A metal weld automatic tracking welding system based on machine vision

Abstract

The present application relates to the technical field of metal welding automation control, in particular to a metal welding seam automatic tracking welding system based on machine vision. It comprises: a data acquisition module for acquiring initial three-dimensional geometric parameters of the welding seam and monitoring welding process parameters including welding current, arc voltage and welding speed; a thermal deformation prediction module for calculating the welding seam thermal deformation displacement vector; a dynamic deviation evaluation module for calculating the dynamic deviation index; a risk classification module for determining the track deviation risk as safe, first-level risk or second-level risk; a defect tendency prediction module for calculating the incomplete fusion index and the hot crack index; a comprehensive risk evaluation module for determining the comprehensive process risk factor; and a trajectory correction module for generating a corrected welding gun target trajectory to perform closed-loop correction control on the welding process. The present application solves the problem of track deviation caused by thermal hysteresis and dynamic deformation, significantly improving the tracking accuracy and stability of the welding path.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of metal welding automation control, in particular to a metal welding seam automatic tracking welding system based on machine vision. BACKGROUND

[0002] In the field of metal welding, the high heat input generated by the electric arc is the core of the process, but it also causes unavoidable physical phenomena; specifically, the local and concentrated high heat during welding will cause the workpiece to produce thermal deformation, so that the actual position of the weld seam dynamically deviates from the pre-set trajectory path during welding; traditional automatic welding systems usually rely on fixed programmed trajectories, and lack the ability to perceive and adapt to such real-time dynamic changes;

[0003] This trajectory deviation caused by thermal deformation is the main reason for serious welding defects such as incomplete fusion and burn-through, directly affecting the forming quality of the weld and the structural strength of the joint, and thus reducing the use performance and service life of the final product; therefore, how to accurately predict and prospectively compensate for thermal deformation during welding to achieve dynamic correction of the welding gun trajectory to ensure that it always accurately tracks the actual position of the weld seam has become a key technical problem to be solved in the field of improving the quality and reliability of automatic welding.

[0004] The above information disclosed in the above background section is only used to strengthen the understanding of the background of the present disclosure, and therefore it can include information that does not constitute prior art known to those of ordinary skill in the art. SUMMARY

[0005] To solve the above technical problems, the present application discloses a metal welding seam automatic tracking welding system based on machine vision, specifically, the technical solution of the present application comprises:

[0006] A data acquisition module for acquiring initial three-dimensional geometric parameters of the weld seam and monitoring welding process parameters including welding current, arc voltage and welding speed;

[0007] A thermal deformation prediction module for calculating the weld seam thermal deformation displacement vector based on the initial three-dimensional geometric parameters of the weld seam and the welding process parameters;

[0008] A dynamic deviation evaluation module for calculating a dynamic deviation index based on the thermal deformation displacement vector and the preset maximum allowed trajectory deviation;

[0009] A risk classification module for determining the trajectory deviation risk as safe, first-level risk or second-level risk according to the dynamic deviation index;

[0010] A defect tendency prediction module for calculating an incomplete fusion index and a hot crack index based on the dynamic deviation index and the welding process parameters;

[0011] a comprehensive risk assessment module configured to determine a comprehensive process risk factor based on the incomplete fusion index and the hot crack index;

[0012] a trajectory correction module configured to generate a corrected welding torch target trajectory for closed-loop correction control of the welding process in combination with the hot deformation displacement vector and the comprehensive process risk factor.

[0013] Optionally, the hot deformation prediction module calculates the hot deformation displacement vector by constructing a physical-geometric coupling prediction model, which combines welding process parameters, a preset geometric influence tensor, a material linear expansion coefficient, an arc heat efficiency coefficient, a material density, a material specific heat capacity, and an effective cross-sectional area of a heat-affected zone.

[0014] Optionally, the dynamic deviation evaluation module calculates a dynamic deviation index, including:

[0015] dividing the module length of the hot deformation displacement vector by a preset maximum allowed trajectory deviation to generate the dynamic deviation index.

[0016] Optionally, the risk grading module determines the trajectory deviation risk, including:

[0017] when the dynamic deviation index is less than or equal to a preset safety threshold, determining that it is safe;

[0018] when the dynamic deviation index is greater than the safety threshold and less than or equal to a preset first-level risk threshold, determining that it is a first-level risk;

[0019] when the dynamic deviation index is greater than the first-level risk threshold, determining that it is a second-level risk.

[0020] Optionally, the defect tendency prediction module calculates the incomplete fusion index, including:

[0021] weighting and summing a geometric alignment deviation represented by the dynamic deviation index and an energy penetration capacity represented by the current actual welding line energy and the reference optimal line energy to determine the incomplete fusion index.

[0022] Optionally, the defect tendency prediction module calculates the hot crack index, including:

[0023] estimating a hot stress level and a cooling speed level based on welding process parameters;

[0024] weighting and summing the hot stress level and the cooling speed level to determine the hot crack index.

[0025] Optionally, the comprehensive risk assessment module determines the comprehensive process risk factor, including:

[0026] weighting and summing the incomplete fusion index and the hot crack index to generate the comprehensive process risk factor.

[0027] Optionally, the trajectory correction module generates a corrected welding torch target trajectory, comprising:

[0028] Based on the comprehensive process risk factor, a compensation amplification coefficient is generated;

[0029] The visually perceived geometric position is superimposed with the thermal deformation displacement vector modulated via the compensation amplification coefficient to determine the corrected welding torch target trajectory.

[0030] Optionally, the trajectory correction module is further used for:

[0031] In response to the trajectory deviation risk being determined as a primary risk, the welding current or arc voltage is increased while the corrected welding torch target trajectory is executed;

[0032] In response to the trajectory deviation risk being determined as a secondary risk, the welding speed is reduced while the corrected welding torch target trajectory is executed.

[0033] Compared with the prior art, the present application has the following beneficial effects:

[0034] 1、The present application can calculate the welding thermal deformation displacement vector in real time and prospectively by constructing a physical-geometric coupling prediction model. This changes the mode of the traditional visual system that can only passively track the current visible weld, realizes the pre-compensation closed-loop control of thermal deformation, and fundamentally solves the problem of trajectory deviation caused by thermal hysteresis and dynamic deformation, significantly improving the tracking accuracy and stability of the welding path.

[0035] 2、The present application establishes a multi-dimensional and prospective risk assessment system from geometric deviation to metallurgical defects. It not only quantifies the risk of trajectory deviation, but also further predicts the tendency of internal defects such as incomplete fusion and hot cracking, and integrates multiple risks into a single comprehensive risk factor. This method enables the system to go beyond simple geometric alignment and achieve deep and preventive control of welding quality.

[0036] 3、The present application proposes a dynamic trajectory correction strategy driven by a comprehensive risk factor. The correction amount is no longer a fixed geometric compensation, but is dynamically amplified according to the real-time evaluation of the comprehensive defect risk. When the predicted defect risk increases, the system will perform a larger prospective compensation, thereby realizing intelligent and adaptive adjustment of control strength and ensuring that optimal avoidance measures can be taken under different risk levels.

[0037] 4、The present application designs a hierarchical response mechanism that coordinates trajectory correction and process parameter adjustment. According to the risk level of trajectory deviation, the system not only corrects the spatial position of the welding torch, but also synchronously optimizes the core process parameters such as welding current, voltage or speed. This comprehensive approach combining geometric control and physical process adjustment realizes three-dimensional, multi-channel intervention of the welding process, and more efficiently guarantees the joint quality under complex working conditions. BRIEF DESCRIPTION OF DRAWINGS

[0038] The present application will be further explained in connection with the accompanying drawings and embodiments:

[0039] Figure 1 is a system structure diagram of the present application. DETAILED DESCRIPTION

[0040] In order to make the purpose, technical scheme and advantages of the present application more clear and explicit, the present application will be further explained in detail below in connection with specific embodiments. Embodiment 1:

[0041] Please refer to Figure 1 A metal weld seam automatic tracking welding system based on machine vision, comprising:

[0042] A data acquisition module for acquiring initial three-dimensional geometric parameters of the weld seam and monitoring welding process parameters including welding current, arc voltage and welding speed;

[0043] A thermal deformation prediction module for calculating weld thermal deformation displacement vectors based on the initial three-dimensional geometric parameters of the weld seam and the welding process parameters;

[0044] A dynamic deviation evaluation module for calculating a dynamic deviation index based on the thermal deformation displacement vectors and a preset maximum allowed trajectory deviation;

[0045] A risk classification module for determining the trajectory deviation risk as safe, first-level risk or second-level risk according to the dynamic deviation index;

[0046] A defect tendency prediction module for calculating an incomplete fusion index and a hot crack index based on the dynamic deviation index and the welding process parameters;

[0047] A comprehensive risk evaluation module for determining a comprehensive process risk factor based on the incomplete fusion index and the hot crack index;

[0048] A trajectory correction module for generating a corrected welding torch target trajectory to perform closed-loop correction control of the welding process in combination with the thermal deformation displacement vectors and the comprehensive process risk factor;

[0049] The embodiment provides a metal weld automatic tracking welding system based on machine vision, which is used for solving the problem of weld track deviation caused by thermal deformation by accurately predicting and prospectively compensating the thermal deformation in the welding process, thereby improving the welding quality and joint service performance; the system provides real-time and accurate initial state and process data for subsequent prediction and control through a data acquisition module; before the welding task starts, the module scans and acquires the initial three-dimensional geometric parameters of the weld of the workpiece to be welded through a machine vision system composed of one or more structured light cameras, and the parameters constitute the geometric reference of initial track path planning; after the welding starts, the module continuously monitors the core welding process parameters, mainly including welding current , arc voltage and welding speed , through real-time communication with the welding power source and the motion controller.

[0050] The system uses a thermal deformation prediction module to calculate the dynamic deviation of the weld geometric position caused by heat input based on real-time input; the module receives the initial three-dimensional geometric parameters of the weld from the data acquisition module and real-time welding process parameters , , , and through an embedded physical-geometric coupling model, the weld thermal deformation displacement vector at the next time step is solved in real time ; the vector represents the displacement size and direction of a point on the weld due to thermal expansion in three-dimensional space;

[0051] The dynamic deviation evaluation module quantifies the risk of the predicted thermal deformation, which aims to convert the deformation displacement with physical dimensions into a standardized risk evaluation index that can be used for decision making; the module receives the thermal deformation displacement vector output by the thermal deformation prediction module, and calculates the dynamic deviation index according to a preset maximum allowed track deviation ; the maximum allowed track deviation refers to the maximum deviation distance in the form of a module length that the physical fusion point and the geometric center line can tolerate to ensure the final weld quality, which serves as a reference for evaluating the severity of track deviation risk, and is a threshold value determined based on a large number of welding experiments and metallographic analysis results;

[0052] Based on the risk index, the risk classification module starts the decision-making mechanism, which aims to provide clear trigger conditions for the subsequent correction control strategy according to the severity of the dynamic deviation; the module compares the dynamic deviation index with the preset threshold value, and judges the risk of track deviation into three levels: safe, first-level risk or second-level risk;

[0053] To achieve a deeper level of quality control, the system also includes a defect tendency prediction module; the purpose of this module is to go beyond pure geometric trajectory control and prospectively predict the risk of potential metallurgical defects; this module not only uses the dynamic deviation index , but also combines real-time welding process parameters , , , and through an embedded defect tendency evolution model, calculates the tendency index of two typical welding defects: the incomplete fusion index and the hot crack index ;

[0054] The comprehensive risk assessment module fuses the multi-dimensional defect risk, and its purpose is to form a single, comprehensive risk measure that can guide the final control execution; this module performs a weighted summation of the incomplete fusion index and the hot crack index output by the defect tendency prediction module, thereby determining a comprehensive process risk factor ; the comprehensive process risk factor is a dimensionless index that fuses multiple potential metallurgical defect risks, and its role is to quantify the overall risk level of producing comprehensive welding defects under the current process conditions, and it is calculated by performing a weighted summation of different defect indices;

[0055] The trajectory correction module, as the execution terminal of the system, completes the entire closed-loop control; the purpose of this module is to generate an optimal welding torch target trajectory based on the results of all previous predictions and assessments, and to adjust the welding process in real time; this module receives the weld geometry position pre-planned by the upper computer, the displacement vector output by the thermal deformation prediction module, and the risk factor output by the comprehensive risk assessment module, and through a risk-driven control algorithm, generates a fully compensated corrected welding torch target trajectory , thereby actively performing closed-loop correction control on the entire welding process. Embodiment 2:

[0056] The thermal deformation prediction module calculates the weld thermal deformation displacement vector by constructing a physical-geometric coupling prediction model; the model combines the welding process parameters, the pre-set geometric influence tensor, the material linear expansion coefficient, the arc heat efficiency coefficient, the material density, the material specific heat capacity, and the effective cross-sectional area of the heat-affected zone;

[0057] Based on Embodiment 1, the thermal deformation prediction module calculates the weld thermal deformation displacement vector by constructing a physical-geometric coupling prediction model ; The model aims to analytically resolve the complex thermo-mechanical coupling phenomena to achieve efficient real-time prediction; its core is an analytical formula derived from the basic physical principle that heat input induces material temperature rise, which in turn leads to thermal strain and macroscopic displacement:

[0058] ;

[0059] wherein, : predicted weld thermal deformation displacement vector at next time step , generated by current process parameters, is a three-dimensional vector calculated by the formula;

[0060] : geometric influence tensor, which physically represents the geometric shape and clamping constraints of the workpiece on the conversion and amplification of thermal strain to displacement, with length dimension; it is obtained through a series of finite element thermal-mechanical coupling simulation analysis; to ensure variable independence, the simulation process uses a set of discrete, pre-set calibration workpiece geometry and constraint conditions{ }, to obtain the corresponding calibration results{ }, and to establish a parameter lookup table or fitting function, which is queried or calculated according to the actual workpiece parameters during model running ;

[0061] : material linear expansion coefficient of the welded part, which is a constant obtained from the standard material manual;

[0062] : thermal efficiency coefficient of the arc, representing the proportion of arc energy converted into workpiece heat, which is a dimensionless coefficient; it is calibrated through welding experiments; to ensure variable independence, the calibration process is carried out under a series of controlled process parameter combinations{ }, the actual absorbed heat is measured by calorimetry , and then calculated by , so as to ensure the determination of independent of real-time input during model running;

[0063] : welding voltage, which is a floating point number monitored in real time by the data acquisition module;

[0064] : welding current, which is a floating point number monitored in real time by the data acquisition module;

[0065] : welding speed, which is a floating point number monitored in real time by the data acquisition module;

[0066] : material density of the welded part, which is a constant obtained from the standard material manual;

[0067] : specific heat capacity of the welded component, constant, obtained from standard material handbook;

[0068] : effective cross-sectional area of the heat-affected zone, float, similar to the way the geometric influence tensor is obtained, through a series of finite element simulation analysis; similarly, to ensure the independence of the variables, this simulation process uses a set of discrete, pre-set plate thicknesses and joint forms , to obtain the corresponding calibration results , and to establish a lookup table or fitting function accordingly;

[0069] This model is suitable for calculation within a pre-set process parameter window, requiring the welding speed to be greater than a minimum threshold to avoid singular values in the model; in system implementation, the input parameters and need to be checked for effectiveness to prevent division by zero errors;

[0070] It should be noted that this model simplifies the complex welding physical process for efficient real-time prediction; for example, the model assumes that the physical performance parameters of the material are constant, and uses a lumped parameter heat input model; the geometric influence tensor linearly maps thermal strain to displacement, mainly capturing the main deformation trend caused by thermal expansion; the accuracy of this model depends on the similarity between the calibration conditions and the actual conditions. Example 3:

[0071] The dynamic deviation evaluation module calculates the dynamic deviation index, including:

[0072] The modulus of the thermal deformation displacement vector is divided by the pre-set maximum allowed trajectory deviation to generate the dynamic deviation index;

[0073] This embodiment is based on embodiment 1, and the method for calculating the dynamic deviation index of the dynamic deviation evaluation module is to convert an absolute displacement vector with physical dimensions into a dimensionless relative risk indicator, so as to facilitate standardized risk level division; the calculation steps are as follows: calculate the modulus of the thermal deformation displacement vector , which represents the size of the predicted weld deviation distance; divide the modulus value by the pre-set maximum allowed trajectory deviation ; the calculation formula is as follows:

[0074] ; Example 4:

[0075] The risk grading module determines the trajectory deviation risk, including:

[0076] When the dynamic deviation index is less than or equal to the preset safety threshold, it is determined to be safe;

[0077] When the dynamic deviation index is greater than the safety threshold and less than or equal to the preset first-level risk threshold, it is determined to be a first-level risk;

[0078] When the dynamic deviation index is greater than the first-level risk threshold, it is determined to be a second-level risk;

[0079] The specific logic of the risk grading module determining the trajectory deviation risk in this embodiment is based on the index output by the dynamic deviation evaluation module , by comparing with a set of preset thresholds, the risk level is divided; in this embodiment, the safety threshold is set to 1, and the first-level risk threshold is set to 1.5; The determination of the first-level risk threshold 1.5 is based on the statistical analysis of historical welding experimental data, it is found that when the deviation index exceeds this value, the probability of structural defects such as incomplete fusion in the weld significantly increases, so the risk level is divided according to this boundary; The determination rules are as follows:

[0080] When the dynamic deviation index , it is determined to be safe, indicating that the predicted trajectory deviation is within the range allowed by the process;

[0081] When the dynamic deviation index , it is determined to be a first-level risk, indicating that the trajectory deviation has exceeded the limit, and there is a risk of deviating from the physical fusion center;

[0082] When the dynamic deviation index , it is determined to be a second-level risk, indicating that the trajectory deviation is severely out of limit, and the risk of producing welding defects is extremely high. Embodiment 5:

[0083] The defect tendency prediction module calculates the incomplete fusion index, including:

[0084] The geometric alignment deviation represented by the dynamic deviation index is weighted and summed with the energy penetration ability represented by the current actual welding line energy and the reference optimal line energy to determine the incomplete fusion index;

[0085] The defect tendency prediction module calculates the hot crack index, including:

[0086] The thermal stress level and the cooling speed level are estimated based on the welding process parameters;

[0087] The thermal stress level and the cooling speed level are weighted and summed to determine the hot crack index;

[0088] Based on Example 1, this embodiment calculates the non-fusion index using the defect tendency prediction module. and hot crack index This method has enabled a deeper understanding of the process, from geometric deviation control to internal metallurgical quality prediction.

[0089] This module calculates the non-fusion index. The specific method is: to use the dynamic deviation index The geometric alignment deviation, as represented, is weighted and summed with the energy penetration capability, represented by the current actual welding line energy and the reference optimal line energy; the calculation formula is as follows:

[0090] ;

[0091] in, Reuse the output of the self-dynamic deviation assessment module; It is based on Calculate the current actual welding line energy; It is the reference optimal line energy set according to the process specifications; The term will make the energy factor dimensionless; These are dimensionless weighting coefficients; to ensure variable independence, the weighting coefficients... The calibration process is based on an independent calibration dataset, which contains measurements from multiple sets of welding experiments. The degree of non-fusion quantified by the corresponding metallographic examination results. By performing multiple linear regression analysis on this dataset, the optimal fit was obtained. value;

[0092] This module calculates the hot crack index. The specific method is as follows: The thermal stress level and cooling rate level are estimated based on welding process parameters, and then these two levels are weighted and summed. To ensure the efficiency of the calculation, both the thermal stress level and cooling rate level are calculated based on the effective temperature rise. For quick estimation, the relationship is as follows:

[0093] ;

[0094] in, This is the dimensionless effective temperature rise coefficient of the heat-affected zone, calibrated experimentally. This simplification aims to capture the influence of process parameter variations on crack tendency, rather than precisely solving for the actual stress and temperature fields. The thermal stress level is represented by a dimensionless thermal stress index. Characterized by the actual cooling rate level With critical cooling rate threshold The ratio is dimensionless, where the actual cooling rate is represented by an empirical formula. To make an estimate, here Let be an empirical cooling coefficient, the dimensions of which are determined by its definition, and be... The final calculation formula is:

[0095] ;

[0096] in, The elastic modulus of the material of the welded component. The yield strength of the welded component at the corresponding temperature. This is the critical cooling rate threshold for the material of the welded component to develop hot cracks. The coefficient of linear expansion of the material of the welded components is a constant, obtained from a handbook or database; weighting coefficients are used to ensure variable independence. The calibration is based on an independent calibration dataset, which includes multiple sets of experimentally measured process parameters, calculated thermal stress and cooling rate indices, and corresponding hot crack susceptibility metallographic analysis results. The cooling coefficient was determined through multiple regression analysis; The cooling rate under different process parameters was measured in the calibration experiment using sensors such as thermocouples. Then according to the relation The physical meaning is obtained by fitting or looking up a table, which ensures the independence of its physical meaning. Example 6:

[0097] The comprehensive risk assessment module identifies comprehensive process risk factors, including:

[0098] The unfusion index and the hot cracking index are weighted and summed to generate a comprehensive process risk factor;

[0099] Based on Example 5, this embodiment uses a comprehensive risk assessment module to determine comprehensive process risk factors. The approach involves merging multiple independent defect propensity indices into a single, guiding comprehensive risk factor; this is achieved by analyzing the unfused index. and hot crack index Perform a weighted summation; the calculation formula is as follows:

[0100] ;

[0101] in, These are the risk weights for different defect indices; risk weights It refers to the coefficient set according to the evaluation of the safety hazard degree of the final structure according to different types of welding defects, which is used to reflect the relative importance of different defects in the comprehensive risk calculation, and is set by experts in the field according to the acceptance level and safety impact evaluation of different defect types in the relevant industry standards such as API1104 or ASME Volume IX. Embodiment 7:

[0102] The trajectory correction module generates a corrected welding gun target trajectory, including:

[0103] Based on the comprehensive process risk factor, a compensation amplification coefficient is generated;

[0104] Superimpose the visually perceived geometric position and the thermal deformation displacement vector modulated by the compensation amplification coefficient to determine the corrected welding gun target trajectory;

[0105] The trajectory correction module is also used for:

[0106] In response to the trajectory deviation risk being determined as a first-level risk, the welding current or arc voltage is increased while the corrected welding gun target trajectory is executed;

[0107] In response to the trajectory deviation risk being determined as a second-level risk, the welding speed is reduced while the corrected welding gun target trajectory is executed;

[0108] This embodiment is based on embodiment 6, and the trajectory correction module generates a corrected welding gun target trajectory and executes a hierarchical correction strategy, which embodies the risk-driven dynamic compensation idea;

[0109] The steps for the module to generate a corrected welding gun target trajectory are: based on the comprehensive process risk factor , a compensation amplification coefficient is generated; the internal logic of this coefficient is that when the system predicts that the comprehensive risk of metallurgical defects increases, the coefficient value is greater than 1, so as to amplify the basic geometric compensation amount and execute more intensive correction to actively avoid risks; the preset next time welding gun target geometric position , which is the welding path planned by the upper computer in advance, is superimposed with the thermal deformation displacement vector modulated by the compensation amplification coefficient that will occur in the next time step to determine the final corrected welding gun target trajectory in the next time; the calculation formula is:

[0110] ;

[0111] The module is also used to execute a grading correction strategy directly linked to the decision result of the risk grading module;

[0112] In response to the trajectory deviation risk being determined as a first-level risk, the system performs trajectory compensation calculated by the above formula at the same time, fine-tuning the welding process parameters, for example, increasing the welding current or arc voltage by 5% to 10%, the adjustment range is set according to the requirements of the welding procedure qualification WPS, on the premise of ensuring the stability of the process window, and is designed to enhance the penetration capacity by moderately increasing the heat input;

[0113] In response to the trajectory deviation risk being determined as a second-level risk, in addition to performing trajectory compensation with risk amplification , the system also significantly adjusts the welding process, for example, reducing the welding speed by 15% to 25%, the adjustment range is also set according to the requirements of the welding procedure qualification WPS, and is designed to ensure fusion by significantly increasing the welding line energy, and possibly introducing advanced process methods such as low-frequency pulse current, to actively control the heat input and crystallization process.

[0114] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present application and not to limit it. Although the present application has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solutions of the present application can be modified or replaced by equivalents without departing from the spirit and scope of the technical solutions of the present application.

Claims

1. A machine vision based automatic tracking welding system for metal welds, characterized in that, The method comprises: a data acquisition module for acquiring initial three-dimensional geometric parameters of a weld and monitoring welding process parameters including welding current, arc voltage and welding speed; a thermal deformation prediction module for calculating a weld thermal deformation displacement vector based on the initial three-dimensional geometric parameters of the weld and the welding process parameters; the calculation formula of the weld thermal deformation displacement vector is: ； wherein, is the predicted weld thermal distortion displacement vector at the next time step resulting from the current process parameters, is the geometric influence tensor, is the material linear expansion coefficient of the welded component, is the thermal efficiency coefficient of the arc, is the welding voltage, is the welding current, is the welding speed, is the material density of the welded component, is the material specific heat capacity of the welded component, is the effective cross-sectional area of the heat-affected zone; a dynamic deviation evaluation module for calculating a dynamic deviation index based on the thermal deformation displacement vector and a preset maximum allowed trajectory deviation; a risk classification module for determining a trajectory deviation risk as safe, a first-level risk or a second-level risk according to the dynamic deviation index; a defect tendency prediction module for calculating an incomplete fusion index and a hot crack index based on the dynamic deviation index and the welding process parameters; the calculation formula of the incomplete fusion index is: ； wherein, is the un-fusion index, is the dynamic bias index, is the actual weld line energy, is the reference optimal line energy, and is the weight coefficient; the calculation formula of the hot crack index is: ； wherein, is the hot cracking index, is the effective temperature rise, is the material elastic modulus of the welded component, is the material linear expansion coefficient of the welded component, is the yield strength of the welded component at the respective temperature, is the welding speed, is the critical cooling rate threshold at which the material of the welded component develops hot cracks, is the cooling coefficient, and is the weight coefficient; a comprehensive risk evaluation module for determining a comprehensive process risk factor based on the incomplete fusion index and the hot crack index; a trajectory correction module for generating a corrected welding gun target trajectory to perform closed-loop correction control on the welding process in combination with the thermal deformation displacement vector and the comprehensive process risk factor.

2. A machine vision-based automatic tracking welding system for metal welds as claimed in claim 1, wherein, The dynamic deviation evaluation module calculates the dynamic deviation index, including: dividing the modulus of the thermal deformation displacement vector by the preset maximum allowed trajectory deviation to generate the dynamic deviation index.

3. The machine vision-based automatic tracking welding system for metal welds of claim 1, wherein, The risk classification module determines the trajectory deviation risk, including: when the dynamic deviation index is less than or equal to a preset safety threshold, it is determined to be safe; when the dynamic deviation index is greater than the safety threshold and less than or equal to a preset first-level risk threshold, it is determined to be a first-level risk; when the dynamic deviation index is greater than the first-level risk threshold, it is determined to be a second-level risk.

4. The machine vision-based automatic tracking welding system for metal weld seams according to claim 1, characterized in that The comprehensive risk evaluation module determines the comprehensive process risk factor, including: performing weighted summation on the incomplete fusion index and the hot crack index to generate the comprehensive process risk factor.

5. The machine vision-based automatic tracking welding system for metal weld seams according to claim 1, characterized in that The trajectory correction module generates the corrected welding gun target trajectory, including: generating a compensation amplification coefficient based on the comprehensive process risk factor; superimposing the geometric position perceived by vision with the thermal deformation displacement vector modulated by the compensation amplification coefficient to determine the corrected welding gun target trajectory.

6. The machine vision-based automatic tracking welding system for metal weld seams according to any one of claims 1, 3, 5, characterized in that, The trajectory correction module is also used for: in response to the trajectory deviation risk being determined to be a first-level risk, increasing the welding current or the arc voltage while executing the corrected welding gun target trajectory; in response to the trajectory deviation risk being determined to be a second-level risk, reducing the welding speed while executing the corrected welding gun target trajectory.

Images