Multi-point intelligent welding system and method

Abstract

This application provides a multi-point intelligent welding system and method. It acquires process state feedback information between the end of each welding torch and the workpiece; then dynamically predicts the thermal deformation of the workpiece and compensates for the initial desired pose of each welding torch accordingly, generating a thermally compensated global pose; next, kinematic calculations are performed based on a predefined weld path, the thermally compensated global pose, and the spatial cooperative constraint relationship between the welding torches to generate the theoretical end pose of each welding torch; for each welding torch, its process state feedback information is input into a process adaptive adjustment model to obtain a local compensation pose to offset process deviations, thus obtaining the final target pose; finally, based on the final target pose of each welding torch, the joint angle of each welding torch is determined, and the joint motor is driven to track the joint angle, achieving coordinated control of the pose and process state of each welding torch. This scheme can achieve high-precision coordinated pose control of multiple welding torches under thermal deformation conditions.

Description

Technical Field

[0001] This application relates to the field of welding technology, and more specifically, to a multi-point intelligent welding system and method. Background Technology

[0002] Welding, a key process in modern manufacturing, is widely used in aerospace, automotive manufacturing, heavy machinery, and other fields. With the improvement of industrial automation, multi-torch collaborative welding systems have gradually become the mainstream method for welding complex structural parts due to their advantages such as high efficiency and good consistency. However, during multi-torch collaborative operation, factors such as uneven heating of the workpiece, differences in the thermophysical properties of materials, and dynamic thermal deformation during the welding process cause deviations between the actual position and theoretical path of the welding torch, which seriously affects the welding quality and forming accuracy.

[0003] Traditional welding systems often rely on pre-programmed paths for control, lacking the ability to perceive and compensate for the real-time state of the workpiece (such as temperature field and thermal deformation). This is particularly pronounced in multi-welding-torch collaborative scenarios, where spatial interference and overlapping heat-affected zones between the torches become more significant. While some existing research attempts to achieve localized adjustments through visual or force feedback, a systematic multi-sensor information fusion and global-local collaborative compensation mechanism are still lacking. This fails to comprehensively suppress thermal deformation and process deviations, leading to unstable welding quality and poor process adaptability. Therefore, achieving high-precision collaborative pose control of multiple welding torches under thermal deformation conditions has become a major challenge for the industry. Summary of the Invention

[0004] This application provides a multi-point intelligent welding system and method, which can realize high-precision collaborative posture control of multiple welding torches under thermal deformation environment.

[0005] In a first aspect, this application provides a multi-point intelligent welding method, comprising the following steps:

[0006] Obtain process status feedback information between the tip of each welding torch and the workpiece;

[0007] Based on real-time monitoring of workpiece temperature field changes and workpiece material properties, the thermal deformation of the workpiece is dynamically predicted, and the initial desired pose of each welding torch is compensated according to the thermal deformation to generate the thermal compensation global pose of each welding torch.

[0008] Based on the predefined weld path, the thermal compensation global attitude, and the spatial cooperative constraint relationship between each welding torch, kinematic calculations are performed to generate the theoretical end pose of each welding torch end.

[0009] For each welding torch, input the process state feedback information of the welding torch into a process self-adaptive adjustment model, and calculate a local compensation pose for offsetting the process deviation; fuse the theoretical end pose of the welding torch with the local compensation pose to obtain a final target pose of the welding torch;

[0010] According to the final target pose of each welding torch, determine the joint angle of each welding torch, and drive the joint motor to track the joint angle to realize the cooperative control of the pose and the process state of each welding torch.

[0011] In some embodiments, obtaining the process state feedback information between the end of each welding torch and the workpiece specifically comprises:

[0012] Deploying a multi-modal sensor at the end of each welding torch, the multi-modal sensor comprising a vision sensor, a force sensor and a temperature sensor;

[0013] Real-time capture the weld image through the vision sensor, and extract the molten pool shape feature;

[0014] Using the force sensor to monitor the contact force between the welding torch and the workpiece, and using the temperature distribution of the weld zone collected by the temperature sensor, integrating the molten pool shape feature, the contact force between the welding torch and the workpiece, and the temperature distribution of the weld zone to form the process state feedback information of each welding torch.

[0015] In some embodiments, based on the real-time monitored workpiece temperature field change and the material properties of the workpiece, dynamically predicting the thermal deformation amount of the workpiece, compensating the initial desired pose of each welding torch according to the thermal deformation amount, and generating a thermal compensation global pose of each welding torch specifically comprises:

[0016] Real-time monitoring of the workpiece temperature field change, constructing a dynamic temperature distribution model, and outputting real-time temperature field data of each region of the workpiece changing with time through the dynamic temperature distribution model;

[0017] Inputting the material property parameters of the workpiece, including the thermal expansion coefficient, the thermal conductivity and the elastic modulus, together with the real-time temperature field data of each region changing with time into a finite element analysis model to predict the thermal deformation amount;

[0018] Converting the predicted thermal deformation amount into a pose compensation matrix, and performing matrix transformation on the initial desired pose of each welding torch through the pose compensation matrix to generate a thermal compensation global pose.

[0019] In some embodiments, kinematics solving is performed according to a predefined weld path, the thermal compensation global pose and the spatial coordination constraint relationship between each welding torch to generate a theoretical end pose of the end of each welding torch specifically comprises:

[0020] Obtaining predefined weld path data, the weld path data comprising path geometry and speed planning;

[0021] based on the spatial coordination constraint relationship between the welding guns, kinematics modeling is performed on the multi-welding gun system;

[0022] According to the welding seam path data and the thermal compensation global pose, inverse solution calculation is performed through the established kinematics model to generate the theoretical end pose of the end of each welding gun.

[0023] In some embodiments, for each welding gun, the process state feedback information is input into a process adaptive adjustment model, and a local compensation pose for offsetting the process deviation is calculated. The theoretical end pose of the welding gun is fused with the local compensation pose to obtain the final target pose of the welding gun, which specifically includes:

[0024] A process adaptive adjustment model is constructed, and the process adaptive adjustment model is used to learn the mapping relationship between the process state feedback information and the deviation compensation;

[0025] The process state feedback information is input into the process adaptive adjustment model, and a local compensation pose is output;

[0026] Through a weighted fusion algorithm, the theoretical end pose and the local compensation pose are added or matrix multiplied to obtain the final target pose.

[0027] In some embodiments, according to the final target pose of each welding gun, the joint angle of each welding gun is determined, and the joint motor is driven to track the joint angle to realize the coordinated control of the pose and the process state of each welding gun, which specifically includes:

[0028] Inverse kinematics calculation is performed on the final target pose of each welding gun to solve the joint angle sequence;

[0029] The joint angle sequence is input into a proportional-integral-derivative controller to generate a driving signal of the joint motor;

[0030] Real-time monitoring of the joint motor feedback is realized to achieve closed-loop tracking and ensure the coordinated consistency of the pose and the process state.

[0031] In some embodiments, the method further includes:

[0032] During the welding process, the coordinated execution data of the multi-welding gun is collected, and the parameters of the process adaptive adjustment model are updated to improve the accuracy of the next welding;

[0033] If it is detected that the process state feedback information is abnormal, a safety interrupt mechanism is triggered to suspend the welding and output an alarm information.

[0034] In a second aspect, the present application provides a multi-point intelligent welding system, which includes:

[0035] A collection module is configured to acquire process state feedback information between the end of each welding gun and a workpiece;

[0036] The processing module is configured to dynamically predict a thermal deformation amount of the workpiece based on the real-time monitored workpiece temperature field change and material properties of the workpiece, compensate the initial desired pose of each welding torch according to the thermal deformation amount, and generate a thermal compensation global pose of each welding torch;

[0037] The processing module is further configured to perform kinematic calculation according to the predefined weld path, the thermal compensation global pose, and a spatial coordination constraint relationship between the welding torches, and generate a theoretical end pose of an end of each welding torch;

[0038] The processing module is further configured to, for each welding torch, input process state feedback information of the welding torch into a process self-adaptive adjustment model, calculate a local compensation pose for offsetting process deviation, and fuse the theoretical end pose of the welding torch with the local compensation pose to obtain a final target pose of the welding torch;

[0039] The execution module is configured to determine joint angles of each welding torch according to the final target pose of each welding torch, and drive joint motors to track the joint angles, so as to realize coordinated control of the pose and process state of each welding torch.

[0040] In a third aspect, the present application provides a computer device, which comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the multi-point intelligent welding method described above.

[0041] In a fourth aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the multi-point intelligent welding method described above.

[0042] The technical scheme provided by the embodiments of the present application has the following beneficial effects:

[0043] In the multi-point intelligent welding system and method of this application, firstly, by real-time monitoring of the workpiece temperature field changes and combining material properties, the amount of thermal deformation is dynamically predicted, and thermal compensation is performed on the initial desired pose of the welding torch to generate a thermally compensated global pose. This effectively overcomes the workpiece deformation problem caused by heat accumulation during welding and improves the accuracy of the welding path. Secondly, kinematic calculations are performed through predefined weld paths, thermally compensated global poses, and spatial collaborative constraints between welding torches to generate theoretical end poses, ensuring the geometric consistency of multi-welding torch collaborative operation. Furthermore, through a process adaptive adjustment model, local compensation poses are calculated based on the process state feedback information of each welding torch. The theoretical end pose and the local compensation pose are fused to obtain the final target pose, realizing real-time cancellation of process deviations during welding and significantly improving the stability and consistency of welding quality. Finally, the joint motor is driven by inverse kinematics calculation to track the target pose, realizing high-precision collaborative pose control of multiple welding torches under thermal deformation environment. Attached Figure Description

[0044] Figure 1 This is an exemplary flowchart of a multi-point intelligent welding method according to some embodiments of this application;

[0045] Figure 2 This is an exemplary flowchart illustrating the generation of thermally compensated global attitude based on thermal deformation, according to some embodiments of this application.

[0046] Figure 3 This is a schematic diagram of the structure of a multi-point intelligent welding system according to some embodiments of this application;

[0047] Figure 4 This is a schematic diagram of the structure of a computer device for implementing a multi-point intelligent welding method according to some embodiments of this application. Detailed Implementation

[0048] To better understand the above technical solutions, the following will provide a detailed explanation of the technical solutions in conjunction with the accompanying drawings and specific implementation methods.

[0049] refer to Figure 1 The figure is an exemplary flowchart of a multi-point intelligent welding method according to some embodiments of this application. The method mainly includes the following steps:

[0050] In step 101, process status feedback information between the end of each welding torch and the workpiece is obtained.

[0051] The process status feedback information is a set of parameters reflecting the real-time status of the welding process, such as the molten pool morphology, contact force, and temperature distribution. This is not specifically limited here. In some embodiments, the process status feedback information between the welding torch tip and the workpiece can be obtained in the following ways:

[0052] deploying multi-modal sensors at each welding torch tip, the multi-modal sensors including a vision sensor, a force sensor, and a temperature sensor;

[0053] capturing a weld seam image in real time through the vision sensor, and extracting a molten pool shape feature, wherein the molten pool shape feature can include a molten pool width, a depth, and a symmetry, etc., and in a specific implementation, the weld seam image contains redundant visual information (such as background workpiece texture, spatter interference, etc.), which can be analyzed by an image processing algorithm to extract quantitative parameters reflecting the essential shape state of the molten pool after removing the interference, which will not be described here;

[0054] monitoring a welding torch and workpiece contact force using the force sensor, and in a specific implementation, for example, three-dimensional torque data of the welding torch tip is collected through the force sensor;

[0055] In addition, the temperature distribution of the weld seam area collected using the temperature sensor is used, and finally the molten pool shape feature, the welding torch and workpiece contact force, and the weld seam area temperature distribution are integrated to form process state feedback information of each welding torch, and in a specific implementation, for example, the above multi-source data is converted into a unified vector representation, and in practice, a multi-source data fusion algorithm such as a particle filtering algorithm or a weighted average algorithm can be used, which is not specifically limited here.

[0056] It should be noted that the multi-modal sensor in the present application can be integrated into the welding torch tip actuator to ensure real-time and accuracy, which will not be described here.

[0057] In step 102, based on the real-time monitored workpiece temperature field change and the material properties of the workpiece, the thermal deformation amount of the workpiece is dynamically predicted, the initial desired pose of each welding torch is compensated according to the thermal deformation amount, and the thermal compensation global pose of each welding torch is generated.

[0058] Wherein, the thermal compensation global pose refers to the global adjustment pose after the initial desired pose is compensated for thermal deformation, and the initial desired pose refers to the welding torch target position and pose preset based on the weld seam path. In some embodiments, referring to Figure 2 The figure is an exemplary flow chart for generating a thermal compensation global pose based on a thermal deformation amount according to some embodiments of the present application, which specifically includes:

[0059] At step 1021, the workpiece temperature field change is monitored in real time, a dynamic temperature distribution model is constructed, and real-time temperature field data of each region of the workpiece changing over time is output through the dynamic temperature distribution model. For example, temperature data can be collected by an infrared thermal imager or other temperature collection equipment, and a dynamic temperature distribution model can be constructed by simulating the temperature field evolution using the finite difference method. The core principle of the finite difference method simulation is to discretize the continuous workpiece temperature field into a finite number of grid nodes, and to replace the differential of the continuous temperature field change with the difference between nodes, so as to convert the continuous heat conduction partial differential equation which is difficult to solve directly into an algebraic equation group which can be solved by numerical calculation, and finally realize the simulation of the temperature field evolution over time. In specific implementation, the following methods can be used, that is:

[0060] The three-dimensional geometric model of the workpiece to be welded is divided into uniform spatial grids (such as cubic grids), and each grid intersection is a temperature calculation node. The coordinates and initial temperature of each node are recorded to realize the conversion of continuous temperature field to discrete node group.

[0061] Based on the thermal conductivity of the workpiece material, a heat conduction partial differential equation describing the change of temperature with time and space is established. At the same time, the boundary conditions (such as temperature input of the welding heat source, heat loss of the workpiece surface and air, heat exchange between the workpiece and the clamp, etc.) are defined, and the temperature change rule of the grid boundary node is determined.

[0062] The collected initial temperature data of the workpiece is used to assign the collected temperature values to the discretized grid nodes as the initial state of the temperature field evolution calculation.

[0063] According to the fixed time step, the finite difference formula (such as central difference or forward difference) is used to approximate the differential term in the heat conduction equation for each grid node, and an algebraic relationship between the node temperature, adjacent node temperature and time step is established. The algebraic equation group is solved to obtain the new temperature value of each node at the current time step, and one temperature field evolution simulation is completed.

[0064] The new temperature data of the workpiece is obtained in real time by the temperature collection equipment, and the temperature value of the grid node is periodically corrected. The iterative calculation is continued to realize the real-time update of the dynamic temperature distribution model, which will not be described here.

[0065] At step 1022, the material characteristic parameters of the workpiece, including the thermal expansion coefficient, the thermal conductivity and the elastic modulus, are input into the finite element analysis model together with the real-time temperature field data of each region of the workpiece changing over time to predict the thermal deformation amount. In specific implementation, the finite element analysis model predicts the thermal deformation amount by first calculating the temperature field, then calculating the thermal stress, and finally obtaining the deformation amount, that is, the three-dimensional geometric model of the workpiece to be analyzed is first discretized into a plurality of independent micro calculation units (such as tetrahedral units, hexahedral units); then, the real-time temperature distribution of each discrete unit at different time nodes is determined by combining the input material thermal conductivity and real-time temperature field data through thermal conduction numerical calculation (that is, the specific temperature value of each unit is determined), which provides the basis for the subsequent thermal strain calculation.

[0066] Based on the above obtained real-time temperature distribution of the unit, the temperature change amount of each unit (that is, the difference between the real-time temperature of the unit and the initial temperature) is first calculated; then, the thermal strain of each unit caused by temperature change (that is, the stretching or deformation trend of the unit) is calculated by combining the input material thermal expansion coefficient (that is, reflecting the stretching ratio of the material under unit temperature change) through the thermal strain formula.

[0067] In addition, the thermal stress of each unit caused by thermal strain is solved by combining the input material elastic modulus (that is, reflecting the ability of the material to resist deformation, the greater the elastic modulus, the greater the stress generated under the same thermal strain) through the structural mechanics balance equation; subsequently, based on the unit correlation of the finite element model, the deformation amounts of all discrete units are superimposed and calculated, and finally the thermal deformation amount of the whole workpiece (including the translation deviation of the workpiece along the X / Y / Z axis, and the rotation deviation such as warping and deflection caused by thermal stress) is obtained. Here, no further description is given;

[0068] At step 1023, the predicted thermal deformation amount is converted into a pose compensation matrix, and the initial desired pose of each welding gun is matrix transformed through the pose compensation matrix to generate a thermal compensation global pose. In specific implementation, for example, the pose compensation matrix can be represented as a 4x4 homogeneous transformation matrix to realize translation and rotation compensation. The homogeneous coordinates of the constructed pose compensation matrix and the initial desired pose are multiplied (4x4 matrix x 4x1 vector) to obtain a new 4x1 homogeneous coordinate; the first 3 elements of the coordinate are the target position after compensation, and the global pose after thermal compensation is jointly constituted by combining the corresponding pose of the rotation matrix.

[0069] It should be noted that in the present application, as a preferred embodiment, the pose compensation matrix is a 4x4 homogeneous transformation matrix constructed based on the thermal deformation of the workpiece, which functions to quantify the translation deviation (e.g., a certain region of the workpiece deviates along the X / Y / Z axis due to heat) and rotation deviation (e.g., the workpiece warps and deflects due to thermal stress) caused by thermal deformation into a mathematical matrix, which is used to accurately correct the initial desired pose (i.e., the preset target position and attitude) of the welding gun, and finally output a thermal compensation global pose that can adapt to the state of the workpiece after thermal deformation.

[0070] In step 103, kinematics calculation is performed according to the predefined weld path, the thermal compensation global pose, and the spatial coordination constraint relationship between the welding guns, to generate the theoretical end pose of the end of each welding gun.

[0071] The spatial coordination constraint relationship refers to the constraint relationship of the position, attitude, and movement limitation between the multiple welding guns, and the theoretical end pose refers to the target position and attitude of the welding gun end calculated based on the kinematics model.

[0072] In some embodiments, the kinematics calculation according to the predefined weld path, the thermal compensation global pose, and the spatial coordination constraint relationship between the welding guns to generate the theoretical end pose of the end of each welding gun can be specifically implemented in the following manner, i.e.:

[0073] Obtain the predefined weld path data, which includes path geometry and speed planning;

[0074] Based on the spatial coordination constraint relationship between the welding guns, the kinematics model of the multi-welding gun system is established. It should be noted that in the present application, the kinematics modeling based on the spatial coordination constraint of the multiple welding guns is to convert the spatial coordination requirements (e.g., anti-interference and attitude synchronization) between the multiple welding guns into mathematical constraint conditions, and combine the kinematics model of a single welding gun (e.g., the degrees of freedom of the mechanical arm of a single welding gun, the joint parameters) to construct the overall multi-body kinematics model of the multi-welding gun system. Specifically, for example: first, build a basic motion model of a single welding gun, i.e., for each welding gun, determine the number of joints (e.g., 6 movable joints), the length of the connecting rod between each joint, the direction parameter of the joint rotation, and then establish the corresponding relationship between the joint rotation angle of the welding gun and the position and attitude of the end of the welding gun, and clarify the rules of the single welding gun movement; secondly, convert the spatial coordination requirements of multiple welding guns into specific rules according to the requirements that the multiple welding guns cannot collide with each other and the attitudes (e.g., the angle of inclination of the welding gun) must be synchronized during welding, to determine the specific coordination rules, i.e., the anti-collision rule and the attitude synchronization rule, wherein the anti-collision rule is to stipulate that the ends of any two welding guns must maintain a safe distance to avoid collision during movement; the attitude synchronization rule is to stipulate that the attitudes (e.g., the angle of inclination of the welding gun) of all welding gun ends must be consistent, and the deviation cannot exceed the set range.

[0075] Finally, the overall motion model of the multi-welding gun system is integrated, that is, based on the motion model of a single welding gun, the determined coordination rules are added as global requirements to build the motion model of the entire multi-welding gun system. Specifically, the joint angles of all welding guns are taken as the calculation variables, and the positions and attitudes that the ends of the welding guns need to reach are taken as the targets to form a set of calculation logic containing the motion rules of a single welding gun and the coordination rules, so that the motions of the multi-welding gun can be associated, and finally the overall motion model of the multi-welding gun system is obtained.

[0076] According to the weld path data and the heat compensation global attitude, the theoretical end poses of the ends of the welding guns are generated by inverse solution calculation through the established kinematic model. Specifically, the basic target pose can be determined by integrating the weld path and heat compensation, that is, the pre-defined weld path data is extracted first to determine the trajectory (such as the direction of a straight weld and the radius of a circular weld) and the motion speed that each welding gun needs to follow during the welding process; then the heat compensation global attitude (which has offset the welding gun overall attitude adjustment amount due to workpiece thermal deformation) is superimposed to adjust the original end pose corresponding to the weld path to the basic target pose after heat compensation, that is, the position and attitude that the end of the welding gun needs to reach, which not only fits the weld trajectory but also adapts to the state after the workpiece thermal deformation; then the basic target pose obtained before is substituted into the previously established multi-welding gun kinematic model (which contains spatial coordination constraints such as anti-collision and attitude synchronization), to check whether the basic target poses of all welding guns meet the coordination rules: for example, whether the basic target poses of any two welding guns maintain a safe distance and whether the attitudes are synchronized; if there is a conflict, the basic target pose of a single welding gun can be fine-tuned (for example, the attitude angle is slightly adjusted without deviating from the weld path) until the target poses of all welding guns meet the coordination requirements to obtain the compliant target pose; finally, the compliant target pose of each welding gun is calculated by inverse solution calculation through the kinematic model: that is, the angle that each joint of the welding gun needs to rotate is calculated by reverse deduction from the compliant target pose that the end of the welding gun needs to reach, and this process relies on the motion rules of a single welding gun in the model to ensure that the deduced joint parameters can accurately drive the end of the welding gun to reach the compliant target pose. According to the joint parameters obtained by inverse solution, the end pose of the welding gun is inversely deduced once through the forward solution (end pose from joint parameters) of the kinematic model to check whether the pose is consistent with the compliant target pose; if it is consistent, the compliant target pose is the final theoretical end pose of each welding gun end; if the deviation exceeds the standard, the inverse solution parameters are adjusted until the requirements are met, which is not described here.

[0077] It should be noted that the compliant target pose in this application is the target pose of the welding gun end without spatial conflict and in accordance with the coordination rules, and the kinematic model can consider the degrees of freedom of the welding gun (such as a 6-DOF robot arm) to ensure the real-time and uniqueness of the calculation, which is not described here.

[0078] In step 104, for each welding torch, the process state feedback information is input into the process adaptive adjustment model to calculate a local compensation pose for offsetting the process deviation; the theoretical end pose of the welding torch is fused with the local compensation pose to obtain the final target pose of the welding torch.

[0079] Wherein, the local compensation pose refers to the fine adjustment amount for the predetermined process deviation, the final target pose refers to the fused welding torch control target, and the process adaptive adjustment model is a model for learning the mapping relationship between the process state feedback information and the local deviation compensation.

[0080] In some embodiments, for each welding torch, the process state feedback information is input into the process adaptive adjustment model to calculate a local compensation pose for offsetting the process deviation; the theoretical end pose of the welding torch is fused with the local compensation pose to obtain the final target pose of the welding torch, which can be specifically implemented in the following manner:

[0081] To construct the process adaptive adjustment model, when implemented, it can be constructed in the following manner, that is, first determine the input and output dimensions of the model, wherein the input is the process state feedback information, that is, the molten pool shape features (width, depth, symmetry), the welding torch and workpiece contact force (three-dimensional torque data), and the weld zone temperature distribution (temperature field quantization data), which are integrated into a unified input vector;

[0082] The output is the key parameters of the local compensation pose, that is, the position offset (such as the small movement amount in X / Y / Z axis direction) and the attitude correction amount (such as the small rotation angle around X / Y / Z axis) of the welding torch end that needs to be fine-tuned, forming an output vector.

[0083] Secondly, collect and preprocess the training data, which can be carried out by multiple groups of welding experiments to record the process state feedback information-actual deviation-effective compensation pose data under different welding conditions:

[0084] Wherein, the collected data can be recorded by multi-modal sensors (vision, force, and temperature sensors) to record the process state, high-precision measurement equipment (such as laser displacement sensor) to record the actual deviation of the welding torch, and artificial or algorithm to determine the effective compensation pose that can offset the deviation;

[0085] The preprocessed data can be used to remove abnormal data, normalize the data (for example, unify the data of different magnitudes such as temperature and contact force to the same range), and divide the training set (which is used for model learning) and the validation set (which is used for verifying the accuracy of the model).

[0086] Thirdly, the model structure is selected and trained offline. The model structure capable of processing dynamic time series data and learning nonlinear mapping relationship is selected, such as a neural network model. The model structure adopts a combination structure of CNN+RNN (wherein CNN extracts spatial features such as molten pool images, and RNN processes time series change features such as temperature and contact force). In specific implementation, a simplified deep learning model can also be used, and the specific model is not limited here.

[0087] The offline training is to input the training set data into the model, to minimize the deviation between the compensation pose output by the model and the actual effective compensation pose, and to iteratively optimize the model parameters until the compensation accuracy of the model on the verification set reaches the preset requirement.

[0088] In specific implementation, the process state feedback information is input into the process adaptive adjustment model, and a local compensation pose for offsetting the process deviation is output. Since the adaptive adjustment model is a pre-trained model, that is, the corresponding relationship between the process state feedback information and the local deviation compensation has been learned through historical welding data. When the real-time process state feedback information is input, the adaptive adjustment model can quickly match the deviation type and degree corresponding to the current process state, determine the direction and range to be corrected, and finally output the local compensation pose of the adaptive adjustment model, which is a small adjustment parameter for a single welding gun and includes two parts: one is position compensation (such as millimeter-level small movement along the X / Y / Z axis, used to correct the position deviation of the welding gun and the weld); the other is attitude compensation (such as small-angle rotation around the X / Y / Z axis, used to correct the inclination angle deviation of the welding gun). The compensation pose is only for the local process deviation of a single welding gun and does not affect the global collaborative relationship of multiple welding guns.

[0089] In addition, in some embodiments, the final target pose of the welding gun is obtained by fusing the theoretical end pose of the welding gun and the local compensation pose through a weighted fusion algorithm. The final target pose is obtained by adding or matrix multiplying the theoretical end pose and the local compensation pose. The addition is for the fusion of the position dimension. Since the theoretical end pose and the local compensation pose both contain position information, the position of the theoretical end pose is a global target position calculated based on the predefined weld path, the global attitude of thermal compensation, and the spatial collaborative constraint of multiple welding guns, which ensures that the welding gun adheres to the weld trajectory, adapts to the workpiece thermal deformation, and does not interfere with other welding guns.

[0090] The position of the local compensation pose is a small position fine adjustment calculated by the process self-adaptive adjustment model for the local process deviation of a single welding gun; the addition is to superimpose the position coordinates of the two, to correct the local deviation on the basis of the global position, to obtain the final position taking into account the compliance and accuracy, and the matrix multiplication is the fusion of the attitude dimension, since the attitude information (for example, the rotation angle of the welding gun around the X, Y and Z axes) of the theoretical end pose and the local compensation pose can be represented by a homogeneous transformation matrix (4x4 matrix, containing rotation and translation information), the attitude matrix of the theoretical end pose is the attitude adjustment result (for example, the overall tilt angle adapted to the thermal deformation of the workpiece) reflecting the global coordination requirement; the attitude matrix of the local compensation pose is a small attitude correction matrix calculated for the local attitude deviation of a single welding gun; the matrix multiplication is to fuse the rotation information of the global attitude and the fine adjustment rotation information of the local attitude (that is, the matrix operation can accurately transmit the rotation logic and avoid attitude conflict) through the mathematical operation of the two attitude matrices, to obtain the final attitude that meets the global coordination of multiple welding guns and corrects the local attitude deviation, which will not be described here.

[0091] In step 105, the joint angles of each welding gun are determined according to the final target pose of each welding gun, and the joint motor is driven to track the joint angles, to realize the coordinated control of the pose and the process state of each welding gun.

[0092] Wherein, the joint angle refers to the rotation angle of each joint of the welding gun mechanical arm, and the coordinated control refers to the real-time matching of the pose adjustment and the process state.

[0093] In some embodiments, the joint angles of each welding gun are determined according to the final target pose of each welding gun, and the joint motor is driven to track the joint angles, to realize the coordinated control of the pose and the process state of each welding gun, which specifically includes:

[0094] Inverse kinematics calculation is performed on the final target pose of each welding gun to solve the joint angle sequence, for example, the geometric method or numerical optimization algorithm is used to calculate multiple solutions and select the optimal solution, which is not limited here, and in specific implementation, the kinematics model of the welding gun has established the positive solution relationship (forward kinematics) of the joint angle to the end pose, and the inverse kinematics calculation is a reverse solution, which may obtain multiple sets of joint angles (i.e. "multiple solutions") satisfying the target pose, and the optimal solution needs to be selected in combination with the actual motion limit of the welding gun (such as the maximum rotation angle of the joint, the motion stability requirement), for example, the angle combination with the smallest joint motion amplitude, no over-limit risk, and no interference between the welding gun and the workpiece / other welding guns is selected, and then arranged according to the time step of the welding process to form a joint angle sequence;

[0095] The joint angle sequence is input into a proportional-differential-integral (PID) controller to generate a driving signal of the joint motor. In specific implementation, the joint angle sequence, i.e., the target angle of each joint of the welding torch at different time steps, is input, and then the actual running state of the joint motor is monitored in real time (e.g., the actual angle of the current joint is obtained through a motor encoder), to form a closed-loop comparison basis of the target angle and the actual angle. The PID controller calculates the deviation between the target angle and the actual angle through three links of proportion (P), integration (I), and differentiation (D), and outputs a control quantity for correcting the deviation. Finally, the PID controller converts the calculated correction control quantity into a driving signal recognizable by the joint motor, which will not be described herein.

[0096] The joint motor feedback is monitored in real time to realize closed-loop tracking and ensure the collaborative consistency of the pose and the process state, i.e., the monitoring-comparison-correction process is continuously circulated to ensure that the joint motor always tracks the target angle, and then the end pose of the welding torch accurately meets the final target pose, and the welding process state (e.g., the molten pool, the contact force, and the temperature) is kept in collaborative consistency. In specific implementation, the encoder feedback can be used to correct the deviation, which will not be described herein.

[0097] In some embodiments, the method further comprises:

[0098] In the welding process, the collaborative execution data of the multiple welding torches are collected, and the parameters of the process adaptive adjustment model are updated to improve the accuracy of the next welding. In the welding process, the key execution data of the multiple welding torches are recorded in real time, including the process state feedback information (the molten pool shape, the contact force, and the weld zone temperature) of each welding torch, the actual end pose (the deviation from the theoretical pose), the joint motor running parameters (e.g., the actual joint angle), and the local compensation pose output by the process adaptive adjustment model and the corresponding compensation effect (e.g., whether the deviation is reduced after compensation). The parameters of the process adaptive adjustment model are updated by arranging the collected collaborative execution data as sample pairs of process state-compensation effect, and fine-tuning the model parameters (e.g., the weights and biases of the neural network) with these new samples. For example, if it is found that the original compensation amount is too small under a certain type of molten pool deviation, the mapping relationship corresponding to this type of process state in the model is corrected through an algorithm to make the model more suitable for the actual welding conditions. Obviously, the updated model can more accurately identify the process deviation in the next welding, output a more suitable local compensation pose, reduce the deviation between the end pose of the welding torch and the ideal state, and thus improve the forming accuracy and quality consistency of the next welding.

[0099] In addition, if it is detected that the process state feedback information is abnormal, a safety interrupt mechanism can be triggered to suspend the welding and output an alarm information, which will not be described herein.

[0100] In addition, another aspect of the present application provides a multi-point intelligent welding system in some embodiments, which is described with reference toFigure 3 The figure is a schematic diagram of the structure of a multi-point intelligent welding system according to some embodiments of this application. The multi-point intelligent welding system 400 includes: a data acquisition module 401, a processing module 402, and an execution module 403, which are described below:

[0101] The acquisition module 401 is used to acquire process status feedback information between the end of each welding torch and the workpiece;

[0102] The processing module 402 is used to dynamically predict the thermal deformation of the workpiece based on the real-time monitoring of the workpiece temperature field change and the material properties of the workpiece, and to compensate the initial expected pose of each welding torch according to the thermal deformation, thereby generating the thermal compensation global pose of each welding torch.

[0103] The processing module 402 is also used to perform kinematic calculations based on the predefined weld path, the thermal compensation global attitude, and the spatial cooperative constraint relationship between each welding torch, to generate the theoretical end pose of each welding torch end.

[0104] The processing module 402 is also used to input the process status feedback information of each welding torch into the process adaptive adjustment model, calculate the local compensation pose for offsetting process deviations, and fuse the theoretical end pose of the welding torch with the local compensation pose to obtain the final target pose of the welding torch.

[0105] The execution module 403 is used to determine the joint angle of each welding torch according to the final target pose of each welding torch, and drive the joint motor to track the joint angle, so as to realize the coordinated control of the pose and process state of each welding torch.

[0106] In addition, this application also provides a computer device, the computer device including a memory and a processor, the memory storing code, and the processor being configured to acquire the code and execute the above-described multi-point intelligent welding method.

[0107] In some embodiments, reference Figure 4 The figure is a schematic diagram of the structure of a computer device implementing a multi-point intelligent welding method according to some embodiments of this application. The method in the above embodiments can be implemented through... Figure 4 The computer device 500 shown is used to implement this, and the computer device 500 includes at least one processor 501, a communication bus 502, a memory 503, and at least one communication interface 504.

[0108] The processor 501 can be a general central processing unit (CPU), an application specific integrated circuit (ASIC), or one or more processors for controlling the execution of the multi-point intelligent welding method in the present application.

[0109] The communication bus 502 can include a path for transmitting information between the above-mentioned components.

[0110] The memory 503 can be a read only memory (ROM) or other type of static storage device that can store static information and instructions, a random access memory (RAM) or other type of dynamic storage device that can store information and instructions, an electrically erasable programmable read only memory (EEPROM), a compact disc read only memory (CDROM) or other optical disk storage, a magnetic disk or other magnetic storage device, or any other medium capable of carrying or storing desired program code in the form of instructions or data structures and that can be accessed by a computer, but not limited to. The memory 503 can exist independently and be connected to the processor 501 through the communication bus 502. The memory 503 can also be integrated with the processor 501.

[0111] The memory 503 is used to store program code for executing the scheme of the present application, and the processor 501 is used to control the execution. The processor 501 is used to execute the program code stored in the memory 503. The program code can include one or more software modules. The execution of the multi-point intelligent welding method in the above-mentioned embodiments can be realized by the processor 501 and one or more software modules in the program code in the memory 503.

[0112] The communication interface 504 uses any transceiver-like device for communicating with other devices or communication networks, such as an Ethernet, a radio access network (RAN), a wireless local area network (WLAN), etc.

[0113] In a particular implementation, as one embodiment, the computer device can include multiple processors, each of which can be a single CPU processor or a multi-CPU processor. A processor herein can refer to one or more devices, circuits, and / or processing cores for processing data, such as computer program instructions.

[0114] The computer device described above can be a general purpose computer device or a special purpose computer device. In a particular implementation, the computer device can be a desktop computer, a laptop computer, a network server, a personal digital assistant (PDA), a mobile phone, a tablet computer, a wireless terminal device, a communication device, or an embedded device. The embodiments of the present application do not limit the type of the computer device.

[0115] In addition, the present application also provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the multi-point intelligent welding method described above.

[0116] Although the preferred embodiments of the present application have been described, those skilled in the art can make additional changes and modifications to the embodiments once they know the basic inventive concept. Therefore, the appended claims are intended to cover all changes and modifications falling within the scope of the present application.

[0117] Obviously, those skilled in the art can make various modifications and variations to the present application without departing from the spirit and scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the claims of the present application and their equivalent technologies, the present application also intends to include these modifications and variations.

Claims

1. A multi-point intelligent welding method, characterized by, The method comprises the following steps: acquiring process state feedback information between each welding torch tip and the workpiece; based on the real-time monitored workpiece temperature field change and the material properties of the workpiece, dynamically predicting the thermal deformation amount of the workpiece, compensating the initial desired pose of each welding torch according to the thermal deformation amount, and generating a thermal compensation global pose of each welding torch; kinematics calculation is performed according to the predefined weld path, the thermal compensation global pose and the spatial coordination constraint relationship between the welding torches, and a theoretical end pose of each welding torch tip is generated; for each welding torch, the process state feedback information of the welding torch is input into a process self-adaptive adjustment model, and a local compensation pose for offsetting the process deviation is calculated; the theoretical end pose of the welding torch is fused with the local compensation pose to obtain a final target pose of the welding torch; based on the final target pose of each welding torch, the joint angle of each welding torch is determined, and the joint motor is driven to track the joint angle, so as to realize the coordinated control of the pose and the process state of each welding torch; wherein the thermal compensation global pose refers to a global adjustment pose after thermal deformation compensation of the initial desired pose, based on the real-time monitored workpiece temperature field change and the material properties of the workpiece, the thermal deformation amount of the workpiece is dynamically predicted, and the initial desired pose of each welding torch is compensated according to the thermal deformation amount, to generate a thermal compensation global pose of each welding torch, which specifically comprises: real-time monitoring of the workpiece temperature field change, construction of a dynamic temperature distribution model, and output of real-time temperature field data of each region of the workpiece changing over time through the dynamic temperature distribution model; inputting the material property parameters of the workpiece, including the thermal expansion coefficient, the thermal conductivity and the elastic modulus, into a finite element analysis model together with the real-time temperature field data of each region changing over time to predict the thermal deformation amount; the predicted thermal deformation amount is converted into a pose compensation matrix, and the initial desired pose of each welding torch is matrix transformed through the pose compensation matrix to generate a thermal compensation global pose.

2. The multi-point intelligent welding method of claim 1, wherein, The acquisition of the process state feedback information between each welding torch tip and the workpiece specifically comprises: deploying multi-modal sensors on each welding torch tip, the multi-modal sensors including visual sensors, force sensors and temperature sensors; real-time capture of weld images through the visual sensors, and extraction of molten pool shape features; using the force sensors to monitor the contact force between the welding torch and the workpiece, and using the temperature distribution of the weld zone collected by the temperature sensors to integrate the molten pool shape features, the contact force between the welding torch and the workpiece and the temperature distribution of the weld zone to form the process state feedback information of each welding torch.

3. The multi-point intelligent welding method of claim 1, wherein, kinematics calculation is performed according to the predefined weld path, the thermal compensation global pose and the spatial coordination constraint relationship between the welding torches, and a theoretical end pose of each welding torch tip is generated; acquiring predefined weld path data, the weld path data including path geometry and speed planning; based on the spatial coordination constraint relationship between the welding torches, kinematics modeling is performed on the multi-welding torch system; according to the weld path data and the thermal compensation global pose, inverse solution calculation is performed through the established kinematics model to generate a theoretical end pose of each welding torch tip.

4. The multi-point intelligent welding method of claim 1, wherein, For each welding torch, input the process state feedback information into the process self-adaptive adjustment model, calculate the local compensation pose for offsetting the process deviation, fuse the theoretical end pose of the welding torch with the local compensation pose to obtain the final target pose of the welding torch, and the final target pose of the welding torch specifically comprises: A process self-adaptive adjustment model is constructed, and the process self-adaptive adjustment model is used to learn the mapping relationship between the process state feedback information and the deviation compensation; The process state feedback information is input into the process self-adaptive adjustment model, and the local compensation pose is output; The theoretical end pose and the local compensation pose are added or matrix multiplied through a weighted fusion algorithm to obtain the final target pose.

5. The multi-point intelligent welding method of claim 1, wherein, According to the final target pose of each welding torch, the joint angle of each welding torch is determined, and the joint motor is driven to track the joint angle to realize the cooperative control of the pose and the process state of each welding torch, and the cooperative control of the pose and the process state of each welding torch specifically comprises: Inverse kinematics calculation is performed on the final target pose of each welding torch to solve the joint angle sequence; The joint angle sequence is input into a proportional-integral-derivative controller to generate a driving signal of the joint motor; Real-time monitoring of the joint motor feedback is realized to achieve closed-loop tracking and ensure the cooperative consistency of the pose and the process state.

6. The multi-point intelligent welding method of claim 1, wherein, The method further comprises: During the welding process, cooperative execution data of multiple welding torches are collected, and parameters of the process self-adaptive adjustment model are updated to improve the accuracy of the next welding; If it is detected that the process state feedback information is abnormal, a safety interrupt mechanism is triggered to suspend the welding and output an alarm information.

7. A multi-point intelligent welding system which adopts the method of any one of claims 1 to 6 for multi-point intelligent welding, characterized by, The system comprises: A collection module is configured to acquire process state feedback information between the end of each welding torch and the workpiece; A processing module is configured to dynamically predict the thermal deformation amount of the workpiece based on the real-time monitored temperature field change of the workpiece and the material properties of the workpiece, compensate the initial desired pose of each welding torch according to the thermal deformation amount, and generate a thermal compensation global pose of each welding torch; The processing module is further configured to perform kinematics calculation according to a predefined weld path, the thermal compensation global pose, and a spatial cooperative constraint relationship between the welding torches to generate a theoretical end pose of the end of each welding torch; The processing module is further configured to input the process state feedback information of each welding torch into the process self-adaptive adjustment model, calculate a local compensation pose for offsetting the process deviation, and fuse the theoretical end pose of the welding torch with the local compensation pose to obtain the final target pose of the welding torch; An execution module is configured to determine the joint angle of each welding torch according to the final target pose of each welding torch, drive the joint motor to track the joint angle, and realize the cooperative control of the pose and the process state of each welding torch.

8. A computer device, comprising: The computer device comprises a memory and a processor, the memory stores a code, and the processor is configured to acquire the code and execute the multi-point intelligent welding method according to any one of claims 1 to 6.

9. A computer-readable storage medium storing a computer program, the computer program comprising instructions that, when executed by a computer, cause the computer to perform the method of any one of claims 1 to 8. The computer program is executed by the processor to realize the multi-point intelligent welding method according to any one of claims 1 to 6.

Images