A smart welding method and system for automotive parts

Abstract

This invention discloses an intelligent welding method and system for automotive parts, relating to the field of intelligent welding. The method includes: determining the basic components of the automotive parts to be welded and generating a welding chain for the automotive parts; connecting a welding robotic arm, introducing front and back welding processes, and extracting welding current data from the welding robotic arm; analyzing the welding distribution characteristics of the basic components; adding welding distribution characteristics to finite element analysis software, and evaluating residual stress and deformation under different welding conditions, combined with the welding current data; and issuing welding defect alerts according to automotive welding quality standards and the residual deformation evaluation results. This application solves the technical problems of insufficient adaptability between welding process parameters and component characteristics, and low accuracy in the quantitative evaluation of welding stress and deformation in existing welding operations, achieving the technical effect of improving the accuracy of the adaptation between welding process parameters and the characteristics of the basic components of the parts, and improving the accuracy of the quantitative evaluation of welding stress and deformation.

Description

Technical Field

[0001] This application relates to the field of intelligent welding, and in particular to an intelligent welding method and system for automotive parts. Background Technology

[0002] The welding quality of automotive parts directly determines the structural strength, assembly accuracy, and service safety of the entire vehicle, making it a core aspect of product quality control in the automotive manufacturing industry. Currently, industry-wide quality control for automotive parts welding largely relies on preset, fixed welding process parameters. This involves using robotic welding arms to perform standardized single- or double-sided welding processes, supplemented by manual sampling and routine stress and deformation testing to determine welding quality. Existing techniques depend solely on empirically based, fixed welding parameters and lack dynamic assessment of the heat accumulation effects across multiple processes. This leads to residual stress concentration, deformation exceeding tolerances, and other process defects in actual welding, impacting the dimensional accuracy and service reliability of parts.

[0003] At present, the welding of automotive parts suffers from technical problems such as insufficient compatibility between welding process parameters and component characteristics, and low accuracy in the quantitative assessment of welding stress and deformation. Summary of the Invention

[0004] This application provides an intelligent welding method and system for automotive parts. It employs a method that determines the base material and weld groove form of the parts to be welded, generates a welding chain, connects a welding robotic arm, introduces forward and reverse welding processes, extracts welding current datasets, analyzes the welding distribution characteristics such as symmetrical and intersecting thermal fields corresponding to the forward and reverse processes, adds these characteristics to finite element analysis software, and evaluates residual stress and deformation under different working conditions based on the current dataset. It also provides welding defect alerts based on automotive welding quality standards and evaluation results. This approach solves the technical problems of insufficient compatibility between welding process parameters and component characteristics, and low accuracy in the quantitative evaluation of welding stress and deformation in existing automotive part welding operations. It achieves the technical effect of improving the accuracy of the compatibility between welding process parameters and the basic component characteristics of the parts, increasing the accuracy of the quantitative evaluation of welding stress and deformation, and thus reducing the welding defect rate.

[0005] This application provides an intelligent welding method for automotive parts, comprising: determining the basic components of the automotive parts to be welded, including the base material, weld bevel form, fixture positioning points, welding electrodes, and connecting fasteners, and generating a welding chain for the automotive parts; based on the welding chain, connecting a welding robot arm, introducing a front welding process and a back welding process, and extracting the welding current dataset of the welding robot arm, wherein the front welding process includes M front welding nodes and the back welding process includes N back welding nodes; based on the front welding process and the back welding process, analyzing the welding distribution characteristics of the basic components, wherein the welding distribution characteristics include the thermal field symmetry distribution characteristics corresponding to mirror-symmetric welds and the thermal field cross distribution characteristics corresponding to cross-symmetric welds; adding the welding distribution characteristics to finite element analysis software, and under different welding conditions, evaluating residual stress and deformation in conjunction with the welding current dataset, and issuing welding defect alerts according to automotive welding quality standards and residual deformation evaluation results.

[0006] In a possible implementation, welding defect alerts are issued according to automotive welding quality standards and residual deformation assessment results, and the following processing is performed: setting impact speed and impact load conditions, and obtaining a weld micro-wear dataset; determining the heat-affected zone width and grain coarsening index based on the weld micro-wear dataset; and determining a first constraint feature based on the heat-affected zone width and grain coarsening index, wherein the first constraint feature is used to improve the accuracy of residual deformation assessment.

[0007] In a possible implementation, the following processing is performed: setting alternating load frequency and amplitude conditions, and obtaining a fatigue response dataset; based on the fatigue response dataset, evaluating the fatigue resistance of the weld, and determining a second constraint feature, which is used to improve the accuracy of the correlation assessment between residual deformation and fatigue failure, so as to improve the reliability of automotive welding quality judgment.

[0008] In a possible implementation, the following processing is performed: connecting an image recognition device to collect a weld micro-wear dataset, which includes grain morphology image samples centered on M+N weld bead control points; using a strain gauge array to monitor the strain response process of the weld area of ​​automotive parts in different impact directions, and obtaining a fatigue response dataset, which includes strain time-series response samples centered on M+N weld bead control points.

[0009] In a possible implementation, under different welding conditions, residual stress and deformation are evaluated using the welding current dataset, and the following processing is performed: Based on the automotive parts to be welded, the thermal stress concentration areas are analyzed, and thermally sensitive areas are identified, including the weld fusion line area, the heat-affected zone transition zone, and the fixture constraint reaction point area; Based on the automotive parts to be welded, thermal deformation is identified, and key deformation points are located, including upper surface warping points, lateral shrinkage points, and angular deformation vertices; Through the thermally sensitive areas, thermal field uniformity analysis is performed with the key deformation points as the center to generate a thermal deformation distribution map, and combined with the welding current dataset, reinforcement learning optimization is performed with the goal of suppressing maximum thermal deformation.

[0010] In a possible implementation, reinforcement learning optimization is performed using the welding current dataset to suppress maximum thermal deformation, and the following processing is performed: a first welding posture and a second welding posture are set for the automotive parts to be welded, wherein the first welding posture is a standard welding posture in a flat welding position and the second welding posture is a standard welding posture in a vertical welding position; based on the first welding posture and the second welding posture, the arc initiation control point and arc termination control point corresponding to the welding robot are compared with the welding current dataset to perform weld bead cell search verification, thereby obtaining multiple combinations of thermal field distributions under the welding quality boundary.

[0011] In a possible implementation, the following processing is performed: the arc initiation control point includes the weld start end and the arc initiation transition section; the arc termination control point includes the weld end and the arc termination backfill section; using the arc initiation control point and the arc termination control point as boundaries, the continuity of the weld bead during different welding posture switching processes is calibrated, and the influence of welding posture switching on residual deformation is evaluated in conjunction with the thermal deformation distribution map, and boundary constraints are applied to the weld bead cell search.

[0012] In a possible implementation, the arc initiation control point and arc termination control point corresponding to the welding robot arm are compared with the welding current dataset to perform weld bead cell search verification, and the following processing is performed: Based on the arc initiation control point and arc termination control point, the welding robot arm servo system is integrated, and the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding gas flow direction are set; based on the first influence factor and the second influence factor, the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding gas flow direction are adaptively and iteratively adjusted to output the optimal combination of process parameters.

[0013] In a possible implementation, the following processing is performed: In the welding robot arm servo system, based on the single-pass welding path from the arc initiation control point to the arc termination control point and the interpass connection path, the influence of the dynamic welding path on the multiple thermal field distribution combinations is evaluated by the weld defect rate to obtain the first influence factor; the influence of thermal cycle changes on the multiple thermal field distribution combinations is evaluated by the temperature gradient characteristics under different welding current inputs to obtain the second influence factor.

[0014] This application also provides an intelligent welding system for automotive parts, comprising: an automotive parts welding chain generation module, used to determine the basic components of the automotive parts to be welded, including the base material, weld bevel form, fixture positioning points, welding electrodes, and connecting fasteners, and generate an automotive parts welding chain; a welding process introduction module, used to connect a welding robot arm based on the automotive parts welding chain, introduce a front welding process and a back welding process, and extract the welding current dataset of the welding robot arm, wherein the front welding process includes M front welding nodes and the back welding process includes N back welding nodes; a welding distribution feature analysis module, used to analyze the welding distribution features of the basic components based on the front welding process and the back welding process, wherein the welding distribution features include the thermal field symmetry distribution features corresponding to mirror-symmetric welds and the thermal field cross distribution features corresponding to cross-symmetric welds; and a welding defect reminder module, used to add the welding distribution features to finite element analysis software, and under different welding conditions, combine the welding current dataset to perform residual stress and deformation assessment, and provide welding defect reminders according to automotive welding quality standards and residual deformation assessment results.

[0015] This application proposes an intelligent welding method and system for automotive parts. First, it identifies the basic components of the automotive parts to be welded, including the base material, weld bevel type, fixture positioning points, welding electrodes, and connecting fasteners, generating a welding chain for the automotive parts. Then, based on this welding chain, a welding robotic arm is connected, introducing front-side welding and back-side welding processes. The welding current dataset of the welding robotic arm is extracted. The front-side welding process includes M front-side welding nodes, and the back-side welding process includes N back-side welding nodes. Next, based on the front-side and back-side welding processes, the welding distribution characteristics of the basic components are analyzed. These welding distribution characteristics include the symmetrical thermal field distribution characteristics corresponding to mirror-symmetric weld beads and the cross-symmetric thermal field distribution characteristics corresponding to cross-symmetric weld beads. Finally, the welding distribution characteristics are added to finite element analysis software, and residual stress and deformation are evaluated under different welding conditions, combined with the welding current dataset. Welding defect alerts are then issued according to automotive welding quality standards and the residual deformation evaluation results. Through the above process, the method and system proposed in this application achieve the technical effect of improving the accuracy of the matching between welding process parameters and the basic component characteristics of the parts, improving the quantitative assessment accuracy of welding stress and deformation, and thus reducing the welding defect rate. Attached Figure Description

[0016] To more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings of the embodiments of the present invention will be briefly described below. Flowcharts are used in this application to illustrate the operations performed by the system according to the embodiments of the present application. It should be understood that the preceding or following operations are not necessarily performed precisely in sequence. Instead, various steps can be processed in reverse order or simultaneously as needed. Furthermore, other operations can be added to these processes, or one or more steps can be removed from these processes.

[0017] Figure 1 This is a flowchart illustrating an intelligent welding method for automotive parts provided in an embodiment of this application.

[0018] Figure 2 This is a structural schematic diagram of an intelligent welding system for automotive parts provided in an embodiment of this application.

[0019] Figure labeling: Automotive parts welding chain generation module 10, welding process introduction module 20, welding distribution feature analysis module 30, welding defect reminder module 40. Detailed Implementation

[0020] To further illustrate the technical means and effects of the present invention in achieving its intended purpose, the following detailed description of the specific implementation methods, structures, features, and effects of the present invention, in conjunction with the accompanying drawings and preferred embodiments, is provided below.

[0021] This application provides an intelligent welding method for automotive parts, such as... Figure 1 As shown, the method includes: Step S100: Determine the base material, weld bevel type, fixture positioning point, welding electrode, and basic components of the automotive parts to be welded, and generate the automotive parts welding chain.

[0022] Specifically, a material spectrometer is used to analyze the material composition of the base material of the automotive parts to be welded, determining its grade and material properties. For example, the base material is identified as automotive 6061 aluminum alloy or automotive DP780 duplex high-strength steel. A bevel profile detector is used to inspect the size and shape of the weld bevel, verifying key parameters such as the bevel angle, blunt edge thickness, and bevel depth. For example, the bevel is determined to be a 60-degree V-shaped bevel with a blunt edge thickness of 2 mm and a bevel depth of 8 mm. A three-dimensional coordinate measuring machine is used to calibrate the three-dimensional coordinates of the fixture positioning points, such as the left front positioning point, right rear positioning point, and lower support point—three core fixture positioning points. A matching welding electrode is selected based on the base material and welding requirements. The welding electrode is the core component for conducting welding current and generating the arc during welding operations. For example, an ER50-6 solid welding wire electrode is used when the base material is high-strength steel, and an ER4043 aluminum-silicon welding wire electrode is used when the base material is aluminum alloy. Connecting fasteners are auxiliary fasteners used to secure the parts to be welded, such as locating pins, clamping blocks, and support seats. Following the general process logic for welding automotive parts, a welding chain is generated in the following sequence: beveling → fixture positioning and clamping → spot welding → continuous welding → weld cooling. Each node is marked with corresponding process requirements. For example, the beveling node is marked with a machining tolerance of ±0.1mm; the fixture positioning and clamping node is marked with a clamping force of 80kN; the spot welding and fixing node is marked with a spot welding current of 200A and a spot welding time of 0.3s; the continuous welding node is marked with a welding travel speed of 300mm / min; and the weld cooling node is marked with natural cooling to room temperature or water cooling to below 200℃.

[0023] Step S200: Based on the automotive component welding chain, connect the welding robot arm, introduce the front welding process and the back welding process, and extract the welding current dataset of the welding robot arm. The front welding process includes M front welding nodes, and the back welding process includes N back welding nodes.

[0024] Specifically, an industrial bus is used to connect the control system of the welding robot arm with the process control system of the automotive parts welding chain. For example, Profinet industrial Ethernet is used for the interface to achieve linkage between welding chain node commands and welding robot arm movements. Based on the weld distribution characteristics of the parts to be welded, front welding and back welding processes are extended after the continuous welding nodes of the welding chain. For example, for the butt weld of the crossbeam of an automotive frame, the front welding process is designed to include 8 front welding nodes, corresponding to 8 segmented welding paths of the front weld, and the back welding process includes 6 back welding nodes, corresponding to 6 segmented welding paths of the back weld. A current sensor is installed at the welding power source of the welding robot arm to collect real-time current values ​​throughout all stages, including no-load, arc initiation, welding travel, and arc termination, forming a welding current dataset. This dataset includes data such as the average current, current fluctuation range, peak current at arc initiation, and current trough at arc termination for each welding node. For example, the average welding current for the first welding node on the front is 180A, with a current fluctuation range of 175A to 185A, a peak current at arc initiation of 220A, and a trough current at arc termination of 150A; the average welding current for the third welding node on the back is 190A, with a current fluctuation range of 186A to 194A, a peak current at arc initiation of 230A, and a trough current at arc termination of 160A. The correlation between welding current and welding zone temperature is as follows: Under conventional welding process conditions, the higher the welding current, the higher the temperature of the welding zone is usually. The core logic is that during welding, the current generates Joule heat through the arc between the electrode and the base material. The heat calculation formula can be simplified to: Welding heat = (Welding current) 2 × Arc resistance × energizing time When parameters such as arc resistance and energizing time are basically stable, an increase in current directly leads to a significant increase in the heat generated by the arc, which in turn raises the temperature of the weld pool and heat-affected zone.

[0025] Step S300: Based on the front welding process and the back welding process, analyze the welding distribution characteristics of the base component. The welding distribution characteristics include the thermal field symmetry distribution characteristics corresponding to mirror-symmetric welds and the thermal field cross distribution characteristics corresponding to cross-symmetric welds.

[0026] Specifically, a 3D laser scanning device is used to create 3D models of the weld beads corresponding to the M welding nodes of the front welding process and the N welding nodes of the back welding process, obtaining the spatial position, length, width, direction, and relative positional relationship of all weld beads. An infrared thermal imager is used to measure the welding area in real time when the welding robot arm simulates the front and back welding processes, collecting data such as the temperature distribution, temperature diffusion direction, and temperature peak position of the heat field during the welding process of each welding node. Using thermal field data analysis software, the collected thermal field data and weld layout data are matched and analyzed. If the spatial positions of the front and back welds are mirror-symmetrical, for example, the front weld is horizontally distributed along the central axis of the component, and the back weld is also horizontally distributed along the central axis and coincides with the front weld, then it is determined to be a mirror-symmetrical weld. The corresponding thermal field shows a symmetrical distribution characteristic with equal temperature on the left and right sides and consistent temperature gradient on the top and bottom, with the central axis as the axis of symmetry. If the front and back welds are arranged in a cross pattern, for example, the front weld is distributed laterally along the component and the back weld is distributed longitudinally along the component, with the intersection of the welds being the core area of ​​the weld, then it is determined to be a cross-symmetrical weld. The corresponding thermal field shows a cross-distribution characteristic with temperature superposition and increase at the intersection and temperature gradually decrease along the extension direction of the weld. The quantitative data of the two types of characteristics, such as the temperature difference of the symmetrical thermal field and the temperature superposition coefficient of the cross thermal field, are organized into a standardized welding distribution characteristic dataset.

[0027] Step S400: In the finite element analysis software, add the welding distribution features, and under different welding conditions, combine the welding current dataset to evaluate residual stress and deformation. Based on the automotive welding quality standards and the residual deformation evaluation results, issue welding defect alerts.

[0028] Specifically, finite element analysis software, such as ANSYS, ABAQUS, and MSC.MARC, is selected to build a 3D solid model of the automotive parts to be welded, scaled to the same size. The model's mesh is refined, with small mesh sizes for the weld area and heat-affected zone, and large mesh sizes for the non-welded areas of the base material, to ensure simulation accuracy. The welding distribution characteristics obtained in step S300 are added to the model as thermal boundary conditions. For example, the symmetrical thermal field distribution characteristics of mirror-symmetric welds are converted into symmetrical temperature loading parameters on both sides of the weld in the model, and the cross-sectional thermal field distribution characteristics of intersecting symmetric welds are converted into temperature superposition loading parameters at the weld intersections in the model. The welding current dataset obtained in step S200 is imported into the model as a heat source input parameter. Combining the aforementioned Joule heating formula, the current values ​​of different welding nodes are converted into corresponding arc heat input values. Different welding conditions are set up, such as normal temperature and pressure, -10°C, 85% humidity, and different welding travel speeds. Welding simulation calculations are performed under each condition. During the simulation, data such as residual stress values, stress distribution locations, deformation magnitude, and deformation direction of the components are extracted in real time. The automotive welding quality standards are retrieved, and the residual stress obtained from the simulation is compared with the deformation assessment results and standard limits. If the residual stress exceeds 80% of the yield strength of the base material, or the deformation exceeds the allowable range of the component assembly tolerance, the corresponding welding defect warning is triggered. All warnings are marked with the specific defect location, defect type, and excess value.

[0029] In one possible implementation, under different welding conditions, residual stress and deformation are evaluated using the welding current dataset. Step S400 further includes step S410, which analyzes the thermal stress concentration areas and identifies heat-sensitive areas based on the automotive parts to be welded. These heat-sensitive areas include the weld fusion line area, the heat-affected zone transition zone, and the fixture constraint reaction point area. Specifically, in the constructed finite element simulation model, the heat input parameters corresponding to the welding current dataset and the thermal field parameters corresponding to the welding distribution characteristics are retrieved. The thermal stress during the welding process is calculated region by region. The thermal stress is calculated using the elastoplastic mechanics formula: thermal stress = elastic modulus × coefficient of linear expansion × temperature change, where the elastic modulus and coefficient of linear expansion are selected based on the base material. The calculated thermal stress distribution is visualized using the stress cloud diagram analysis function of the finite element software. The red highlighted areas in the stress cloud diagram represent the thermal stress concentration areas. Based on the definition of heat-sensitive areas, three types of heat-sensitive areas are identified in the stress cloud map. These include: the weld fusion line region, the boundary between the weld metal and the base metal, which is the area with the largest temperature gradient and the most likely stress concentration during welding; for example, a ring-shaped area with a weld edge width of 1mm to 2mm in the stress cloud map; the heat-affected zone transition zone, the transition part between the area of ​​the base metal affected by the welding thermal cycle and the unaffected base metal area, where both mechanical properties and metallographic structure are unstable; for example, a band-shaped area with a width of 3mm to 5mm outside the fusion line; and the fixture constraint reaction point region, the clamping node where the component is constrained by the clamping force, and the thermal deformation generated by welding creates a reverse constraint stress, easily leading to stress concentration; for example, a circular area with a radius of 5mm at the fixture positioning point. The three-dimensional coordinates, area range, and stress values ​​of the three types of heat-sensitive areas are compiled into a heat-sensitive area dataset, and the stress sensitivity level of each area is labeled.

[0030] Step S420: Based on the automotive parts to be welded, thermal deformation is identified, and key deformation points are located. These key deformation points include upper surface warping points, lateral shrinkage points, and angular deformation vertices. Specifically, in the finite element simulation model, the deformation analysis module is activated. Combined with the heat-sensitive area dataset from step S410, the three-dimensional deformation of the parts during welding is calculated in real time. The deformation is calculated using a displacement vector analysis method, that is, by comparing the changes in the three-dimensional coordinates of each point on the parts before and after welding, the displacement and direction of each point are obtained. The deformation distribution is visualized using the deformation cloud map analysis function of the finite element software. The red highlighted areas in the deformation cloud map are the core points with the largest deformation. Based on the definition of critical deformation points, three types of critical deformation points are located in the deformation cloud map. These include: **Upper surface warping points:** These are extreme points where the upper surface of the component warps upwards due to welding thermal expansion. These are the core points where the component's flatness exceeds tolerances. For example, a single point on the centerline of the weld on the upper surface of the component has an upward warping deformation of 0.3 mm. **Lateral contraction points:** These are extreme points where the sides of the weld contract towards the weld due to welding thermal contraction. These are the core points where the component's dimensional accuracy exceeds tolerances. For example, two symmetrically distributed points on both sides of the weld have a contraction deformation of 0.25 mm towards the weld. **Angular deformation vertices:** These are the vertices where the component deflects due to uneven distribution of the welding heat field. These are the core points where the component's form and position tolerances exceed tolerances. For example, two diagonally opposite points at the end of the component have an angular deflection of 0.5 degrees. The three-dimensional coordinates, deformation magnitude, and deformation direction of these three types of critical deformation points are compiled into a critical deformation point dataset, and the deformation influence weight of each point is labeled.

[0031] Step S430: Using the heat-sensitive area, perform thermal field uniformity analysis centered on the key deformation points to generate a thermal deformation distribution map. Combined with the welding current dataset, perform reinforcement learning optimization with the goal of suppressing maximum thermal deformation. Specifically, using each key deformation point located in step S420 as the center of a sphere, a thermal field analysis area is delineated with a preset radius. This area must completely cover the heat-sensitive area identified in step S410. The thermal field analysis module of finite element software is used to calculate the temperature distribution uniformity within each analysis area. The calculation index for temperature distribution uniformity is the ratio of the maximum temperature difference to the average temperature within the area; the smaller the ratio, the better the thermal field uniformity. Match the thermal field uniformity data and deformation data of all analysis areas to generate a thermal deformation distribution map. This map marks the location of each key deformation point, the magnitude of the deformation, the corresponding thermal field uniformity value, and the thermal stress distribution. Darker areas in the map represent poorer thermal field uniformity and greater deformation. A reinforcement learning optimization model aimed at suppressing maximum thermal deformation was constructed. The model's input layer contains all parameters from the welding current dataset, the hidden layer comprises a three-layer fully connected neural network, the activation function is ReLU, and the output layer contains the optimized welding current parameter combination and its corresponding maximum thermal deformation. The model is trained using a Markov decision process, treating each welding current parameter combination as a decision state and the maximum thermal deformation as the reward value; the smaller the deformation, the higher the reward value. Through iterative optimization, the optimal welding current parameter combination is obtained, effectively suppressing thermal deformation.

[0032] In one possible implementation, the welding current dataset is used to perform reinforcement learning optimization with the goal of suppressing maximum thermal deformation. Step S430 further includes step S431, setting a first welding posture and a second welding posture for the automotive component to be welded. The first welding posture is a standard welding posture in a flat welding position, and the second welding posture is a standard welding posture in a vertical welding position. Specifically, based on the structural characteristics and weld layout of the automotive component to be welded, two standardized welding posture parameters are preset in the control system of the welding robot arm. The first welding posture is a standard welding posture in a flat welding position, specifically referring to a welding posture where the weld of the component to be welded is in a horizontal position, and the welding torch of the welding robot arm is perpendicular to or at a 45° angle to the weld. This is the most basic and commonly used posture in automotive component welding. The second welding posture is a standard welding posture in a vertical welding position, specifically referring to a welding posture where the weld of the component to be welded is in a vertical position, and the welding torch of the welding robot arm is at a 90° or 60° angle to the weld. This is the core posture for vertical welds in automotive component welding. For example, the parameters for the first welding posture are set as follows: the component is placed on a horizontal fixture, the weld axis is parallel to the horizontal plane, the spatial angle of the welding torch is 45° with respect to the vertical direction of the weld and 30° with respect to the direction of weld travel, and the height of the welding torch is 10mm from the weld surface; the parameters for the second welding posture are set as follows: the component is fixed by a vertical fixture, the weld axis is perpendicular to the horizontal plane, the spatial angle of the welding torch is 90° with respect to the vertical direction of the weld and 15° with respect to the direction of weld travel, and the height of the welding torch is 8mm from the weld surface. The parameters for both welding postures are stored in the posture database of the welding robot arm, and the weld type and welding node corresponding to each posture are labeled. For example, the first welding posture corresponds to the bottom horizontal weld of the component, and the second welding posture corresponds to the side vertical weld of the component.

[0033] Step S432: Based on the first welding posture and the second welding posture, and referring to the arc initiation control point and arc termination control point corresponding to the welding robot arm, weld bead cell search verification is performed in conjunction with the welding current dataset to obtain multiple thermal field distribution combinations under the welding quality boundary constraints. The arc initiation control point includes the weld start end and the arc initiation transition section; the arc termination control point includes the weld end and the arc termination backfill section. Using the arc initiation control point and the arc termination control point as boundaries, the continuity of the weld bead during different welding posture switching processes is calibrated. The influence of welding posture switching on residual deformation is evaluated in conjunction with the thermal deformation distribution map, and boundary constraints are applied to the weld bead cell search. Specifically, the three-dimensional coordinate parameters of the arc initiation control point and the arc termination control point are retrieved from the welding robot arm's control system. The parameters of the arc initiation control point include the coordinates of the weld start end, the length of the arc initiation transition section, and the travel speed; the parameters of the arc termination control point include the coordinates of the weld end, the length of the arc termination backfill section, and the travel speed. The two welding posture parameters from step S431 are matched with the arc initiation and arc termination control point parameters to define the search boundary of the weld bead within the working range of the welding robot arm. The welding current dataset from step S200 is split into weld bead cells, where each weld bead cell is the smallest independent welding unit of the weld. Each weld bead cell corresponds to a set of independent current parameters. A grid search method is used to traverse and verify the current parameters of each cell, with the verification indicators being weld formation quality and residual deformation. During the search and verification process, for scenarios involving welding posture switching, such as switching from the first welding posture to the second welding posture, the movement path of the welding torch is calibrated during the switching process to ensure seamless connection between the arc termination control point of the previous weld and the arc initiation control point of the next weld, and that the weld bead width deviation at the connection meets the requirements. Based on the thermal deformation distribution map in step S430, the influence of the change in thermal field distribution after attitude switching on the residual deformation is evaluated. For example, if the deformation of the thermal field superposition area increases after attitude switching, boundary constraints are applied to the weld cell search in that area to reduce the fluctuation range of the current parameter. Finally, the thermal field distribution combination that meets the welding quality boundary limit is obtained through search verification. Each combination includes the corresponding welding attitude, current parameter, weld cell layout and residual deformation data.

[0034] In one possible implementation, by comparing the arc initiation control point and arc termination control point corresponding to the welding robot arm with the welding current dataset, weld bead cell search verification is performed. Step S432 further includes step S4321, which integrates the welding robot arm servo system based on the arc initiation control point and arc termination control point, and sets the welding torch oscillation frequency and amplitude, wire feed angle, and shielding gas flow direction. Specifically, the servo motor, servo driver, position sensor of the welding robot arm are integrated with the control system of the arc initiation / arc termination control point, and a servo bus is used to complete the communication connection. Based on the welding process requirements of the arc initiation and arc termination control points, the core parameters of the welding torch are initially set. The torch oscillation frequency is the number of times the torch oscillates left and right during welding, for example, 3 times per second. The torch oscillation amplitude is the maximum distance the torch oscillates left and right, for example, 4 mm. The wire feed angle is the angle between the welding wire and the weld when the wire is fed from the wire feed nozzle, for example, 15°. The shielding gas flow direction is the flow direction of the shielding gas when it is ejected from the torch, for example, 30° inclined along the weld travel direction. The shielding gas flow rate is set to 18 L / min. The initial parameters are verified through the trial operation function of the welding robot arm. During the trial operation, data such as the torch oscillation trajectory, wire feed stability, and shielding gas flow coverage are collected to ensure that the torch oscillates smoothly, feeds wire continuously, and the shielding gas flow does not deviate throughout the entire welding path from arc initiation to arc termination. Finally, the verified initial parameters are stored in the parameter database of the servo system as the benchmark values ​​for adaptive iterative adjustment.

[0035] Step S4322: Based on the first and second influence factors, adaptive iterative adjustments are made to the welding torch oscillation frequency and amplitude, wire feed angle, and shielding gas flow direction to output the optimal combination of process parameters. Specifically, in the welding robotic arm servo system, based on the single-pass welding path from the arc initiation control point to the arc termination control point, and the inter-pass connection path, the influence of the dynamic welding path on the multiple thermal field distribution combinations is evaluated using the weld defect rate to obtain the first influence factor. Furthermore, the influence of thermal cycle changes on the multiple thermal field distribution combinations is evaluated using the temperature gradient characteristics under different welding current inputs to obtain the second influence factor. Specifically, a two-factor evaluation model is established. The first step evaluates the first influence factor, which is the influence coefficient of the dynamic welding path on the thermal field distribution combination. This is quantified using the weld defect rate. The dynamic welding path includes a single-pass welding path and an inter-pass connection path. The single-pass welding path is from the arc initiation control point to the arc termination control point, and the inter-pass connection path is from the previous arc termination control point to the next arc initiation control point. In the welding robotic arm servo system, all parameters of the single-pass welding path and the inter-pass connection path are retrieved. High-speed camera equipment acquires weld formation images during the welding process, and defects are identified in the images. Defect types include porosity, slag inclusions, lack of fusion, and undercut. The weld defect rate for each path is calculated, and the formula for the defect rate is the ratio of the number of defects to the total weld length. The defect rate is converted into a first influencing factor; the higher the defect rate, the larger the factor value. The second step evaluates the second influencing factor, which is the influence coefficient of thermal cycling changes on the thermal field distribution combination. This is quantified using temperature gradient characteristics, which represent the temperature variation of the welding area over time and space under different welding current inputs. The welding current dataset from step S200 is retrieved, and the temperature gradient under different current inputs is calculated. The formula for the temperature gradient is the temperature change per unit distance. The temperature gradient characteristics are converted into a second influencing factor; the larger the temperature gradient, the larger the factor value. An adaptive iterative adjustment algorithm was constructed. The algorithm's inputs were the first and second influencing factors and the initial parameters of the welding torch. The iteration rules were as follows: when the first influencing factor was greater than a first preset influencing factor threshold, the welding torch oscillation amplitude and oscillation frequency were reduced; when the second influencing factor was greater than a second preset influencing factor threshold, the welding current and wire feed angle were reduced. The iteration terminated when the weld defect rate was lower than a preset defect rate threshold and the temperature gradient was lower than a preset gradient threshold. After iterative adjustment, the optimal combination of process parameters was output, which met the welding quality requirements.

[0036] In one possible implementation, welding defect alerts are issued according to automotive welding quality standards and residual deformation assessment results. Step S400 further includes step S440, setting impact velocity and impact load conditions to obtain a weld micro-wear dataset. This involves connecting an image recognition device to collect the weld micro-wear dataset, which includes grain morphology image samples centered on M+N weld bead control points. Specifically, a pendulum impact testing machine is selected as the testing equipment to conduct impact tests on the finished welded parts of the automotive components to be welded. Standardized impact parameters are set before the test: the impact velocity is the speed at which the impact head contacts the weld during the impact test, and the impact load is the force applied to the weld by the impact head. An image recognition device consisting of an industrial-grade metallographic microscope and a high-speed camera is installed on the impact testing machine to collect the weld micro-morphology. The focal center of the image acquisition is the M front weld bead control points and the N back weld bead control points. Impact tests were conducted sequentially according to preset impact parameters. After each test, metallographic samples were prepared from the weld area, including grinding, polishing, and etching. Grain morphology images were collected using image recognition equipment within a preset radius centered on each weld control point. Each sample contained microscopic feature data such as grain size, grain morphology, and wear marks in the weld area. All image samples and corresponding impact parameters and wear data were compiled into a weld micro-wear dataset.

[0037] Step S450: Based on the weld micro-wear dataset, determine the width of the heat-affected zone (HAZ) and the grain coarsening index. Specifically, metallographic image analysis software is used to process all image samples in the weld micro-wear dataset. The first step is to measure the HAZ width, identifying the boundary between the weld fusion line and the original microstructure of the base material in the image. The distance between the two boundary lines is measured using the software's length measurement tool. Multiple measurements are performed on samples at each weld control point, and the average value is taken. The second step is to calculate the grain coarsening index. A representative area of ​​the HAZ is selected in the image, and the number of grains and the total grain area within that area are statistically analyzed. The average grain size is calculated, and then the average size of the original grains in the base material is retrieved. The formula for calculating the grain coarsening index is the average grain size of the HAZ divided by the original grain size of the base material. The HAZ width and grain coarsening index of all weld control points are statistically analyzed to obtain a corresponding numerical distribution table, and the acceptable threshold for each parameter is marked.

[0038] Step S460: Based on the width of the heat-affected zone and the grain coarsening index, determine the first constraint feature. This first constraint feature is used to improve the accuracy of residual deformation assessment. Specifically, the first constraint feature is a mechanical property constraint parameter established based on the width of the heat-affected zone and the grain coarsening index. Its core function is to correlate the performance degradation of the weld microstructure with the macroscopic residual deformation, filling the gap in the assessment model that relies solely on thermal field and stress analysis, and achieving a joint assessment of macroscopic deformation and microscopic properties. A correlation model is established between the width of the heat-affected zone, the grain coarsening index, and the mechanical properties of the weld. For example, the correlation rule is: for every 0.1 mm increase in the width of the heat-affected zone, the yield strength of the weld decreases by 2 MPa; for every 0.1 increase in the grain coarsening index, the tensile strength of the weld decreases by 3 MPa. The correlation rule is obtained by fitting experimental data of automotive welded parts. This correlation model is then transformed into specific parameters of the first constraint feature, in the form of material property correction coefficients in the finite element simulation model. Add the first constraint feature to the finite element simulation model in step S400, apply the correction coefficient to the heat-affected zone and weld area of ​​the model, and re-execute the residual stress and deformation evaluation calculation. The accuracy of the corrected evaluation will be significantly improved.

[0039] In one possible implementation, step S400 further includes step S470, setting alternating load frequency and amplitude conditions, and obtaining a fatigue response dataset. This involves using a strain gauge array to monitor the strain response process of the weld area of ​​the automotive component in different impact directions, thereby obtaining the fatigue response dataset. The fatigue response dataset includes strain time-series response samples centered on M+N weld bead control points. Specifically, an electro-hydraulic servo fatigue testing machine is used as the testing equipment to conduct fatigue tests on the finished welded parts of the automotive component to be welded. Before the test, standardized alternating load parameters are set, where the alternating load frequency is the number of periodic load changes, and the alternating load amplitude is the difference between the maximum and minimum load values. A strain gauge array is attached to the weld area. The strain gauge array is a sensing array composed of multiple strain gauges used to monitor the strain changes in the weld area in real time. The number of strain gauges is the sum of the M weld bead control points on the front and the N weld bead control points on the back, with each strain gauge attached at a position corresponding to a weld bead control point. Fatigue tests were conducted sequentially according to preset alternating load parameters. Each test lasted until microcracks appeared in the weld or the preset number of cycles was reached. During the test, the strain values ​​of each strain gauge were collected in real time to form a strain-time response curve. All strain curves, strain peak values, and fatigue life data were compiled into a fatigue response dataset.

[0040] Step S480: Based on the fatigue response dataset, evaluate the fatigue resistance of the weld and determine the second constraint feature. The second constraint feature is used to improve the accuracy of the correlation assessment between residual deformation and fatigue failure, thereby improving the reliability of automotive welding quality judgment. Specifically, the second constraint feature is a failure correlation constraint parameter established based on the weld fatigue resistance. Its core function is to correlate macroscopic residual deformation with microscopic fatigue failure risk, so that welding quality judgment considers not only whether the deformation amount meets the standard, but also whether the risk of fatigue failure is controllable. The first step is to evaluate the weld fatigue resistance, which is the weld's ability to resist fatigue crack initiation and propagation under alternating loads. The fatigue life of each weld control point is calculated using the strain-time response curve. The fatigue life calculation formula is the number of cycles based on the Miner cumulative damage criterion. The second step is to calculate the weld fatigue strength, which is the maximum load amplitude at which the weld does not produce cracks under 1 million cycles. A correlation model between residual deformation and fatigue failure is established. For example, the correlation rule is: for every 0.01 mm increase in residual deformation of a component, the fatigue life of the weld decreases by 1%, and the fatigue strength decreases by 2 MPa. The correlation rule is obtained by fitting fatigue test data of automotive welded parts. This correlation model is transformed into specific parameters of the second constraint feature, which are in the form of fatigue failure correction coefficients in the finite element simulation model. The second constraint feature is added to the finite element simulation model in step S400, and works together with the first constraint feature to re-execute the joint evaluation of residual stress, deformation, and fatigue failure. The final corrected evaluation model can simultaneously output the residual deformation of the component, the microstructure properties of the weld, and the fatigue resistance of the weld. The judgment result can accurately reflect the actual service performance of the welded part and effectively avoid the quality hazard of ignoring the risk of fatigue failure due to only considering the deformation.

[0041] This application's embodiments employ basic components such as determining the base material and weld bevel form of the parts to be welded and generating a welding chain. Based on this, a welding robotic arm is connected, positive and negative welding processes are introduced, and welding current datasets are extracted. Welding distribution characteristics such as the symmetry and cross distribution of the thermal field corresponding to the positive and negative processes are analyzed. These characteristics are added to the finite element analysis software. Residual stress and deformation are evaluated under different working conditions in conjunction with the current dataset. Welding defect alerts are provided based on automotive welding quality standards and evaluation results. These technical means solve the technical problems of insufficient compatibility between welding process parameters and component characteristics, and low accuracy of quantitative evaluation of welding stress and deformation in existing automotive parts welding operations. This achieves the technical effect of improving the accuracy of the compatibility between welding process parameters and the basic component characteristics of parts, improving the accuracy of quantitative evaluation of welding stress and deformation, and thus reducing the welding defect rate.

[0042] In the above text, refer to Figure 1 A smart welding method for automotive parts according to an embodiment of the present invention is described in detail. Next, reference will be made to... Figure 2 A smart welding system for automotive parts is described according to an embodiment of the present invention.

[0043] An intelligent welding system for automotive parts according to an embodiment of the present invention addresses the technical problems of insufficient compatibility between welding process parameters and component characteristics, and low accuracy in the quantitative assessment of welding stress and deformation in existing automotive part welding operations. The system aims to improve the accuracy of the compatibility between welding process parameters and the basic component characteristics of the parts, enhance the quantitative assessment accuracy of welding stress and deformation, and thereby reduce the welding defect rate. The intelligent welding system for automotive parts includes: an automotive part welding chain generation module 10, a welding process introduction module 20, a welding distribution characteristic analysis module 30, and a welding defect alert module 40.

[0044] The automotive component welding chain generation module 10 is used to determine the basic components of the automotive component to be welded, including the base material, weld bevel form, fixture positioning points, welding electrodes, and connecting fasteners, and generate an automotive component welding chain. The welding process introduction module 20 is used to connect a welding robot arm based on the automotive component welding chain, introduce a front welding process and a back welding process, and extract the welding current dataset of the welding robot arm. The front welding process includes M front welding nodes, and the back welding process includes N back welding nodes. The welding distribution feature analysis module 30 is used to analyze the welding distribution features of the basic components based on the front welding process and the back welding process. The welding distribution features include the thermal field symmetry distribution features corresponding to mirror-symmetric welds and the thermal field cross distribution features corresponding to cross-symmetric welds. The welding defect reminder module 40 is used to add the welding distribution features to the finite element analysis software, and under different welding conditions, combine the welding current dataset to perform residual stress and deformation assessment, and provide welding defect reminders according to automotive welding quality standards and residual deformation assessment results.

[0045] The specific configuration of the welding defect alert module 40 is described in detail below: As mentioned above, welding defect alerts are issued according to automotive welding quality standards and residual deformation assessment results. The welding defect alert module 40 may further include: a weld micro-wear dataset acquisition unit for setting impact speed and impact load conditions to acquire a weld micro-wear dataset; a heat-affected zone width determination unit for determining the heat-affected zone width and grain coarsening index based on the weld micro-wear dataset; and a first constraint feature determination unit for determining a first constraint feature based on the heat-affected zone width and grain coarsening index, wherein the first constraint feature is used to improve the accuracy of residual deformation assessment.

[0046] The welding defect alert module 40 may further include: a fatigue response dataset acquisition unit for setting alternating load frequency and amplitude conditions to acquire a fatigue response dataset; and a second constraint feature determination unit for evaluating the fatigue resistance of the weld and determining a second constraint feature based on the fatigue response dataset. The second constraint feature is used to improve the accuracy of the correlation assessment between residual deformation and fatigue failure, thereby improving the reliability of automotive welding quality judgment.

[0047] The weld micro wear dataset acquisition unit may further include: connecting to an image recognition device to collect a weld micro wear dataset, wherein the weld micro wear dataset includes grain morphology image samples centered on M+N weld bead control points; The fatigue response dataset acquisition unit may further include: using a strain gauge array to monitor the strain response process of the weld area of ​​automotive parts in different impact directions, and acquiring a fatigue response dataset, wherein the fatigue response dataset includes strain time-series response samples with M+N weld control points as the focus.

[0048] In this module, under different welding conditions, residual stress and deformation are evaluated using the welding current dataset. The welding defect alert module 40 may further include: a heat-sensitive area identification unit for analyzing the thermal stress concentration area and identifying heat-sensitive areas based on the automotive parts to be welded, including weld fusion line areas, heat-affected zone transition zones, and fixture constraint reaction point areas; a key deformation point positioning unit for identifying thermal deformation and locating key deformation points based on the automotive parts to be welded, including upper surface warping points, lateral shrinkage points, and angular deformation vertices; and a thermal deformation distribution map generation unit for performing thermal field uniformity analysis centered on the key deformation points within the heat-sensitive areas, generating a thermal deformation distribution map, and combining it with the welding current dataset to perform reinforcement learning optimization with the goal of suppressing maximum thermal deformation.

[0049] Specifically, the heat deformation distribution map generation unit, in conjunction with the welding current dataset, performs reinforcement learning optimization with the goal of suppressing maximum thermal deformation. The unit may further include: a welding posture setting subunit for setting the first and second welding postures of the automotive component to be welded, wherein the first welding posture is a standard welding posture in a flat welding position, and the second welding posture is a standard welding posture in a vertical welding position; and a weld bead cell search and verification subunit for performing weld bead cell search and verification based on the first and second welding postures, referring to the arc initiation and arc termination control points corresponding to the welding robot arm, and combining the welding current dataset to obtain multiple thermal field distribution combinations under the welding quality boundary constraints.

[0050] The weld bead cell search verification subunit may further include: the arc initiation control point includes the weld start end and the arc initiation transition section; the arc termination control point includes the weld end and the arc termination backfill section; using the arc initiation control point and the arc termination control point as boundaries, the continuity of the weld bead during different welding posture switching processes is calibrated, and the influence of welding posture switching on residual deformation is evaluated in conjunction with the thermal deformation distribution map, thereby constraining the weld bead cell search.

[0051] Specifically, the weld bead cell search verification is performed by comparing the arc initiation control point and arc termination control point corresponding to the welding robot arm with the welding current dataset. The weld bead cell search verification subunit may further include: a welding torch parameter setting component for integrating the welding robot arm servo system based on the arc initiation control point and arc termination control point, and setting the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding gas flow direction; and an adaptive iterative adjustment component for adaptively iteratively adjusting the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding gas flow direction based on a first influence factor and a second influence factor, and outputting the optimal combination of process parameters.

[0052] The adaptive iterative adjustment component may further include: a first influence factor acquisition subcomponent used in the welding robot arm servo system to evaluate the influence of the dynamic welding path on the multiple combinations of thermal field distributions based on the single-pass welding path from the arc initiation control point to the arc termination control point and the interpass connection path, and to acquire the first influence factor; and a second influence factor acquisition subcomponent used to evaluate the influence of thermal cycle changes on the multiple combinations of thermal field distributions based on the temperature gradient characteristics under different welding current inputs, and to acquire the second influence factor.

[0053] The intelligent welding system for automotive parts provided in this embodiment of the invention can execute the intelligent welding method for automotive parts provided in any embodiment of the invention, and has the corresponding functional modules and beneficial effects of the method.

[0054] Although this application makes various references to certain modules in the system according to the embodiments of this application, any number of different modules can be used and run on user terminals and / or servers. The various units and modules included are only divided according to functional logic, but are not limited to the above division, as long as the corresponding functions can be achieved; in addition, the specific names of each functional unit are only for easy distinction between each other and are not used to limit the scope of protection of this invention.

[0055] The above description is merely a preferred embodiment of the present invention and is not intended to limit the present invention in any way. Although the present invention has been disclosed above with reference to preferred embodiments, it is not intended to limit the present invention. Any person skilled in the art can make some modifications or alterations to the above-disclosed technical content to create equivalent embodiments without departing from the scope of the present invention. Any modifications, equivalent changes, and alterations made to the above embodiments based on the technical essence of the present invention without departing from the scope of the present invention shall still fall within the scope of the present invention.

Claims

1. A smart welding method for automotive parts, characterized in that, The method includes: The basic components of the automotive parts to be welded, including the base material, weld bevel type, fixture positioning points, welding electrodes, and connecting fasteners, are determined to generate the automotive part welding chain. Based on the automotive component welding chain, a welding robot arm is connected, a front welding process and a back welding process are introduced, and the welding current dataset of the welding robot arm is extracted. The front welding process includes M front welding nodes, and the back welding process includes N back welding nodes. Based on the front welding process and the back welding process, the welding distribution characteristics of the basic component are analyzed. The welding distribution characteristics include the thermal field symmetry distribution characteristics corresponding to mirror symmetric welds and the thermal field cross distribution characteristics corresponding to cross symmetric welds. In the finite element analysis software, the welding distribution features are added, and under different welding conditions, residual stress and deformation are evaluated in conjunction with the welding current dataset. Welding defects are alerted according to the automotive welding quality standards and residual deformation evaluation results. Among them, the method for evaluating residual stress and deformation under different welding conditions, combined with the welding current dataset, includes: Based on the automotive parts to be welded, the areas of thermal stress concentration are analyzed and the heat-sensitive areas are identified. The heat-sensitive areas include the weld fusion line area, the heat-affected zone transition zone, and the area of ​​the fixture constraint reaction point. Based on the automotive parts to be welded, thermal deformation is identified, and key deformation points are located. The key deformation points include upper surface warping points, lateral shrinkage points, and angular deformation vertices. Through the heat-sensitive area, thermal field uniformity analysis is performed with the key deformation point as the center to generate a thermal deformation distribution map. Combined with the welding current dataset, reinforcement learning optimization is performed with the goal of suppressing maximum thermal deformation. The method, which combines the welding current dataset and optimizes the process by suppressing maximum thermal deformation through reinforcement learning, includes: The first welding posture and the second welding posture of the automotive parts to be welded are set, wherein the first welding posture is the standard welding posture in the flat welding position and the second welding posture is the standard welding posture in the vertical welding position. Based on the first welding posture and the second welding posture, and by referring to the arc start control point and arc end control point of the welding robot arm, and combining the welding current dataset, we can perform weld cell search and verification to obtain multiple thermal field distribution combinations under the welding quality boundary. The arc initiation control point includes the weld initiation end and the arc initiation transition section; the arc termination control point includes the weld termination end and the arc termination backfill section. Using the arc initiation control point and the arc termination control point as boundaries, the continuity of the weld bead during different welding posture switching processes is calibrated. The influence of welding posture switching on residual deformation is evaluated in conjunction with the thermal deformation distribution map, and boundary constraints are applied to the weld bead cell search.

2. The intelligent welding method for automotive parts as described in claim 1, characterized in that, Based on automotive welding quality standards and residual deformation assessment results, the method for issuing welding defect alerts also includes: Set the impact velocity and impact load conditions to obtain a dataset of weld micro-wear. Based on the aforementioned weld micro-wear dataset, the width of the heat-affected zone and the grain coarsening index are determined. Based on the width of the heat-affected zone and the grain coarsening index, a first constraint feature is determined, which is used to improve the accuracy of residual deformation assessment.

3. The intelligent welding method for automotive parts as described in claim 2, characterized in that, The method includes: Set the alternating load frequency and amplitude conditions to obtain the fatigue response dataset; Based on the fatigue response dataset, the fatigue resistance of the weld is evaluated, and a second constraint feature is determined. The second constraint feature is used to improve the accuracy of the correlation assessment between residual deformation and fatigue failure, thereby improving the reliability of automotive welding quality judgment.

4. The intelligent welding method for automotive parts as described in claim 3, characterized in that, The method further includes: Connect an image recognition device to collect a weld micro wear dataset, which includes grain morphology image samples centered on M+N weld bead control points; A strain gauge array is used to monitor the strain response process of the weld area of ​​automotive parts under different impact directions, and to obtain a fatigue response dataset. The fatigue response dataset includes strain time-series response samples with M+N weld control points as the focus.

5. The intelligent welding method for automotive parts as described in claim 1, characterized in that, The method further includes: comparing the arc initiation control point and arc termination control point corresponding to the welding robot arm, and performing weld bead cell search verification in conjunction with the welding current dataset; Based on the arc initiation control point and arc termination control point, an integrated welding robot arm servo system is established, and the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding gas flow direction are set. Based on the first and second influence factors, the welding torch oscillation frequency and amplitude, wire feeding angle, and shielding airflow direction are adaptively and iteratively adjusted to output the optimal combination of process parameters.

6. The intelligent welding method for automotive parts as described in claim 5, characterized in that, The method includes: In the welding robotic arm servo system, based on the single-pass welding path from the arc initiation control point to the arc termination control point, and the interpass connection path, the influence of the dynamic welding path on the multiple thermal field distribution combinations is evaluated by the weld defect rate, and the first influence factor is obtained. By evaluating the temperature gradient characteristics under different welding current inputs, the influence of thermal cycling changes on the combination of multiple thermal field distributions is assessed, and the second influencing factor is obtained.

7. An intelligent welding system for automotive parts, characterized in that, The system is used to implement the intelligent welding method for automotive parts according to any one of claims 1-6, the system comprising: The automotive parts welding chain generation module is used to determine the basic components of the automotive parts to be welded, including the base material, weld bevel type, fixture positioning point, welding electrode, and connecting fastener, and to generate the automotive parts welding chain. The welding process introduction module is used to connect the welding robot arm based on the automotive component welding chain, introduce the front welding process and the back welding process, and extract the welding current dataset of the welding robot arm. The front welding process includes M front welding nodes and the back welding process includes N back welding nodes. The welding distribution feature analysis module is used to analyze the welding distribution features of the base component based on the front welding process and the back welding process. The welding distribution features include the thermal field symmetry distribution features corresponding to mirror-symmetric welds and the thermal field cross distribution features corresponding to cross-symmetric welds. The welding defect alert module is used to add the welding distribution features to the finite element analysis software, and to evaluate residual stress and deformation under different welding conditions by combining the welding current dataset, and to issue welding defect alerts according to automotive welding quality standards and residual deformation evaluation results.

Images