding process is obtained. When the weld offset is greater than the preset offset threshold, the weld point trajectory is corrected.

7. The method for calling welding procedure with weld profile matching function according to claim 1, characterized in that: Methods for obtaining classification results include: Preset G thermal expansion coefficient thresholds, divide according to the preset G thermal expansion coefficient thresholds, obtain G+1 thermal expansion coefficient intervals, and number each thermal expansion coefficient interval in ascending order; preset M thermal conductivity coefficient thresholds, divide according to the preset M thermal conductivity coefficient thresholds, obtain M+1 thermal conductivity coefficient intervals, and number each thermal conductivity coefficient interval in ascending order; traverse and compare the thermal expansion coefficient and thermal conductivity coefficient of the welding material with the G+1 thermal expansion coefficient intervals and the M+1 thermal conductivity coefficient intervals, respectively, to obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number to which the welding material belongs as the classification result.

8. The method for calling welding procedure specifications with weld profile matching function according to claim 1, characterized in that: Methods for obtaining the material grade of welding materials include: Obtain the thermal expansion coefficient interval number and thermal conductivity coefficient interval number corresponding to the base material, and select the material corresponding to the material brand whose thermal expansion coefficient interval number and thermal conductivity coefficient interval number are the same as or adjacent to the thermal expansion coefficient interval number and thermal conductivity coefficient interval number of the base material as the material brand of the welding material.

9. A welding procedure specification calling system with a weld profile matching function, used to implement a welding procedure specification calling method with a weld profile matching function as claimed in any one of claims 1 to 8, characterized in that: include: Material matching module: collects the material grade of the parent material, and indexes it in the preset material database according to the material grade, obtains the thermal expansion coefficient and thermal conductivity coefficient corresponding to the parent material, classifies the welding materials according to the thermal expansion coefficient and thermal conductivity coefficient, and obtains the classification results; Match the welding materials to the parent materials according to the classification results to obtain the material grades of the welding materials; Model prediction module: Based on the selected base material and welding material, a 3D finite element model is established, the thermal expansion coefficient and thermal conductivity of the material are collected and input, the optimization variable range is set, and the 3D finite element model is optimized; the optimization variables include current, voltage, and welding speed; based on the optimized 3D finite element model, the heat-affected zone width and residual stress are output; Feedback adjustment module: obtains the Cr and Ni content in the molten pool, fits and establishes a Cr / Ni element calibration equation, and triggers a feedback mechanism to adjust the optimization variables when the Cr content is detected to be lower than the preset content threshold; Contour matching module: collects 3D point cloud data of welds, extracts weld contour features, matches weld contour features with contours in a preset contour library, and obtains weld contour matching results; performs trajectory correction on weld contours based on the weld contour matching results; Procedure calling module: Build a multi-level rule library based on welding material matching, optimization variable optimization, feedback adjustment of optimization variables, and weld contour matching. Define process procedures based on the multi-level rule library for dynamic calling of process procedures in the later stage.

Images