A method and system for compensating terminal posture errors of a five-axis hybrid robot

Abstract

The present invention relates to the technical field of error compensation for five-axis hybrid robots, and in particular to a method and system for compensating the terminal posture error of a five-axis hybrid robot. The method comprises the following steps: obtaining the welding demand data of the vehicle body and analyzing the welding strength difference between the welding points; then, matching the theoretical contact force of the manipulator according to the welding strength difference data, and simulating the thermal conduction toughness loss of adjacent welding points, and further identifying the toughness influence gradient of the thermal chain reaction; then, performing regional gridding processing on the toughness influence gradient data, regulating the contact force of the manipulator, and learning the terminal posture error compensation strategy; finally, designing the automation firmware based on the compensation strategy, and embedding it into the bottom layer of the robot to ensure that the terminal posture error is accurately compensated. The present invention makes the error compensation technology of the five-axis hybrid robot more perfect by optimizing the error compensation technology of the five-axis hybrid robot.

Description

Technical Field

[0001] The present invention relates to the technical field of error compensation for a five-axis hybrid robot, and in particular to a method and system for compensating a terminal position and posture error of a five-axis hybrid robot. Background Art

[0002] Five-axis hybrid robots play an important role in the field of industrial automation, especially in applications such as high-precision welding and assembly. As the manufacturing industry continues to increase its requirements for product quality and production efficiency, the accuracy requirements for robots in performing complex tasks are also getting higher and higher. Terminal posture error compensation has become one of the key technologies to improve robot accuracy and work efficiency. In response to this challenge, a terminal posture error compensation method for a five-axis hybrid robot has emerged. This method analyzes the error source of the robot's end effector and combines the robot's kinematics and dynamics models to adopt various compensation strategies, such as real-time feedback control, model predictive control, and data-driven error correction, to reduce the position and posture errors of the end effector. Through this technology, even if there are errors when the robot performs tasks, it can maintain a high-precision motion trajectory after compensation, thereby ensuring work quality and production efficiency. However, a traditional terminal posture error compensation method for a five-axis hybrid robot cannot accurately analyze the toughness loss caused by the thermal stress of the current welding point to other welding points, so that it cannot accurately adjust the robot's welding behavior dynamically, resulting in a large error in the terminal posture error compensation. Summary of the invention

[0003] Based on this, it is necessary to provide a five-axis hybrid robot terminal posture error compensation method and system to solve at least one of the above technical problems.

[0004] To achieve the above object, a method for compensating the terminal position error of a five-axis hybrid robot is provided, the method comprising the following steps:

[0005] Step S1: obtaining body welding demand design data; performing welding strength analysis between different welding points on the body welding demand design data to obtain welding strength difference data between different welding points;

[0006] Step S2: Based on the welding strength difference data between different welding spots, the theoretical contact force of the end manipulator of the five-axis hybrid robot is matched to obtain the theoretical contact force data of the manipulator; based on the welding strength difference data between different welding spots and the theoretical contact force data of the manipulator, the thermal conduction toughness loss between adjacent welding spots is simulated to obtain the thermal conduction welding spot toughness loss data; the thermal chain reaction toughness influence gradient data between adjacent welding spots is identified for the thermal conduction welding spot toughness loss data to obtain the thermal chain reaction toughness influence gradient data;

[0007] Step S3: regionally grid the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform logical control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logical control data of the welding contact force; and learn the terminal posture error compensation strategy based on the logical control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator;

[0008] Step S4: Based on the manipulator end position error compensation strategy, perform automated firmware design to obtain the manipulator end position error compensation firmware, and embed the manipulator end position error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position error compensation.

[0009] Preferably, step S1 comprises the following steps:

[0010] Step S11: Obtaining vehicle body welding requirement design data;

[0011] Step S12: performing welding point sequence identification processing on the vehicle body welding requirement design data to obtain vehicle body welding point sequence identification data;

[0012] Step S13: analyzing the welding strength between different welding points on the vehicle body welding requirement design data to obtain welding strength difference data between different welding points.

[0013] Preferably, step S2 comprises the following steps:

[0014] Step S21: matching the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points to obtain the theoretical contact force data of the manipulator;

[0015] Step S22: extracting the welding path trajectory of the end manipulator of the five-axis hybrid robot according to the sequential identification data of the welding points of the vehicle body, and obtaining the welding path trajectory of the manipulator;

[0016] Step S23: based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, the manipulator welding path trajectory is simulated for the thermal conduction toughness loss between adjacent welding points to obtain the thermal conduction welding point toughness loss data;

[0017] Step S24: identifying the thermal chain reaction toughness influence gradient data between adjacent solder joints on the thermal conduction solder joint toughness loss data to obtain the thermal chain reaction toughness influence gradient data.

[0018] Preferably, step S23 includes the following steps:

[0019] Step S231: performing instantaneous heat energy intensity analysis of the weld points based on the welding strength difference data between different weld points, and obtaining instantaneous heat energy intensity data of the weld points between different weld points;

[0020] Step S232: performing force-heat coupling density analysis according to the instantaneous heat energy intensity data of the soldering point and the theoretical contact force data of the manipulator to obtain the force-heat coupling density data of the soldering point;

[0021] Step S233: performing CFL conditional incremental relationship analysis on the solder joint force-heat coupling density data to obtain a force-heat density conditional incremental relationship;

[0022] Step S234: performing conditional constraint two-dimensional thermal equation calculation on the force-heat coupling density data of the solder joint according to the force-heat density condition increment relationship to obtain a force-heat density constraint two-dimensional thermal equation;

[0023] Step S235: Acquire body material data; simulate the thermal conduction toughness loss between adjacent welding points based on the body material data and the robot welding path trajectory based on the force-heat density constraint two-dimensional thermal equation to obtain thermal conduction welding point toughness loss data.

[0024] Preferably, step S3 comprises the following steps:

[0025] Step S31: performing regional gridding processing on the thermal chain reaction toughness influence gradient data to obtain the thermal chain reaction toughness influence grid data;

[0026] Step S32: performing logic control of the manipulator contact force on the manipulator theoretical contact force data according to the hot chain reaction toughness influence grid data to obtain the welding contact force logic control data;

[0027] Step S33: dynamically adjusting the welding angle according to the hot chain reaction toughness influence grid data and the welding contact force logic control data to obtain the welding angle dynamic adjustment data;

[0028] Step S34: Based on the welding contact force logic control data and the welding angle dynamic adjustment data, the terminal position and posture error compensation strategy is learned to obtain the manipulator terminal position and posture error compensation strategy.

[0029] Preferably, step S32 includes the following steps:

[0030] Step S321: Derivation and measurement of the linear relationship of the theoretical contact force data of the manipulator according to the grid data of the thermal chain reaction toughness influence, to obtain the contact force thermal chain toughness reaction relationship derivation data between different grids;

[0031] Step S322: performing heat source toughness attenuation diffusion position anisotropy analysis on the contact force thermal chain toughness response relationship derivation data between different grids to obtain heat source toughness diffusion position anisotropy data;

[0032] Step S323: based on the different data of heat source toughness diffusion orientation, the contact force heat chain toughness reaction relationship derivation data between different grids are used to locate the reaction concentration point, and the toughness trend abnormal increment of the reaction concentration point is identified to obtain the toughness trend abnormal increment data of the concentration point;

[0033] Step S324: Based on the abnormal incremental data of the toughness trend at the concentrated point, the contact force of the robot is logically controlled to obtain the welding contact force logical control data.

[0034] Preferably, the present invention further provides a five-axis hybrid robot terminal position and posture error compensation system, which is used to execute the five-axis hybrid robot terminal position and posture error compensation method as described above, and the five-axis hybrid robot terminal position and posture error compensation system comprises:

[0035] The welding strength difference analysis module is used to obtain the body welding demand design data; the welding strength analysis between different welding points is performed on the body welding demand design data to obtain the welding strength difference data between different welding points;

[0036] The module for identifying the impact of thermal chain reaction toughness is used to match the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points, and obtain the theoretical contact force data of the manipulator; to simulate the thermal conduction toughness loss between adjacent welding points based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, and obtain the toughness loss data of the thermal conduction welding points; to identify the thermal chain reaction toughness impact gradient data between adjacent welding points on the toughness loss data of the thermal conduction welding points, and obtain the thermal chain reaction toughness impact gradient data;

[0037] The terminal posture error compensation strategy learning module is used to perform regional grid processing on the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform the logic control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logic control data of the welding contact force; and perform the terminal posture error compensation strategy learning based on the logic control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator;

[0038] The automated firmware design module is used to perform automated firmware design based on the manipulator end position and posture error compensation strategy, obtain the manipulator end position and posture error compensation firmware, and embed the manipulator end position and posture error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position and posture error compensation.

[0039] The beneficial effect of the present invention is that by obtaining the design data of the body welding requirements and performing welding strength analysis between welding points, the bearing capacity and quality difference of each welding point can be accurately understood. This analysis provides basic data for subsequent robot control, so that the contact force can be optimized according to the strength difference between different welding points, thereby ensuring that the strength and stability of each welding point meet the design requirements, which helps to improve the overall welding quality and structural strength. By matching the contact force based on the welding strength difference data, the theoretical contact force of the manipulator is matched with the requirements of different welding points. This helps to reduce the problem of poor welding caused by uneven contact force. In addition, by simulating the thermal conduction toughness loss of adjacent welding points and identifying the gradient of the toughness influence of the thermal chain reaction, the influence of temperature on the contact part during welding can be predicted, thereby optimizing welding parameters, reducing the influence of heat loss on the quality of welding points, and improving the consistency and reliability of welding. Regional gridding of the gradient data of the toughness influence of the thermal chain reaction can further refine the contact force control of the manipulator during welding, so that the force distribution is more reasonable. The refined control of this process can effectively avoid welding defects caused by uneven contact force, and at the same time, the terminal posture error compensation strategy based on the data can be learned to improve the accuracy of the robot welding operation. By adjusting the end position of the manipulator, accurate docking during the welding process is ensured to avoid damage to the workpiece or unqualified welding quality due to errors. Based on the learning of the end position error compensation strategy, the automated firmware is designed and developed and embedded into the bottom layer of the five-axis hybrid robot. The implementation of this step can correct the end position error of the robot during the welding process in real time, thereby improving the adaptability and accuracy of the robot in actual welding tasks. Through the automated firmware, the robot can automatically compensate for errors when performing complex tasks, ensuring that welding accuracy is guaranteed, improving overall production efficiency and product quality, and reducing manual intervention and maintenance costs. Therefore, the present invention provides a five-axis hybrid robot terminal posture error compensation method, which is an improvement on a traditional five-axis hybrid robot terminal posture error compensation method. It solves the problem that the traditional five-axis hybrid robot terminal posture error compensation method cannot accurately analyze the toughness loss caused by the thermal stress of the current welding point to other welding points, and thus cannot accurately perform dynamic adjustment of the robot's welding behavior, resulting in a large error in the terminal posture error compensation. It improves the accuracy of the analysis of the toughness loss caused by the thermal stress of the current welding point to other welding points, thereby accurately performing dynamic adjustment of the robot's welding behavior and reducing the error in terminal posture error compensation. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A schematic diagram of the steps of a method for compensating the terminal position and posture error of a five-axis hybrid robot;

[0041] Figure 2 for Figure 1Detailed implementation steps of step S2 in the flowchart;

[0042] Figure 3 for Figure 1 Detailed implementation steps of step S3 in FIG. DETAILED DESCRIPTION

[0043] The following is a clear and complete description of the technical method of the present invention in conjunction with the accompanying drawings. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by technicians in this field without creative work are within the scope of protection of the present invention.

[0044] In addition, the accompanying drawings are only schematic illustrations of the present invention and are not necessarily drawn to scale. The same reference numerals in the figures represent the same or similar parts, and their repeated description will be omitted. Some of the block diagrams shown in the accompanying drawings are functional entities and do not necessarily correspond to physically or logically independent entities. The functional entities can be implemented in software form, or implemented in one or more hardware modules or integrated circuits, or implemented in different networks and / or processor methods and / or microcontroller methods.

[0045] It should be understood that, although the terms "first", "second", etc. may be used herein to describe various units, these units should not be limited by these terms. These terms are used only to distinguish one unit from another unit. For example, without departing from the scope of the exemplary embodiments, the first unit may be referred to as the second unit, and similarly the second unit may be referred to as the first unit. The term "and / or" used herein includes any and all combinations of one or more of the listed associated items.

[0046] To achieve this, please refer to Figures 1 to 3 A method for compensating the terminal position error of a five-axis hybrid robot is provided, the method comprising the following steps:

[0047] Step S1: obtaining body welding demand design data; performing welding strength analysis between different welding points on the body welding demand design data to obtain welding strength difference data between different welding points;

[0048] Step S2: Based on the welding strength difference data between different welding spots, the theoretical contact force of the end manipulator of the five-axis hybrid robot is matched to obtain the theoretical contact force data of the manipulator; based on the welding strength difference data between different welding spots and the theoretical contact force data of the manipulator, the thermal conduction toughness loss between adjacent welding spots is simulated to obtain the thermal conduction welding spot toughness loss data; the thermal chain reaction toughness influence gradient data between adjacent welding spots is identified for the thermal conduction welding spot toughness loss data to obtain the thermal chain reaction toughness influence gradient data;

[0049] Step S3: regionally grid the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform logical control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logical control data of the welding contact force; and learn the terminal posture error compensation strategy based on the logical control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator;

[0050] Step S4: Based on the manipulator end position error compensation strategy, perform automated firmware design to obtain the manipulator end position error compensation firmware, and embed the manipulator end position error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position error compensation.

[0051] In the embodiment of the present invention, reference Figure 1 The above is a schematic flow chart of the steps of a method for compensating the terminal posture error of a five-axis hybrid robot of the present invention. In this example, the method for compensating the terminal posture error of the five-axis hybrid robot includes the following steps:

[0052] Step S1: obtaining body welding demand design data; performing welding strength analysis between different welding points on the body welding demand design data to obtain welding strength difference data between different welding points;

[0053] In an embodiment of the present invention, first, an electronic file containing vehicle body structure information, welding point position coordinates, welding process parameters (e.g., welding current amplitude, welding voltage, welding speed), and welding standard specifications is exported from the digital design system of the vehicle engineering department, and the file format is extensible markup language (XML). The XML file contains the three-dimensional model data of all sheet metal parts constituting the vehicle body, and the model data is stored in STEP format. Secondly, the XML file is parsed to extract the precise three-dimensional spatial coordinates of each welding point, and the coordinate system is established based on the vehicle body coordinate system. At the same time, the welding process parameters associated with each welding point are extracted. For example, for thin plate lap welding, the process parameters include: the welding current setting value is 300 amperes, the welding voltage setting value is 12 volts, and the welding speed setting value is 5 mm per second. The welding standard specification specifies the minimum tensile strength requirements of the welded joint under different material combinations and plate thicknesses, in megapascals (MPa). Subsequently, a database containing all welding point information is constructed, and the fields of the database table include: welding point number (unique identifier), X coordinate, Y coordinate, Z coordinate, welding current setting value, welding voltage setting value, welding speed setting value, and minimum tensile strength requirement.

[0054] In another embodiment, first, the body welding requirement design data is obtained. This data is derived from the design drawings of the body and the welding process requirements, and contains information such as the location of the welding point, welding type, material properties and welding strength requirements. Next, the welding strength analysis between different welding points is performed, and the finite element analysis method is used to simulate the welding points and calculate the welding strength of each welding point. This process analyzes the stress conditions of the welding points in detail by setting boundary conditions and applying loads, and obtains the welding strength difference data between different welding points. Through the quantitative analysis of the welding strength difference, it is possible to identify which welding points will face greater stress concentration or failure risks during the production process. The welding strength difference data finally generated at this stage provides important input for subsequent steps and lays the foundation for the contact force matching and heat conduction analysis of the manipulator.

[0055] Step S2: Based on the welding strength difference data between different welding spots, the theoretical contact force of the end manipulator of the five-axis hybrid robot is matched to obtain the theoretical contact force data of the manipulator; based on the welding strength difference data between different welding spots and the theoretical contact force data of the manipulator, the thermal conduction toughness loss between adjacent welding spots is simulated to obtain the thermal conduction welding spot toughness loss data; the thermal chain reaction toughness influence gradient data between adjacent welding spots is identified for the thermal conduction welding spot toughness loss data to obtain the thermal chain reaction toughness influence gradient data;

[0056] In the embodiment of the present invention, for the welding strength difference data between different welding points obtained in step S1, a mapping relationship between the welding strength difference and the theoretical contact force of the manipulator at the end of the manipulator is established; wherein the formula of the mapping relationship is:

[0057] ;

[0058] In the formula, It represents the theoretical contact force of the robot end, in Newton (N). Represents the difference in welding strength between two welding spots, in Newtons; the mapping relationship is defined by a piecewise linear function. When the welding strength difference is less than 100 Newtons, the theoretical contact force of the manipulator is set to 100 Newtons. When the welding strength difference is between 100 Newtons and 500 Newtons, the theoretical contact force of the manipulator is set to 100 Newtons plus 0.2 times the welding strength difference. When the welding strength difference is greater than 500 Newtons, the theoretical contact force of the manipulator is set to 200 Newtons plus 0.1 times the welding strength difference. For example, if the welding strength difference between welding spot C and welding spot D is 300 Newtons, then according to the mapping relationship, the theoretical contact force of the manipulator between the two welding spots is set to 100 Newtons + 300 Newtons multiplied by 0.2 = 160 Newtons. Perform the above calculations for all adjacent welding spot pairs to obtain the theoretical contact force data of the manipulator at each welding spot position, and store the data in the form of a list, where each element in the list corresponds to a welding spot and its theoretical contact force value. Based on the welding strength difference data between different welding points obtained in step S1, the welding heat input of each welding point is calculated using the first law of thermodynamics. The welding heat input is proportional to the square of the welding current, the welding voltage, and the welding time. The formula is: welding heat input = η×the square of the welding current×the welding voltage×the welding time, where η is the welding thermal efficiency, and its value is obtained from the welding process manual.

[0059] Secondly, combined with the theoretical contact force data of the manipulator obtained in step S2, a finite difference heat conduction model is established for two adjacent welds. Considering that the welding process is a dynamic process, the force generated by the robot end manipulator during the welding process includes not only contact stress but also dynamic friction. Therefore, the dynamic force at the weld point should be modeled based on the motion trajectory of the manipulator and the welding speed. Dynamic contact stress modeling: During the welding process, the contact force at the end of the manipulator is decomposed into two parts: static contact stress and dynamic friction: static contact stress The theoretical contact force designed by welding requirements is directly given as the initial force condition of welding. It is determined by the robot moving speed v, the material surface friction coefficient μ and the contact area A during the welding process, and its expression is: =μ A dynamic friction force is further converted into equivalent dynamic stress and included in the solder joint force model. This model simultaneously considers the effects of heat conduction, heat radiation, friction and contact stress.

[0060] Heat conduction is described by Fourier's law, and thermal radiation is described by Stefan-Boltzmann's law. Fourier's law describes the phenomenon of heat conduction, and the formula is: , where q is the heat flux (unit: W / m²), k is the thermal conductivity of the material (unit: W / m·K), is the temperature gradient; the Stefan-Boltzmann law describes the thermal radiation process, and the formula is: Where σ is the Stefan-Boltzmann constant, ε is the emissivity of the material, T is the surface temperature, is the ambient temperature. The boundary condition of the model is set to the temperature at the weld point equal to the weld pool temperature, which is determined by the melting point of the weld material. The initial condition is set to the initial temperature of the workpiece, which is usually room temperature. Then, the explicit finite difference method is used to solve the heat conduction model; the explicit finite difference method solution formula is: ,in, is the temperature at position (i, j) at time n, k is the thermal conductivity, Δt is the time step, ρ is the density of the material, is the specific heat capacity of the material, and Δx is the spatial step length. According to the CFL condition, the maximum allowable value of the time step length is determined to ensure the stability of the calculation. In each time step, the temperature change of each unit is calculated. When the welding of a solder joint is completed, the heat generated by it will diffuse to the surrounding area, affecting the temperature distribution of the next solder joint area. Finally, based on the calculated temperature field distribution, the toughness-temperature relationship curve of the material is used to evaluate the toughness loss between adjacent solder joints. This curve describes the law that the toughness value of the material decreases with increasing temperature. The toughness value is expressed by the absorbed energy value obtained by the Charpy impact test, and the unit is joule (J). For example, at a certain time point, the temperature of a unit between solder joints E and solder joints F is 200 degrees Celsius. By consulting the toughness-temperature relationship curve of the material, the toughness value of the material at this temperature is 50 joules, while the toughness value at room temperature is 80 joules, and the toughness loss of the unit is 30 joules. The above calculation is performed on the units between all adjacent solder joints to obtain the toughness loss data of the heat conduction solder joints, and the data is stored in the form of a matrix. The elements of the matrix represent the toughness loss caused by heat conduction between two adjacent solder joints. The toughness loss data of the heat conduction solder joint obtained in step S2 is subjected to Gaussian filtering to remove noise interference and smooth the toughness loss distribution. The standard deviation of the Gaussian filter is set according to the spacing between solder joints. Secondly, for each solder joint, the toughness loss gradient within a certain range around it is calculated. The gradient calculation is implemented using the Sobel operator, and the toughness loss change rate in the horizontal and vertical directions is calculated respectively. The amplitude and direction of the gradient indicate the severity and direction of the toughness loss change. For example, if the toughness loss change rate in the horizontal direction around a solder joint is 5 joules per millimeter, and the toughness loss change rate in the vertical direction is 3 joules per millimeter, then the toughness loss gradient amplitude around the solder joint is (5 squared + 3 squared) joules per millimeter under the square root, and the gradient direction is the inverse tangent (3 / 5). Then, the toughness loss gradient direction between adjacent solder joints is analyzed. If the gradient directions of adjacent solder joints are approximately the same (for example, the angle is less than 30 degrees), it is considered that there is a thermal chain reaction, that is, the thermal influence of the previous solder joint significantly affects the toughness of the next solder joint. On the contrary, if the gradient direction difference is large, it is considered that the thermal chain reaction is not obvious. Finally, for pairs of adjacent welds with thermal chain reactions, calculate their toughness influence gradient. The toughness influence gradient is defined as the toughness loss value of the latter weld minus the toughness loss value of the previous weld, divided by the distance between the two welds. For example, the toughness loss of weld G is 40 joules, the toughness loss of the adjacent weld H is 50 joules, and the distance between the two welds is 10 mm. The toughness influence gradient of the thermal chain reaction from weld G to weld H is (50 joules - 40 joules) / 10 mm = 1 joule per millimeter. The toughness influence gradient data of all pairs of adjacent welds with thermal chain reactions are stored in a list, where each element contains the numbers of the two welds and the toughness influence gradient value between them.

[0061] Step S3: regionally grid the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform logical control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logical control data of the welding contact force; and learn the terminal posture error compensation strategy based on the logical control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator;

[0062] In an embodiment of the present invention, the range of the area that needs to be gridded is determined, and the range covers all welding points. The boundary of the gridded area is determined according to the maximum size and the minimum size of the vehicle body. For example, the range of the vehicle body in the X-axis direction is 0 to 5000 mm, and the range in the Y-axis direction is 0 to 2000 mm. Secondly, determine the grid division method and grid size. Use a uniform rectangular grid for division. The grid size is set according to the actual accuracy requirements, for example, the side length of the grid is set to 50 mm. Then, traverse all grids to determine whether each grid contains welding points. For grids containing welding points, mark them as "valid grids". For valid grids, calculate the average toughness influence gradient of all welding point pairs with thermal chain reactions in the grid. For example, a grid contains two welding point pairs with thermal chain reactions, and their toughness influence gradients are 0.8 joules per millimeter and 1.2 joules per millimeter, respectively. Then, the average toughness influence gradient of the grid is (0.8 + 1.2) / 2 = 1 joule per millimeter. The average toughness influence gradient value of each valid grid is stored to obtain the thermal chain reaction toughness influence grid data. The data can be represented in the form of a two-dimensional array, and each element of the array corresponds to a grid, storing the average toughness influence gradient value of the grid. For a grid that does not contain a solder joint, its toughness influence gradient value is set to zero. For the theoretical contact force data of the manipulator obtained in step S2 and the toughness influence grid data of the thermal chain reaction obtained in step S3, a manipulator contact force control rule is designed. The rule is based on the following logic: when the average toughness influence gradient of the grid is positive and the absolute value is large, it indicates that the toughness of the next solder joint is greatly affected by the previous solder joint, and the material is prone to thermal brittle cracking. At this time, the actual contact force of the manipulator needs to be reduced. On the contrary, when the average toughness influence gradient is close to zero or negative, the actual contact force of the manipulator can be maintained or appropriately increased. The specific control rule is defined by a piecewise linear function. For example, when the average toughness influence gradient of the grid is greater than 0.5 joules per millimeter, the actual contact force of the manipulator of the solder joint in the grid is reduced by 10% based on the theoretical contact force. When the average toughness influence gradient of the grid is between -0.5 joules per millimeter and 0.5 joules per millimeter, the actual contact force of the manipulator remains unchanged from the theoretical value. When the average toughness influence gradient of the grid is less than -0.5 joules per millimeter, the actual contact force of the manipulator increases by 5% on the basis of the theoretical contact force. Traverse all the welds, adjust the contact force of the manipulator according to the average toughness influence gradient of the grid where they are located, and obtain the welding contact force logic control data. The data is stored in the form of a list, and each element in the list corresponds to a weld and its adjusted actual contact force value of the manipulator. The learning process of the manipulator end posture error compensation strategy is as follows: a training data set containing input features and target outputs is constructed through the forward kinematics model.The input features include: the three-dimensional coordinates of the weld, the logic control data of the welding contact force (i.e., the actual contact force of the manipulator after adjustment), and the average hot chain reaction toughness influence gradient of the grid where the weld is located. The target output is the posture error compensation of the end of the manipulator, including the position error compensation (the translation along the X, Y, and Z axes) and the posture error compensation (the rotation angle around the X, Y, and Z axes). The posture error compensation can be obtained by using a high-precision external measurement system (e.g., a laser tracker) to measure the actual posture of the end of the manipulator in real time during the actual welding process, and compare it with the theoretical posture to obtain the posture error. A large number of input features and corresponding posture error data under different welding conditions (different welding parameters, different weld position, different hot chain reaction toughness influence gradient) are collected to construct a training data set. Then, a machine learning algorithm is selected for model training. A regression model based on the gradient boosting decision tree (GBDT) is used. The GBDT model trains multiple decision trees in an iterative manner, and weighted sums the prediction results of all decision trees to obtain the final prediction result. The goal of model training is to minimize the mean square error between the predicted pose error and the actual pose error. During the model training process, the cross-validation method is used to evaluate the performance of the model, and the hyperparameters of the model (for example, the number of decision trees, the depth of the decision tree, and the learning rate) are adjusted to obtain the best prediction performance. After the training is completed, the manipulator end pose error compensation strategy is obtained, which is stored in the form of trained GBDT model parameters. The model receives the three-dimensional coordinates of the solder joint, the adjusted actual contact force of the manipulator, and the average thermal chain reaction toughness influence gradient of the grid where the solder joint is located as input, and outputs the corresponding error compensation amount of the manipulator end position and posture; the calculation formula for the manipulator end pose error compensation amount is:. in, Indicates the position error compensation amount of the robot end, indicating the correction amount required to correct the position and posture of the robot end, including compensation for displacement and rotation; The area of ​​the grid where the weld point is located represents the spatial range of the integration, usually covering the entire grid of the weld area; The three-dimensional coordinate vector of the solder joint indicates the position of the solder joint in space; Indicates the target position of the robot end, usually the position that the robot should reach under ideal conditions; represents the actual contact force vector of the manipulator at the welding point position r, which represents the actual force applied at each welding point; Indicates the target contact force, which is usually the ideal contact force set by the control system according to welding requirements; It represents the average thermal chain reaction toughness influence gradient of the grid where the solder joint is located. It indicates the influence of the thermal chain reaction around the solder joint on the material toughness, which usually depends on heat conduction and material properties. The weight factor representing the contact force difference. It is used to adjust the effect of the difference between the contact force and the target contact force on the amount of posture error compensation; Thermal chain attenuation factor. It is used to control the influence of thermal chain effect on the compensation amount of manipulator end position error. The influence of welding points at a longer distance on posture compensation is weakened; The distance scale that represents the impact of the heat chain effect; it describes the extent of the heat chain's influence on the area surrounding the solder joint and is usually related to the heat transfer characteristics of the soldering process.

[0063] Step S4: Based on the manipulator end position error compensation strategy, perform automated firmware design to obtain the manipulator end position error compensation firmware, and embed the manipulator end position error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position error compensation.

[0064] In an embodiment of the present invention, the manipulator end position error compensation strategy obtained by training in step S3 is converted into a C code that can be run at the bottom layer of the five-axis hybrid robot control system. This can be achieved by using a model export tool, for example, exporting the trained Scikit-learn model to ONNX format, and then converting it into C code using the ONNX Runtime library. The C code contains the structural information of the GBDT model (nodes, branch conditions, values ​​of leaf nodes of the decision tree) and model parameters (weights of each decision tree). Secondly, the main loop program of the firmware is designed. The main loop program runs at a fixed frequency (for example, 1000 Hz). In each cycle, the current weld point coordinate information and welding parameter information are first obtained from the sensor of the robot control system. Then, according to the current weld point coordinates, the grid to which it belongs is determined, and the average thermal chain reaction toughness influence gradient value of the grid is read from the regional grid data obtained in step S3. At the same time, the adjusted actual contact force value of the manipulator of the current weld point is obtained from the welding contact force logic control data obtained in step S3. The weld point coordinates, the adjusted actual contact force of the manipulator, and the average hot chain reaction toughness influence gradient of the grid are used as inputs, and the converted GBDT model C code is called to calculate the posture error compensation of the manipulator end, including the translation compensation values ​​along the X, Y, and Z axes and the rotation compensation values ​​around the X, Y, and Z axes. Finally, the calculated posture error compensation is superimposed on the robot's target posture instruction to generate a corrected target posture instruction, which is sent to the robot's motion control module to drive the robot to perform precise welding operations. The manipulator end posture error compensation firmware embedding process is as follows: First, according to the hardware architecture and software platform of the five-axis hybrid robot control system, the firmware embedding method is determined. Usually, the robot control system uses a real-time operating system (RTOS), and the firmware needs to be compiled into an executable file compatible with the RTOS. Use a cross-compilation tool chain to compile the designed C code into a binary file for the target platform. Then, upload the compiled firmware file to the controller of the robot control system through a network connection or a USB interface.

[0065] Step S1 includes the following steps:

[0066] Step S11: Obtaining vehicle body welding requirement design data;

[0067] Step S12: performing welding point sequence identification processing on the vehicle body welding requirement design data to obtain vehicle body welding point sequence identification data;

[0068] Step S13: analyzing the welding strength between different welding points on the vehicle body welding requirement design data to obtain welding strength difference data between different welding points.

[0069] In an embodiment of the present invention, the process of obtaining the body welding demand design data includes extracting welding-related data from the design document. First, a three-dimensional model file of the body is opened using computer-aided design (CAD) software, and the file format is STEP or IGES. Through a data parsing algorithm, the weld information in the model is read, including the three-dimensional coordinates of the weld, the chemical composition of the welding material, the welding process parameters, and the corresponding welding strength requirements. During the data extraction process, it is ensured that all weld information is complete and accurate to form a unified welding demand data set for subsequent processing. The obtained welding demand design data is subjected to weld sequence identification processing. First, the connection relationship and welding sequence of each weld are identified, and the welds are sorted by an algorithm. A topological sorting algorithm is used to generate weld sequence identification data from the starting weld to the ending weld according to the welding process requirements and the welding path. During the sorting process, a priority queue or a linked list is used to store the welds and their sequence information to ensure that the welding process is carried out in a predetermined order, thereby forming a structured weld sequence data set, which provides a basis for subsequent welding path planning and control. The welding strength analysis is performed on the body welding demand design data, and the stress condition of the weld is simulated by the finite element method. First, a welding model is established to define the material properties of the solder joint, such as elastic modulus, yield strength, and thermal conductivity. Then, boundary conditions and loading conditions are set during the welding process to simulate the thermal and mechanical stresses on the solder joint during the actual welding process. The welding model is meshed and the numerical solver is used for calculation to obtain the welding strength value of each solder joint. The analysis results show the difference in welding strength between different solder joints, forming a welding strength difference data set, which provides an important basis for subsequent welding contact force matching and heat conduction analysis.

[0070] Step S2 includes the following steps:

[0071] Step S21: matching the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points to obtain the theoretical contact force data of the manipulator;

[0072] Step S22: extracting the welding path trajectory of the end manipulator of the five-axis hybrid robot according to the sequential identification data of the welding points of the vehicle body, and obtaining the welding path trajectory of the manipulator;

[0073] Step S23: based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, the manipulator welding path trajectory is simulated for the thermal conduction toughness loss between adjacent welding points to obtain the thermal conduction welding point toughness loss data;

[0074] Step S24: identifying the thermal chain reaction toughness influence gradient data between adjacent solder joints on the thermal conduction solder joint toughness loss data to obtain the thermal chain reaction toughness influence gradient data.

[0075] As an example of the present invention, refer to Figure 2 As shown, in this example, step S2 includes:

[0076] Step S21: matching the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points to obtain the theoretical contact force data of the manipulator;

[0077] In an embodiment of the present invention, the welding strength difference data between different welding spots obtained in step S13 is read. A mapping relationship between the welding strength difference and the theoretical contact force of the end of the manipulator is established. When the absolute value of the welding strength difference is less than or equal to 100 Newtons, the theoretical contact force of the manipulator is set to 150 Newtons. When the absolute value of the welding strength difference is greater than 100 Newtons and less than or equal to 300 Newtons, the theoretical contact force of the manipulator is set to 150 Newtons plus 0.1 times the absolute value of the welding strength difference. When the absolute value of the welding strength difference is greater than 300 Newtons, the theoretical contact force of the manipulator is set to 170 Newtons plus 0.05 times the absolute value of the welding strength difference. For example, if the welding strength difference between adjacent welding spots A and B is 200 Newtons, the theoretical contact force of the manipulator between the two welding spots is set to 150 Newtons plus 200 Newtons multiplied by 0.1, that is, 170 Newtons. Traverse all adjacent welding point pairs in the welding sequence, determine the corresponding manipulator theoretical contact force according to their welding strength differences using the above piecewise linear mapping relationship, and store the manipulator theoretical contact force value corresponding to each welding point in a list.

[0078] In another embodiment, in this step, first, a contact mechanics model between the manipulator and the welding spot is established based on the welding strength difference data obtained in step S13. The model takes into account the geometric shape, material properties and welding process requirements of the welding spot. Then, the theoretical contact force of each welding spot is calculated using a mechanical analysis method. In the specific operation, the static equilibrium condition is used to convert the difference in welding strength into the contact force requirement. By establishing a linear equation group, the welding strength of the welding spot is associated with the contact force of the manipulator to obtain the theoretical contact force data of each welding spot. These data will be used for subsequent welding path planning and contact control to ensure that the manipulator can provide appropriate welding force during the welding process.

[0079] Step S22: extracting the welding path trajectory of the end manipulator of the five-axis hybrid robot according to the sequential identification data of the welding points of the vehicle body, and obtaining the welding path trajectory of the manipulator;

[0080] In the embodiment of the present invention, the welding point sequence identification data generated in step S12 is first read to determine the order of welding. Next, the welding path trajectory of the manipulator is generated using an interpolation algorithm based on the three-dimensional coordinate information of the welding point. The spline interpolation method (such as cubic spline interpolation) is usually used to smooth the path between the welding points to ensure that the manipulator maintains a stable motion trajectory during the welding process. The adjustment of the welding posture needs to be considered during the path extraction process to ensure that the end of the manipulator always maintains the optimal welding angle and position. The welding path trajectory finally generated will be recorded in the form of a sequence of coordinate points to provide a basis for subsequent welding process control.

[0081] Step S23: based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, the manipulator welding path trajectory is simulated for the thermal conduction toughness loss between adjacent welding points to obtain the thermal conduction welding point toughness loss data;

[0082] In the embodiment of the present invention, each straight line segment in the robot welding path trajectory extracted in step S22 represents the welding process between two adjacent welding points. The welding strength difference data between the corresponding welding points obtained in step S13 and the theoretical contact force data of the robot obtained in step S21 are read. Based on the welding strength difference data, the heat input in the welding process is calculated using the welding heat input calculation formula: the heat input is equal to the welding current multiplied by the welding voltage multiplied by the welding time. The welding current and welding voltage take the target values ​​of the corresponding welding points, and the welding time is calculated according to the distance between the two welding points and the set welding speed. The heat input is used as the heat source, and the one-dimensional heat conduction equation is used to simulate the heat diffusion process around the welding point. The equation is solved using the finite difference method. The area around the welding point is discretized into a number of small units, and the initial temperature is set to room temperature. The boundary condition is set as the center temperature of the welding point calculated according to the heat input. In each time step, the temperature change of each unit is calculated. According to the calculated temperature distribution, the toughness-temperature curve of the welding material is consulted to obtain the material toughness value at the corresponding temperature. The toughness loss is equal to the initial toughness value of the material minus the toughness value at the current temperature. For example, if the temperature at a certain welding point rises to 150 degrees Celsius, the toughness value of the material at 150 degrees Celsius is 50 joules, and the initial toughness value is 80 joules, then the toughness loss at the welding point is 30 joules. The heat conduction process is simulated for each welding path trajectory to obtain the heat conduction toughness loss value at each welding point, and these values ​​are stored in a list.

[0083] Step S24: identifying the thermal chain reaction toughness influence gradient data between adjacent solder joints on the thermal conduction solder joint toughness loss data to obtain the thermal chain reaction toughness influence gradient data.

[0084] In an embodiment of the present invention, the list of toughness loss data of the heat conduction weld obtained in step S23 is read. According to the welding sequence, the toughness loss gradient between two adjacent welds is calculated. The toughness loss gradient is equal to the toughness loss value of the latter weld minus the toughness loss value of the previous weld, divided by the distance between the two welds. The distance between the welds is obtained by calculating the Euclidean distance of their three-dimensional coordinates. For example, the toughness loss of weld C in the welding sequence is 25 joules, the toughness loss of weld D is 30 joules, and the distance between weld C and weld D is 10 mm. The thermal chain reaction toughness influence gradient from weld C to weld D is (30 joules-25 joules) / 10 mm, that is, 0.5 joules per mm. Traverse all adjacent weld pairs in the welding sequence, calculate the thermal chain reaction toughness influence gradient between them, and store the results in a list, each element of the list contains the numbers of the two adjacent welds and the thermal chain reaction toughness influence gradient value between them. In another embodiment, the thermal chain reaction toughness influence gradient data between adjacent welds is identified by analyzing the thermal conduction weld toughness loss data. First, the heat conduction loss data is combined with the geometric position and heat distribution of the solder joints, and the thermal gradient analysis method is applied to calculate the thermal chain reaction between adjacent solder joints. The specific operation is to use the differential method to process the heat conduction loss data, identify the gradient of the heat loss change, and form the thermal chain reaction toughness influence gradient data between solder joints. This process requires accurate calculation of the temperature field and stress field changes in the area around the solder joint, and numerical calculation of the temperature gradient and stress gradient using the finite difference or finite element method. Finally, through data processing, the thermal chain reaction toughness influence gradient data between each pair of adjacent solder joints is obtained. These data provide a basis for optimizing the welding contact force and welding path adjustment. The thermal chain reaction refers to the process in which the thermal influence of a solder joint affects the toughness of adjacent solder joints in a chain reaction manner. By identifying the gradient data, the degree and changing trend of the thermal influence can be more accurately understood.

[0085] Step S23 includes the following steps:

[0086] Step S231: performing instantaneous heat energy intensity analysis of the weld points based on the welding strength difference data between different weld points, and obtaining instantaneous heat energy intensity data of the weld points between different weld points;

[0087] Step S232: performing force-heat coupling density analysis according to the instantaneous heat energy intensity data of the soldering point and the theoretical contact force data of the manipulator to obtain the force-heat coupling density data of the soldering point;

[0088] Step S233: performing CFL conditional incremental relationship analysis on the solder joint force-heat coupling density data to obtain a force-heat density conditional incremental relationship;

[0089] Step S234: performing conditional constraint two-dimensional thermal equation calculation on the force-heat coupling density data of the solder joint according to the force-heat density condition increment relationship to obtain a force-heat density constraint two-dimensional thermal equation;

[0090] Step S235: Acquire body material data; simulate the thermal conduction toughness loss between adjacent welding points based on the body material data and the robot welding path trajectory based on the force-heat density constraint two-dimensional thermal equation to obtain thermal conduction welding point toughness loss data.

[0091] In the embodiment of the present invention, the welding strength difference data between each welding point is first obtained, and the instantaneous heat energy intensity of each welding point is calculated. By analyzing the heat input and heat output of the welding point during the welding process, combined with the welding strength difference, the heat conduction equation is used to solve the instantaneous heat energy intensity. In the specific implementation, the heat input of the welding point (such as welding current, voltage, welding speed and other parameters) is used to calculate the instantaneous heat energy generated during the welding process. Then, the heat conduction model is used to simulate the diffusion process of heat energy in the welding point and its adjacent areas according to the geometric shape of the welding point, the thermal conductivity of the material and the welding process parameters. Through these analyses, the instantaneous heat energy intensity data of each welding point is obtained, which reflects the change of instantaneous heat of different welding points during the welding process, and provides a basis for the subsequent force-heat coupling density analysis. Combined with the instantaneous heat energy intensity data of the welding point and the theoretical contact force data of the manipulator, the force-heat coupling density analysis is performed. The force-heat coupling density refers to the density of the interaction between force and heat during the welding process, reflecting the joint influence of heat input and mechanical contact force on the welding quality of the welding point. In the implementation process, the coupling effect between heat and contact force in the stress area of ​​the solder joint is first calculated based on the instantaneous thermal energy intensity data and contact force data of the solder joint. Through numerical analysis, the coupled thermodynamic model is used to calculate the force-heat coupling density of each solder joint area. This model involves a comprehensive analysis of factors such as heat conduction, thermal expansion, and mechanical stress. The finite element analysis method (FEA) is usually used to numerically solve the coupling process of force and heat to obtain the force-heat coupling density data. This data provides a basis for subsequent welding path optimization and welding contact force regulation. The CFL condition (Courant-Friedrichs-Lewy condition) is used to analyze the incremental relationship of the solder joint force-heat coupling density data. The CFL condition is a stability condition in numerical calculation, which is usually used to solve partial differential equations involving time and space transformations. In the welding process, due to the complex interaction between force and heat, the force-heat coupling density of the solder joint must be discretized in time and space to ensure the stability of the solution process. In the specific implementation, the solder joint force-heat coupling density data is first discretized in time and space, and the incremental relationship of the force-heat density at each moment and each spatial position is calculated. Through the CFL condition, the calculation step size and spatial grid of each solder joint area are adjusted to ensure the stability of the numerical calculation. Through the analysis of this incremental relationship, the change law of the force-heat density condition of the solder joint is obtained, which further provides data support for the thermal conduction simulation and the calculation of the toughness loss of the solder joint. The force-heat coupling density data of the solder joint is calculated by the conditional two-dimensional thermal equation using the incremental relationship of the force-heat density condition. The two-dimensional thermal equation is an important mathematical model to describe the heat conduction process, which is usually solved by the finite difference method or the finite element method. In the implementation process, the force-heat density incremental relationship is first converted into a two-dimensional heat conduction equation and processed by mathematical formulas.The specific steps include determining the heat source distribution in the solder joint area, calculating the temperature change of each point, and applying the heat conduction equation to obtain the temperature field distribution of each solder joint and its surrounding area. In this process, the force-heat coupling density of the solder joint plays a constraining role, limiting the range of change of the temperature field and ensuring the rationality and accuracy of the simulation results; specifically, it describes a quantity of interaction between force and heat during welding, affecting the temperature distribution and mechanical properties of the solder joint. Its calculation is usually carried out through numerical methods, combining the coupling effect of mechanical force and heat input; the final force-heat density constrained two-dimensional thermal equation is:. in, is the temperature at position (x,y) and time t, , represents thermal diffusivity, which indicates the ability of a material to conduct heat, where k is thermal conductivity and ρ is material density. is the specific heat capacity; and It represents the second-order derivative of temperature with respect to the spatial coordinates x and y, and describes the spatial rate of change of the temperature field in the x and y directions. represents the contact force at position (x, y), which represents the force applied by the manipulator during the welding process. represents the local temperature at the position (x, y), usually the temperature of the welding spot area, and λ represents the force-heat coupling coefficient, which represents the coupling strength between mechanical force and thermal energy. This coefficient is usually obtained through historical experiments or theoretical models, reflecting the influence of mechanical contact force on heat transfer. First, the body material data is obtained, including the thermophysical properties (such as thermal conductivity, specific heat capacity, etc.) and mechanical performance data of various parts of the body. Then, the thermal conduction toughness loss of the body material and the robot welding path trajectory is simulated by combining the two-dimensional heat equation constrained by force-heat density. In the specific implementation, the heat transfer and toughness loss in the welding process are simulated according to the material data of different parts of the body. During the welding process, the instantaneous heat input of the welding spot is closely related to the thermal properties of the material, and these heat conductions will affect the toughness of the welding spot area. By combining the heat conduction model with the material data, the thermal-mechanical coupling model is used to calculate the heat conduction process between adjacent welding spots to obtain the toughness loss data of the welding spot. This simulation needs to consider factors such as temperature gradient and stress field changes, and then predict the toughness reduction of each welding spot caused by heat conduction, providing a basis for subsequent welding path optimization and mechanical compensation strategies.

[0092] Step S24 includes the following steps:

[0093] Step S241: performing multi-point spatial thermal field difference analysis on the heat conduction solder joint toughness loss data to obtain multi-point thermal field difference data;

[0094] Step S242: fitting the numerical law of thermal stress alternation characteristics between adjacent solder joints to the thermal conduction solder joint toughness loss data based on the multi-point thermal field difference data to obtain fitting data of the numerical law of thermal stress alternation;

[0095] Step S243: performing heat capacity transfer concentration range measurement on the thermal stress alternating numerical law fitting data to obtain heat capacity transfer concentration range measurement data;

[0096] Step S244: identifying the thermal chain reaction toughness influence gradient data between adjacent solder joints according to the thermal stress alternation numerical law fitting data and the heat capacity transfer concentration range measurement data, and obtaining the thermal chain reaction toughness influence gradient data.

[0097] In an embodiment of the present invention, firstly, the temperature field difference caused by heat conduction between different solder joints is analyzed from the obtained heat conduction solder joint toughness loss data. Through the spatial thermal field difference analysis, the heat distribution difference of each solder joint and its surrounding area is further quantified to form thermal field difference data. In the implementation process, the temperature field modeling of multiple solder joints on the welding path is first performed, and the heat source distribution and heat flow direction of each solder joint area are assumed. Through the numerical simulation method, the heat conduction equation (such as a two-dimensional or three-dimensional heat conduction equation) is used for calculation to obtain the temperature distribution of different solder joints. Then, the thermal field difference between each solder joint is calculated, specifically, the temperature difference between adjacent solder joints and the temperature gradient change between the solder joint and the surrounding area are compared. Through these data, multi-point thermal field difference data is formed for subsequent thermal stress analysis. Combined with the multi-point thermal field difference data and the solder joint toughness loss data, the alternating characteristics of thermal stress between adjacent solder joints are analyzed. During the welding process, the temperature field change caused by the heat input will cause the material to expand or contract thermally, thereby generating thermal stress. Thermal stress has an alternating characteristic during the welding process, which changes with time and space. In the specific implementation, the multi-point thermal field difference data is first combined with the mechanical properties of the solder joint (such as the thermal expansion coefficient and elastic modulus of the material) to calculate the thermal stress generated by each solder joint during the welding process. Then, the curve fitting or regression analysis method is used to fit the numerical law of the thermal stress alternation characteristics of adjacent solder joints. Through numerical optimization algorithms such as the least squares method, the law of thermal stress variation with time and space, that is, the numerical law of thermal stress alternation, is obtained. This law describes the change of thermal stress between different solder joints, and provides data support for the subsequent heat capacity transfer analysis and thermal chain reaction toughness influence gradient calculation. Based on the data fitting of the numerical law of thermal stress alternation, the heat capacity transfer aggregation range is measured. The heat capacity transfer aggregation range refers to the spatial range of heat transfer during welding, which usually affects the thermal effect and stress distribution in the area around the solder joint. In the implementation, firstly, based on the fitted thermal stress alternation law, the propagation range of thermal stress in different areas is calculated. By analyzing the relationship between thermal stress and heat capacity, the heat capacity transfer model is used to calculate the thermal energy distribution around the solder joint. The heat capacity transfer aggregation range measurement can be achieved by the integral method or the spatial weighted average method to calculate the propagation range of thermal energy and its influence. For example, the welding point is selected as the center, and the spatial distribution of the thermal stress alternation data is considered to calculate the degree of aggregation of heat capacity transfer in the area. Through this process, the heat capacity transfer aggregation range measurement data can be obtained, which describes how heat is accumulated and diffused in the welding area, thereby affecting the toughness of the surrounding area. Combining the thermal stress alternation numerical law fitting data and the heat capacity transfer aggregation range measurement data, the thermal chain reaction toughness influence gradient between adjacent welding points is identified. The thermal chain reaction toughness influence gradient refers to the gradient of toughness change caused by the interaction between thermal stress and heat capacity transfer between adjacent welding points during the welding process.In the specific implementation, the numerical law of thermal stress alternation is first combined with the measurement data of the heat capacity transfer aggregation range to calculate the thermal chain reaction between solder joints (it means that during the welding process, the thermal influence of one solder joint will affect the toughness of adjacent solder joints in a chain reaction manner. By identifying the gradient data, the degree and change trend of the thermal influence can be understood more accurately.). According to the distribution of thermal stress and heat capacity, numerical analysis methods such as finite difference method or finite element method are used to simulate the influence of thermal stress and heat capacity transfer on solder joint toughness. By performing gradient identification on the thermal chain reaction of each solder joint in the welding area, the gradient data of the thermal chain reaction toughness influence is finally obtained. This data describes the change in toughness between adjacent solder joints due to the interaction of heat conduction, thermal stress and heat capacity during welding, and provides an important basis for subsequent welding path optimization and error compensation strategies.

[0098] Step S242 includes the following steps:

[0099] Perform spatial coordinate distribution analysis on multi-point thermal field difference data to generate thermal field data for each soldering point; perform time series fluctuation analysis on the thermal field data for each soldering point to obtain thermal field time series fluctuation data for each soldering point;

[0100] Based on the thermal field time series fluctuation data of each solder joint, the thermal energy-toughness response time delay variance between adjacent solder joints is calculated for the toughness loss data of the heat conduction solder joint, and the thermal energy-toughness response time delay variance is obtained;

[0101] The thermal energy-toughness fluctuation extreme value derivative difference calculation is performed according to the thermal energy-toughness response delay variance to obtain the thermal energy-toughness fluctuation extreme value difference derivative;

[0102] Based on the extreme differential derivative of thermal energy-toughness fluctuation, thermal energy boundary-limited cubic spline fitting is performed on the thermal field time series fluctuation data of each solder joint to obtain thermal energy-limited cubic spline fitting data;

[0103] Based on the thermal energy limited cubic spline fitting data, the numerical law fitting of the thermal stress alternation characteristics between adjacent solder joints is performed to obtain the fitting data of the numerical law of thermal stress alternation.

[0104] In an embodiment of the present invention, a spatial coordinate distribution analysis is performed on the multi-point thermal field difference data to generate the thermal field data of each solder joint. This step obtains the temperature field information of different solder joint positions by analyzing the temperature distribution data of multiple solder joints. In the specific operation, each solder joint on the welding path is selected, the temperature data of its adjacent area is obtained, and it is mapped into a three-dimensional coordinate system to facilitate the analysis of the thermal distribution difference between each solder joint during the welding process. Using the heat conduction equation, the thermal field data of each solder joint position is obtained by numerical calculation, and these data are expressed as the spatial distribution of the temperature field around different solder joints. Through the spatial coordinate distribution analysis, the thermal field distribution data of different solder joints are visualized to help understand the heat transfer and change trend during the welding process. Next, the thermal field data of each solder joint is analyzed in time series to obtain the thermal field time series fluctuation data of each solder joint. This step analyzes the change of the thermal field over time during the welding process. By performing time series analysis on the thermal field data of the solder joint, the temperature fluctuation of each solder joint during the welding process is recorded, and then the time series fluctuation characteristics of the solder joint are obtained. During implementation, the temperature data of each solder joint is first arranged in chronological order to form a time series. Next, the time series is subjected to volatility analysis. Statistical methods such as standard deviation and variance can be used to quantify the fluctuation amplitude and fluctuation law of the thermal field of the solder joint. Through this step, the thermal field time series fluctuation data of each solder joint is obtained, which provides a basis for the subsequent response delay analysis. Then, based on the thermal field time series fluctuation data of each solder joint, the thermal energy-toughness response delay variance between adjacent solder joints is calculated for the toughness loss data of the heat conduction solder joint, and the thermal energy-toughness response delay variance is obtained. The purpose of this step is to quantify the response relationship between thermal energy change and solder joint toughness during welding and its delay effect. In specific implementation, based on the time series fluctuation data of the thermal field and the toughness loss data, the time series correlation analysis method is used to calculate the response delay between thermal energy change and toughness loss. By calculating the time delay variance between the thermal energy change and the toughness change between each pair of adjacent solder joints, the variance data of the thermal energy and toughness response delay is obtained. This calculation reflects the hysteresis effect of thermal energy change on solder joint toughness and its range of variation, and further provides the necessary statistical data for subsequent fluctuation analysis and fitting. Next, the extreme derivative of thermal energy-toughness fluctuation is differentially calculated based on the variance of thermal energy-toughness response delay to obtain the extreme differential derivative of thermal energy-toughness fluctuation. This step aims to further analyze the fluctuation characteristics between thermal energy and toughness and quantify their rate of change. In the specific operation, the extreme points of fluctuation between thermal energy and toughness are first identified, and then the derivative of each extreme point of fluctuation is calculated to analyze its rate of change. Through the differential calculation method, the extreme derivative of thermal energy-toughness fluctuation is obtained, thereby revealing the dynamic change trend between thermal energy and toughness fluctuations during welding. Extreme derivative differential helps to identify the key time points of toughness change in the weld and reveal the nonlinear relationship between thermal energy and toughness loss during heat conduction.Then, based on the extreme differential derivative of thermal energy-toughness fluctuation, the thermal field time series fluctuation data of each solder joint is fitted with a thermal energy boundary-limited cubic spline to obtain thermal energy-limited cubic spline fitting data. This step uses spline fitting technology to smooth the thermal energy fluctuation data by limiting boundary conditions, so as to accurately describe the distribution and fluctuation of thermal energy between different solder joints. First, according to the extreme derivative of thermal energy-toughness fluctuation obtained by the above calculation, the boundary range of thermal energy data is determined. Then, the cubic spline interpolation method is used to fit the thermal field time series fluctuation data of each solder joint under the boundary condition. The cubic spline fitting method generates a smooth fitting curve by minimizing the fitting error, which reflects the thermal energy change trend of the solder joint at different time points. Through the fitting result, the thermal energy fluctuation characteristics of the solder joint during the welding process can be more accurately described. Finally, based on the thermal energy-limited cubic spline fitting data, the numerical law of the thermal stress alternation characteristics between adjacent solder joints is fitted to obtain the numerical law fitting data of the thermal stress alternation. The purpose of this step is to derive the alternation law of thermal stress during welding by fitting the thermal energy fluctuation data. By further processing the thermal energy limited cubic spline fitting data, the numerical fitting method (such as least squares method or nonlinear regression analysis) is used to calculate the numerical law of the change of thermal stress between welds over time. This law describes the change characteristics of thermal stress during welding under different welding conditions, including the amplitude and frequency of thermal stress. In this way, the alternating characteristics of thermal stress during welding can be accurately simulated and predicted, providing reliable data support for subsequent heat conduction and toughness analysis.

[0105] Step S3 includes the following steps:

[0106] Step S31: performing regional gridding processing on the thermal chain reaction toughness influence gradient data to obtain the thermal chain reaction toughness influence grid data;

[0107] Step S32: performing logic control of the manipulator contact force on the manipulator theoretical contact force data according to the hot chain reaction toughness influence grid data to obtain the welding contact force logic control data;

[0108] Step S33: dynamically adjusting the welding angle according to the hot chain reaction toughness influence grid data and the welding contact force logic control data to obtain the welding angle dynamic adjustment data;

[0109] Step S34: Based on the welding contact force logic control data and the welding angle dynamic adjustment data, the terminal position and posture error compensation strategy is learned to obtain the manipulator terminal position and posture error compensation strategy.

[0110] As an example of the present invention, refer to Figure 3 As shown, in this example, step S3 includes:

[0111] Step S31: performing regional gridding processing on the thermal chain reaction toughness influence gradient data to obtain the thermal chain reaction toughness influence grid data;

[0112] In an embodiment of the present invention, the hot chain reaction toughness influence gradient data is subjected to regional gridding processing to obtain hot chain reaction toughness influence grid data. In this step, the welding area is spatially discretized according to the hot chain reaction toughness influence gradient data. Specifically, the geometric shape of the welding workpiece is first divided into several small areas, which correspond to different welding points of the vehicle body. The hot chain reaction toughness influence gradient data corresponding to each area will be converted into a set of discrete grid data points through a gridding method. These grid data points represent the changes in the hot chain reaction toughness influence at different positions. During the operation, a suitable grid size is selected so that the hot chain reaction toughness influence of each weld point and its surroundings can be accurately captured and provide the necessary spatial resolution in subsequent analysis. This step uses a gridding technique such as a finite difference method to convert the hot chain reaction toughness gradient in the welding area into a two-dimensional or three-dimensional discrete grid, and the generated grid data is the hot chain reaction toughness influence grid data. In another embodiment, the hot chain reaction toughness influence gradient data is subjected to regional gridding processing. First, it is necessary to construct a three-dimensional spatial distribution model of the welding area based on the three-dimensional coordinate information of the spatial distribution of the weld points and the gradient data of the thermal reaction between the weld points. Using the heat conduction data between the solder joints, the grid of the welding area is divided according to the distance between the solder joints and the law of thermal gradient change. Specifically, a grid division method based on the thermal gradient change rate is adopted, and the areas with large thermal gradients or more drastic changes are set as high-density grids, and the areas with small thermal gradient changes are set as low-density grids to ensure that the fineness of the gridding is consistent with the thermal gradient change characteristics. In the process of grid division, multi-point interpolation calculations are required for the thermal conduction characteristics of the solder joints, the thermal gradient value of the grid boundary is determined by the interpolation method, and the node distribution of the grid boundary is refined by the linear interpolation method. Finally, the hot chain reaction toughness influence grid data is generated, including the thermal gradient value of each grid node, the correlation relationship between nodes, and the adjacency matrix.

[0113] Step S32: performing logic control of the manipulator contact force on the manipulator theoretical contact force data according to the hot chain reaction toughness influence grid data to obtain the welding contact force logic control data;

[0114] In an embodiment of the present invention, the contact force logic control of the manipulator is performed on the theoretical contact force data of the manipulator according to the grid data affected by the toughness of the thermal chain reaction. First, the node thermal gradient is statistically analyzed for the grid data affected by the toughness of the thermal chain reaction, and the thermal conduction intensity and the cumulative thermal stress level of each grid node are calculated. Then, combined with the theoretical contact force data of the manipulator, the response relationship between the contact force distribution of the manipulator at each welding point and the thermal gradient of the grid node is calculated. Specifically, the thermal gradient data of the grid node is normalized by the force-heat coupling analysis method to obtain the standardized thermal sensitivity coefficient of the node. By establishing a contact force control equation, the thermal sensitivity coefficient of each grid node is substituted into the equation to calculate the contact force adjustment value of the manipulator at different nodes. The core logic of the equation is: when the thermal gradient value of a node is large, the contact force of the manipulator needs to be reduced to a certain extent to avoid excessive contact force causing overheating or toughness loss of the welding point; when the thermal gradient value of a node is small, the contact force is appropriately increased to improve the stability and reliability of welding. Finally, according to the node contact force adjustment value, the welding contact force logic control data is generated, which includes the contact force adjustment parameters of each welding point and the corresponding control logic.

[0115] Step S33: dynamically adjusting the welding angle according to the hot chain reaction toughness influence grid data and the welding contact force logic control data to obtain the welding angle dynamic adjustment data;

[0116] In an embodiment of the present invention, the welding angle is dynamically adjusted according to the grid data affected by the toughness of the thermal chain reaction and the logic control data of the welding contact force. First, based on the grid data affected by the toughness of the thermal chain reaction, the welding angle requirement of each grid node is extracted. Specifically, the initial value of the welding angle is calculated using the thermal gradient direction and thermal stress distribution characteristics of each welding point node. The calculation formula of the initial welding angle is: the angle between the normal vector of the thermal gradient direction of the welding point and the normal vector of the contact surface of the welding tool at the end of the manipulator. Subsequently, according to the logic control data of the welding contact force, the initial welding angle is dynamically adjusted to ensure that the welding angle can adapt to the change of the thermal stress distribution of the welding point. The angle adjustment method adopts a welding angle optimization algorithm based on dynamic response, and the specific steps are: parsing the contact force adjustment parameters of each welding point, combining the thermal gradient direction of the welding point, calculating the optimal welding angle change value of the welding point; superimposing the optimal welding angle change value with the initial welding angle to obtain the final welding angle dynamic adjustment value. Finally, the welding angle dynamic adjustment data is generated, which includes the optimal welding angle value of each welding point and its dynamic adjustment change trend.

[0117] Step S34: Based on the welding contact force logic control data and the welding angle dynamic adjustment data, the terminal position and posture error compensation strategy is learned to obtain the manipulator terminal position and posture error compensation strategy.

[0118] In an embodiment of the present invention, the terminal posture error compensation strategy is learned based on the welding contact force logic control data and the welding angle dynamic adjustment data. First, according to the welding contact force logic control data, the contact force adjustment parameters of the manipulator at different welding points are extracted, and the welding angle change value in the welding angle dynamic adjustment data is combined to establish a mathematical model for posture error compensation. The model uses the theoretical posture of the end of the manipulator as a reference to analyze the influence of the contact force change and the welding angle change on the actual posture of the end. The specific calculation steps are as follows: the contact force adjustment value and the welding angle adjustment value of each welding point are weighted averaged to obtain the comprehensive influence value of the force and heat on the end of the manipulator; the dynamic equation of the manipulator is used to calculate the real-time offset of the end posture caused by the force and heat; wherein, in the dynamic analysis of the manipulator, the Lagrange equation or the Newton-Euler equation is usually used to describe the mechanical system. Assume that the Newton-Euler equation is used to establish a dynamic model, which is a commonly used method and is particularly suitable for the motion analysis of multi-degree-of-freedom manipulators. The dynamic equation of the manipulator can be described in the following form: in, Indicates the real-time offset, Represents the position vector or joint angle vector (usually the joint angle) of the end effector of the manipulator; Represents the acceleration vector of the joint; Represents the mass matrix of the manipulator, describing the inertia of the manipulator joints; Table Coriolis force matrix, taking into account the influence of joint velocity on the system; is the gravity matrix, describing the effect of gravity on the manipulator; It is the torque vector applied by external heating, which is generated here by the change of contact force, welding angle and welding heat value.

[0119] Through three-dimensional coordinate transformation, the offset is converted into a spatial vector of the terminal posture error. Then, the posture error compensation strategy is trained using an error compensation strategy optimization algorithm based on gradient descent. The optimization algorithm aims to minimize the terminal posture error and gradually approaches the optimal compensation strategy by iteratively adjusting the compensation strategy parameters. Finally, the terminal posture error compensation strategy of the manipulator is obtained, which includes the posture error compensation parameters of each welding point and its mapping relationship with the contact force and welding angle, which is used to guide the terminal position and posture adjustment of the manipulator.

[0120] Step S32 includes the following steps:

[0121] Step S321: Derivation and measurement of the linear relationship of the theoretical contact force data of the manipulator according to the grid data of the thermal chain reaction toughness influence, to obtain the contact force thermal chain toughness reaction relationship derivation data between different grids;

[0122] Step S322: performing heat source toughness attenuation diffusion position anisotropy analysis on the contact force thermal chain toughness response relationship derivation data between different grids to obtain heat source toughness diffusion position anisotropy data;

[0123] Step S323: based on the different data of heat source toughness diffusion orientation, the contact force heat chain toughness reaction relationship derivation data between different grids are used to locate the reaction concentration point, and the toughness trend abnormal increment of the reaction concentration point is identified to obtain the toughness trend abnormal increment data of the concentration point;

[0124] Step S324: Based on the abnormal incremental data of the toughness trend at the concentrated point, the contact force of the robot is logically controlled to obtain the welding contact force logical control data.

[0125] In an embodiment of the present invention, the theoretical contact force data of the manipulator is deduced and measured according to the grid data of the hot chain reaction toughness influence, so as to obtain the contact force hot chain toughness reaction relationship derivation data between different grids. During implementation, the hot chain reaction toughness influence data and the manipulator contact force data in each grid are first statistically analyzed. Using the linear regression model, the linear relationship between the two is derived by calculating the correlation between the change of the hot chain reaction toughness in the welding area and the contact force of the manipulator. Specifically, the hot chain reaction toughness value of each welding point in the grid and the corresponding contact force data are selected, and regression analysis is performed by the least squares method to obtain the linear relationship equation between the hot chain toughness reaction and the contact force. This derivation process helps to quantify the dependence between the contact force and the hot chain reaction toughness between different grids, and obtains the contact force hot chain toughness reaction relationship derivation data. Next, the heat source toughness attenuation diffusion azimuth anisotropy analysis is performed on the contact force hot chain toughness reaction relationship derivation data between different grids, and the heat source toughness diffusion azimuth different data is obtained. In this step, the azimuth difference of the heat source toughness attenuation is identified by analyzing the distribution of the heat source in different grid areas and the direction of heat energy transfer. First, the heat source position, heat energy attenuation rate and contact force data are spatially analyzed, and combined with physical models (such as heat conduction equations), the diffusion path and thermal attenuation characteristics of the heat source in space are identified. For each grid, the degree of heat energy attenuation from its center point to the surrounding area is calculated, and then the diffusion characteristics of the heat source in different directions are obtained. Specifically, the Fourier transform method is used to analyze the propagation wavelength of the heat source, and the degree of diffusion in different directions is evaluated in combination with the toughness data of the thermal chain reaction. Finally, the data of the different directions of the heat source toughness diffusion are obtained, reflecting the characteristics of the heat source attenuation in different directions. Then, based on the different directions of the heat source toughness diffusion data, the contact force thermal chain toughness reaction relationship derivation data between different grids are used to locate the reaction concentration point, and the abnormal increment of the toughness trend of the reaction concentration point is identified to obtain the abnormal increment data of the toughness trend of the concentration point. The purpose of this step is to identify the concentrated areas where there are abnormal changes in the thermal chain reaction. During implementation, the concentration points of the thermal chain reaction in each grid are determined by analyzing the results of the heat source toughness diffusion direction and combining the contact force thermal chain toughness reaction relationship derivation data. These concentration points refer to areas where the toughness changes of the heat source are relatively large. By further calculating the trend of changes in the toughness of the hot chain near the concentration point, the difference algorithm is used to perform incremental analysis on the toughness changes, and the abnormal toughness increments that appear in certain areas are identified. Specifically, statistical methods, such as standard deviation calculation, are used to evaluate the toughness fluctuations in a certain area, and abnormal points are identified through incremental analysis. Finally, the abnormal incremental data of the toughness trend at the concentration point is obtained for subsequent contact force adjustment. Finally, the contact force of the manipulator is logically controlled based on the abnormal incremental data of the toughness trend at the concentration point to obtain the logical control data of the welding contact force. The purpose of this step is to adjust the contact force of the manipulator to cope with the toughness changes in the hot chain reaction.During implementation, the contact force of the manipulator is dynamically adjusted according to the abnormal increment data of the toughness trend at the concentrated point. Specifically, in areas where there are abnormal increments of toughness, the contact force of the manipulator is appropriately increased or decreased to ensure that the heat transfer and mechanical stress distribution during welding are more uniform. The control process uses a weighted average algorithm to adjust the size of the contact force according to the degree of abnormal increment. For each welding point, the contact force distribution of the manipulator is adjusted in real time through a logical control algorithm to adapt to the thermal chain reaction conditions in different areas. Finally, the welding contact force logic control data is obtained for use in mechanical control in subsequent welding processes.

[0126] Step S33 includes the following steps:

[0127] Step S331: performing solder joint blunt edge thickness matching according to the hot chain reaction toughness influence grid data to obtain toughness associated solder joint blunt edge thickness matching data;

[0128] Step S332: dynamically adjusting the feed angle based on the welding contact force logic control data and the toughness-related weld blunt edge thickness matching data to obtain welding feed angle dynamic adjustment data;

[0129] Step S333: adjusting intermittent penetration welding according to the toughness-associated weld blunt edge thickness matching data and the welding feed angle dynamic adjustment data to obtain intermittent penetration welding adjustment data;

[0130] Step S334: dynamically adjust the welding angle based on the toughness-related weld blunt edge thickness matching data, the welding feed angle dynamic adjustment data, and the intermittent penetration welding adjustment data to obtain the welding angle dynamic adjustment data.

[0131] In an embodiment of the present invention, the blunt edge thickness of the weld is matched according to the hot chain reaction toughness influence grid data to obtain the toughness-related weld blunt edge thickness matching data. When implemented, the hot chain reaction toughness data of each weld is first processed to identify the blunt edge area related to the change of hot chain toughness. In this process, the known weld geometric parameters (such as the size and shape of the weld) are used in combination with the heat conduction model to calculate the hot chain reaction toughness influence area around each weld. Then, by analyzing the blunt edge thickness of these areas, the blunt edge matching algorithm (such as a matching algorithm based on similarity) is used to reasonably match the blunt edge thickness of the weld, thereby obtaining the toughness-related weld blunt edge thickness matching data. This data reflects the relationship between the hot chain toughness characteristics of different welds and their blunt edge thickness, providing a basis for subsequent welding adjustments. Next, the feed angle is dynamically adjusted based on the welding contact force logic control data and the toughness-related weld blunt edge thickness matching data to obtain the welding feed angle dynamic adjustment data. When implemented, first, by analyzing the welding contact force data and combining the hot chain reaction toughness characteristics of the weld, the change trend of welding quality under different feed angles is calculated. The adjustment of the feed angle depends on the matching result of the blunt edge thickness and the contact force. Specifically, the weighted average method is used to dynamically adjust the feed angle according to the matching data of the blunt edge thickness of the toughness-related weld and the logical control data of the welding contact force. During the adjustment process, the welding feed angle is continuously adjusted by an optimization algorithm (such as the gradient descent method) to ensure that the contact pressure and heat conduction effect during the welding process are optimally balanced, thereby obtaining the dynamic adjustment data of the welding feed angle. Then, intermittent penetration welding adjustment is performed according to the matching data of the blunt edge thickness of the toughness-related weld and the dynamic adjustment data of the welding feed angle, thereby obtaining the intermittent penetration welding adjustment data. During implementation, the penetration under different welding parameters is evaluated based on the previously calculated relationship between the feed angle and the blunt edge thickness. For the penetration during the welding process, a time series analysis is first performed to identify the variation law of the penetration, and the penetration is intermittently adjusted by the set parameters (such as time, temperature, feed speed, etc.). A step-by-step adjustment strategy is adopted to implement different welding intensities and time intervals at each weld point to control the variation of the penetration. During the adjustment process, numerical simulation methods (such as the finite element method) are used to accurately calculate the changes in penetration depth, and the welding parameters are optimized in real time according to the dynamic adjustment data of the blunt edge thickness and the feed angle to ensure that the penetration depth during welding meets the needs of different areas. Finally, the welding angle is dynamically adjusted based on the matching data of the blunt edge thickness of the toughness-related weld, the dynamic adjustment data of the welding feed angle, and the intermittent penetration welding adjustment data to obtain the dynamic adjustment data of the welding angle. During implementation, the welding feed angle and the penetration adjustment results are first combined to calculate the change trend of the welding angle of each welding point. The basis for adjusting the welding angle includes the geometry of the welding point, the change in toughness of the hot chain reaction, the blunt edge thickness, and the dynamic adjustment data of the feed angle.In specific operations, the welding angle is dynamically adjusted under the premise of ensuring welding quality through optimization algorithms (such as genetic algorithms). The adjustment of the welding angle not only takes into account the requirements of the local welding points, but also the overall heat conduction during the welding process. Finally, through algorithm adjustment and optimization, the dynamic adjustment data of the welding angle is obtained to ensure that the angle distribution during the entire welding process can match the changes in the toughness of the thermal chain reaction and the welding contact strength.

[0132] The present invention also provides a five-axis hybrid robot terminal position and posture error compensation system, which is used to execute the five-axis hybrid robot terminal position and posture error compensation method as described above. The five-axis hybrid robot terminal position and posture error compensation system comprises:

[0133] The welding strength difference analysis module is used to obtain the body welding demand design data; the welding strength analysis between different welding points is performed on the body welding demand design data to obtain the welding strength difference data between different welding points;

[0134] The module for identifying the impact of thermal chain reaction toughness is used to match the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points, and obtain the theoretical contact force data of the manipulator; to simulate the thermal conduction toughness loss between adjacent welding points based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, and obtain the toughness loss data of the thermal conduction welding points; to identify the thermal chain reaction toughness impact gradient data between adjacent welding points on the toughness loss data of the thermal conduction welding points, and obtain the thermal chain reaction toughness impact gradient data;

[0135] The terminal posture error compensation strategy learning module is used to perform regional grid processing on the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform the logic control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logic control data of the welding contact force; and perform the terminal posture error compensation strategy learning based on the logic control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator;

[0136] The automated firmware design module is used to perform automated firmware design based on the manipulator end position and posture error compensation strategy, obtain the manipulator end position and posture error compensation firmware, and embed the manipulator end position and posture error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position and posture error compensation.

[0137] Therefore, the embodiments should be regarded as illustrative and non-restrictive from all points, and the scope of the present invention is limited by the appended claims rather than the above description, and it is therefore intended that all changes falling within the meaning and range of equivalent elements of the application documents are included in the present invention.

[0138] The above description is only a specific embodiment of the present invention, so that those skilled in the art can understand or implement the present invention. Various modifications to these embodiments will be apparent to those skilled in the art, and the general principles defined herein may be implemented in other embodiments without departing from the spirit or scope of the present invention. Therefore, the present invention will not be limited to the embodiments shown herein, but should conform to the widest scope consistent with the principles and novel features invented herein.

Claims

1. A method for compensating the terminal position error of a five-axis hybrid robot, characterized in that: The following steps are involved: Step S1: obtaining body welding demand design data; performing welding strength analysis between different welding points on the body welding demand design data to obtain welding strength difference data between different welding points; Step S2: Based on the welding strength difference data between different welding spots, the theoretical contact force of the end manipulator of the five-axis hybrid robot is matched to obtain the theoretical contact force data of the manipulator; based on the welding strength difference data between different welding spots and the theoretical contact force data of the manipulator, the thermal conduction toughness loss between adjacent welding spots is simulated to obtain the thermal conduction welding spot toughness loss data; the thermal chain reaction toughness influence gradient data between adjacent welding spots is identified for the thermal conduction welding spot toughness loss data to obtain the thermal chain reaction toughness influence gradient data; Step S3: performing regional gridding processing on the thermal chain reaction toughness influence gradient data to obtain the thermal chain reaction toughness influence grid data; According to the hot chain reaction toughness influence grid data, the manipulator contact force logic control is performed on the manipulator theoretical contact force data to obtain the welding contact force logic control data; Based on the welding contact force logic control data, the terminal posture error compensation strategy is learned to obtain the terminal posture error compensation strategy of the manipulator; Step S4: Based on the manipulator end position error compensation strategy, perform automated firmware design to obtain the manipulator end position error compensation firmware, and embed the manipulator end position error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position error compensation.

2. The five-axis hybrid robot terminal position error compensation method according to claim 1 is characterized in that: Step S1 includes the following steps: Step S11: Obtaining vehicle body welding requirement design data; Step S12: performing welding point sequence identification processing on the vehicle body welding requirement design data to obtain vehicle body welding point sequence identification data; Step S13: analyzing the welding strength between different welding points on the vehicle body welding requirement design data to obtain welding strength difference data between different welding points.

3. The five-axis hybrid robot terminal position error compensation method according to claim 2 is characterized in that: Step S2 includes the following steps: Step S21: matching the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points to obtain the theoretical contact force data of the manipulator; Step S22: extracting the welding path trajectory of the end manipulator of the five-axis hybrid robot according to the sequential identification data of the welding points of the vehicle body, and obtaining the welding path trajectory of the manipulator; Step S23: based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, the manipulator welding path trajectory is simulated for the thermal conduction toughness loss between adjacent welding points to obtain the thermal conduction welding point toughness loss data; Step S24: identifying the thermal chain reaction toughness influence gradient data between adjacent solder joints on the thermal conduction solder joint toughness loss data to obtain the thermal chain reaction toughness influence gradient data.

4. The five-axis hybrid robot terminal position error compensation method according to claim 3 is characterized in that: Step S23 includes the following steps: Step S231: performing instantaneous heat energy intensity analysis of the weld points based on the welding strength difference data between different weld points, and obtaining instantaneous heat energy intensity data of the weld points between different weld points; Step S232: performing force-heat coupling density analysis according to the instantaneous heat energy intensity data of the soldering point and the theoretical contact force data of the manipulator to obtain the force-heat coupling density data of the soldering point; Step S233: performing CFL conditional incremental relationship analysis on the solder joint force-heat coupling density data to obtain a force-heat density conditional incremental relationship; Step S234: performing conditional constraint two-dimensional thermal equation calculation on the force-heat coupling density data of the solder joint according to the force-heat density condition increment relationship to obtain a force-heat density constraint two-dimensional thermal equation; Step S235: Acquire body material data; simulate the thermal conduction toughness loss between adjacent welding points based on the body material data and the robot welding path trajectory based on the force-heat density constraint two-dimensional thermal equation to obtain thermal conduction welding point toughness loss data.

5. The five-axis hybrid robot terminal position error compensation method according to claim 1 is characterized in that: Step S3 includes the following steps: Step S31: performing regional gridding processing on the thermal chain reaction toughness influence gradient data to obtain the thermal chain reaction toughness influence grid data; Step S32: performing logic control of the manipulator contact force on the manipulator theoretical contact force data according to the hot chain reaction toughness influence grid data to obtain the welding contact force logic control data; Step S33: dynamically adjusting the welding angle according to the hot chain reaction toughness influence grid data and the welding contact force logic control data to obtain the welding angle dynamic adjustment data; Step S34: Based on the welding contact force logic control data and the welding angle dynamic adjustment data, the terminal position and posture error compensation strategy is learned to obtain the manipulator terminal position and posture error compensation strategy.

6. The five-axis hybrid robot terminal position error compensation method according to claim 5 is characterized in that: Step S32 includes the following steps: Step S321: Derivation and measurement of the linear relationship of the theoretical contact force data of the manipulator according to the grid data of the thermal chain reaction toughness influence, to obtain the contact force thermal chain toughness reaction relationship derivation data between different grids; Step S322: performing heat source toughness attenuation diffusion position anisotropy analysis on the contact force thermal chain toughness response relationship derivation data between different grids to obtain heat source toughness diffusion position anisotropy data; Step S323: based on the different data of heat source toughness diffusion orientation, the contact force heat chain toughness reaction relationship derivation data between different grids are used to locate the reaction concentration point, and the toughness trend abnormal increment of the reaction concentration point is identified to obtain the toughness trend abnormal increment data of the concentration point; Step S324: Perform logic control of the robot contact force based on the abnormal incremental data of the toughness trend of the concentrated point to obtain logic control data of the welding contact force.

7. A five-axis hybrid robot terminal position error compensation system, characterized in that: Used to execute the five-axis hybrid robot terminal position and posture error compensation method as claimed in claim 1, the five-axis hybrid robot terminal position and posture error compensation system comprises: The welding strength difference analysis module is used to obtain the body welding demand design data; the welding strength analysis between different welding points is performed on the body welding demand design data to obtain the welding strength difference data between different welding points; The module for identifying the impact of thermal chain reaction toughness is used to match the theoretical contact force of the end manipulator of the five-axis hybrid robot based on the welding strength difference data between different welding points, and obtain the theoretical contact force data of the manipulator; to simulate the thermal conduction toughness loss between adjacent welding points based on the welding strength difference data between different welding points and the theoretical contact force data of the manipulator, and obtain the toughness loss data of the thermal conduction welding points; to identify the thermal chain reaction toughness impact gradient data between adjacent welding points on the toughness loss data of the thermal conduction welding points, and obtain the thermal chain reaction toughness impact gradient data; The terminal posture error compensation strategy learning module is used to perform regional grid processing on the gradient data of the toughness influence of the thermal chain reaction to obtain the grid data of the toughness influence of the thermal chain reaction; perform the logic control of the contact force of the manipulator on the theoretical contact force data of the manipulator according to the grid data of the toughness influence of the thermal chain reaction to obtain the logic control data of the welding contact force; and perform the terminal posture error compensation strategy learning based on the logic control data of the welding contact force to obtain the terminal posture error compensation strategy of the manipulator; The automated firmware design module is used to perform automated firmware design based on the manipulator end position and posture error compensation strategy, obtain the manipulator end position and posture error compensation firmware, and embed the manipulator end position and posture error compensation firmware into the bottom layer of the five-axis hybrid robot to perform end position and posture error compensation.

Images