Automatic welding robot system suitable for different industries

Abstract

The invention relates to the technical field of welding robot systems. An automatic welding robot system suitable for different industries comprises a welding seam recognition module used for recognizing a welding seam, a welding path point extraction module used for extracting welding path points, a path optimization module used for optimizing a welding path and a path optimization module used for obtaining three-dimensional contour data of the welding seam, and processing the three-dimensional contour data, extracting the shape and position change parameters of the welding seam, fitting and smoothing the shape and position change parameters of the welding seam based on the three-dimensional contour data, extracting a welding path, and controlling a welding gun of the welding robot to move along the optimal path. According to the welding seam recognition module, a laser sensor is used for obtaining three-dimensional contour data of a welding seam, median filtering and Gaussian filtering are sequentially carried out, noise interference is effectively removed, then the edge of the welding seam is accurately extracted through coordinate transformation and a Canny algorithm, and the shape and position change of the welding seam are determined.

Description

Technical Field

[0001] The present invention relates to the technical field of welding robot systems, and specifically to an automated welding robot system applicable to different industries. Background Art

[0002] In the prior art, a patent document with the publication number CN113245752B discloses a weld seam recognition system and a welding method for intelligent welding, including a weld seam recognition device for intelligent welding, a computer, and an intelligent welding robot. The weld seam recognition device for intelligent welding includes a housing, in which there are an electronic gyroscope, a temperature and humidity sensor, an array displacement sensor, a length sensor, a guide wheel with an acceleration sensor, a control switch, and a control circuit; the computer includes a graphics processing unit, a control system, and a welding machine system. In the above device, an operator uses the weld seam recognition device to recognize the weld seam of the welding workpiece and the working environment, etc., and the rest is processed by the computer, and the intelligent welding robot completes the welding, greatly reducing the requirements for the operation ability of intelligent welding. However, the above system has the following technical problems when in use:

[0003] 1. The operation difficulty during weld seam recognition is large and the degree of automation is low;

[0004] 2. It is not convenient to find the optimal welding path under the premise of meeting the welding process, and it is not possible to improve the welding efficiency, reduce the welding time and energy consumption;

[0005] 3. The applicable scenarios are relatively limited;

[0006] 4. It is not convenient to dynamically adjust the welding parameters and ensure the stability of the welding quality;

[0007] Based on this, the present invention provides an automated welding robot system applicable to different industries to solve the technical problems raised in the above background art. Summary of the Invention

[0008] Aiming at the technical problems existing in the prior art, the present invention provides an automated welding robot system applicable to different industries to solve the technical problems of large operation difficulty and low degree of automation during the weld seam recognition of existing welding robots, not being convenient to find the optimal welding path under the premise of meeting the welding process, not being able to improve the welding efficiency, reduce the welding time and energy consumption, and relatively limited applicable scenarios.

[0009] The technical solution of the present invention to solve the above technical problems is as follows: An automated welding robot system applicable to different industries, including:

[0010] A weld seam recognition module for recognizing weld seams;

[0011] A welding path point extraction module for extracting welding path points;

[0012] A path optimization module is used to optimize the welding path. The path optimization module obtains the three-dimensional contour data of the weld, processes the three-dimensional contour data and extracts the weld shape and position change parameters, fits and smoothes the weld shape and position change parameters based on the three-dimensional contour data, extracts the welding path, controls the welding robot's welding gun to move along the optimal path, and uses an algorithm to optimize the path;

[0013] Robot kinematics calculation module, used to calculate the kinematic parameters of the welding robot;

[0014] Welding parameter adjustment module, used for adaptive adjustment of welding parameters;

[0015] Monitoring module, used for weld quality monitoring;

[0016] The modules communicate via an industrial bus.

[0017] The beneficial effects of the present invention are:

[0018] 1. The weld recognition module of the present invention uses a laser sensor to obtain the three-dimensional contour data of the weld, which is processed by median filtering and Gaussian filtering to effectively remove noise interference, and then accurately extracts the weld edge through coordinate transformation and Canny algorithm to determine the shape and position changes of the weld. The point cloud data integration unit and the point cloud denoising unit further process the data, which improves the accuracy and stability of weld recognition, and then effectively solves the problems of difficult operation and low degree of automation of weld recognition in the prior art.

[0019] 2. The path optimization module of the present invention adopts the A-star algorithm to perform local path optimization, combines the welding speed and the welding gun posture angle to calculate the motion energy consumption, and determines the shortest path from the current point to the end point through heuristic estimation. At the same time, it uses the dynamic programming algorithm for global optimization, searches for the global optimal path under the constraints of speed and welding gun posture angle, and searches for the optimal path, thereby effectively improving the welding efficiency of the system and reducing the welding time and energy consumption.

[0020] 3. The welding parameter monitoring unit of the present invention monitors various parameters in the welding process and parameters of the workpiece to be welded in real time through auxiliary sensors. The welding parameter adjustment unit automatically adjusts the current, voltage, welding speed and wire feeding speed according to data feedback. The intelligent control module adopts machine learning algorithm and fuzzy control algorithm to adjust the welding current and voltage according to the material and thickness of the workpiece. At the same time, each unit works together to generate the final welding parameter combination and updates the model weights online. By realizing the above-mentioned technical effects, dynamic adjustment of welding parameters is realized and the stability of welding quality is guaranteed.

[0021] Based on the above technical solution, the present invention can also be improved as follows.

[0022] Furthermore, the weld seam recognition module includes the following units:

[0023] The acquisition unit includes a laser sensor. The laser sensor scans the surface of the workpiece at a set frequency and scanning angle to obtain the three-dimensional contour data of the weld seam. First, median filtering is used to preliminarily process the collected raw data to remove discrete interference points such as pulse noise. Then, Gaussian filtering is used to remove high-frequency noise to smooth the data. Then, through coordinate transformation, the data is converted to a unified coordinate system. Using the Canny algorithm, the weld seam edge is accurately extracted, and then the shape and position changes of the weld seam are determined;

[0024] The point cloud data integration unit: The laser sensor emits a laser beam, and the reflected light is received by a CCD camera to form the three-dimensional point cloud data of the weld seam. The point cloud data is expressed as:

[0025] P = pi(xi, yi, zi), i = 1, 2,..., N

[0026] where pi(xi, yi, zi) represents the three-dimensional coordinates of the i-th point, and N is the total number of point cloud data;

[0027] The point cloud denoising unit: Gaussian filtering is used to perform preliminary noise reduction processing on measurement errors and environmental noise, and then the RANSAC algorithm is used to fit the weld seam curve and remove abnormal points:

[0028] ax + by + cz + d = 0

[0029] where a, b, c, and d are fitting parameters, and x, y, and z represent the normal vector and intercept of the plane;

[0030] Calculate the distance from all points to the fitted plane:

[0031]

[0032] If D i < T threshold , then the point p i is regarded as an inlier, and outliers greater than the threshold are removed;

[0033] where T threshold is the distance threshold for determining whether a point is an inlier;

[0034] 0.5mm ≤ T threshold ≤ 2mm, and the specific value of T threshold is limited by the accuracy of the laser sensor and the shape of the weld seam;

[0035] This step solves the technical problem of accurately obtaining the shape and position information of the weld seam in a complex environment. Compared with the prior art, its beneficial effect is that it can more accurately identify the weld seam, reduce the influence of data noise interference on weld seam identification, and improve the accuracy and stability of weld seam identification;

[0036] For example, in an automobile manufacturing workshop where the environmental light is complex and the mechanical vibration is large, this module can effectively extract the weld seam information from the raw data with noise, providing a reliable basis for subsequent welding work.

[0037] Furthermore, the welding path point extraction module includes the following units:

[0038] Weld seam shape analysis unit, which uses the least squares method to fit and calculate the weld seam center line for the three-dimensional point cloud data collected by the point cloud data integration unit;

[0039] y = ax 2 + bx + c

[0040] Solve the coefficients a, b, c by minimizing the sum of squared errors:

[0041]

[0042] where x i , y i are the coordinates of the data points in the point cloud;

[0043] Spline curve fitting unit, which uses the B-spline curve for path fitting. The expression of the B-spline curve is:

[0044]

[0045] where P i is the control point, is the cubic Bernstein basis function:

[0046]

[0047] Use the least squares method to optimize the control points P i of the B-spline curve to minimize the fitting error. The control points P i correspond to the shape of the weld seam;

[0048] Through this step, the welding path can be more accurately fitted, improving the smoothness and accuracy of the welding path, reducing the path deviation during the welding process, and thus enhancing the welding quality;

[0049] Taking the welding of large structural parts in shipbuilding as an example, this module can accurately plan the welding path according to the complex weld seam shape, avoiding welding defects caused by unreasonable welding paths. Furthermore, the path optimization module includes the following units:

[0050] Path calculation unit: The A-star algorithm is used for local path optimization. The cost function of the A-star algorithm is:

[0051] f(n) = g(n) + h(n)

[0052] Among them, g(n) is the cumulative motion energy consumption from the starting point to the current point, usually the motion energy consumption from the current point to the next point;

[0053] Assume that the cost of each movement is related to the welding speed v and the torch attitude angle θ. Therefore:

[0054]

[0055] Among them, t is time, C v and C θ are constant coefficients related to the welding speed and angle change;

[0056] h(n) is a heuristic estimate, the shortest path from the current point to the end point:

[0057]

[0058] Among them, x and y are coordinate data;

[0059] Global optimization unit, using the dynamic programming algorithm to search for the global optimal path, with the constraint conditions:

[0060] The speed v min ≤ v ≤ v max , and the torch attitude angle θ min ≤ θ ≤ θ max ;

[0061] Among them, v min is the minimum value of the welding speed, v max is the maximum value of the welding speed, θ min is the minimum value of the angle change, θ max is the maximum value of the angle change;

[0062] Perform global optimal search for the path through dynamic programming.

[0063] The beneficial effect of adopting the above further solution is that this step solves the technical problem of finding the optimal welding path under the premise of meeting the welding process requirements. Compared with the prior art, the prominent beneficial effect is that it can significantly improve the welding efficiency, reduce the welding time and energy consumption, and at the same time ensure that the movement of the torch during the welding process is more reasonable, reducing the risk of welding deformation;

[0064] In the welding of aerospace components, this module can improve production efficiency and reduce production costs while ensuring high-precision welding.

[0065] Furthermore, the robot kinematics calculation module includes:

[0066] Forward kinematics calculation unit: Forward kinematics calculates the motion parameters of the welding torch through the position and attitude of the end effector of the six-degree-of-freedom robotic arm (1), and the forward kinematics model is defined using the D-H parameter method.

[0067] Forward kinematics equation:

[0068]

[0069] Where:

[0070] θ i is the angle of the i-th joint, α i is the twist angle of the i-th joint, a i is the link length of the i-th joint, d i is the offset of the i-th joint;

[0071] Inverse kinematics calculation unit: Inverse kinematics is used to calculate the desired position of the end of the welding torch and then reverse calculate the angle of each joint. The reverse calculation is performed through iterative solution using the Newton-Raphson method or numerical optimization method, and finally the angles of each joint θ1, θ2,.., θ6 are solved. After the solution is completed, the end effector of the robot reaches the desired spatial position and attitude:

[0072]

[0073] Where, T desired is the planned welding terminal position.

[0074] The beneficial effect of adopting the above further solution is that this step solves the technical problem of how to automatically adjust welding parameters under different welding conditions to ensure welding quality;

[0075] Compared with the prior art, the prominent beneficial effect is that it can adaptively adjust welding parameters in real time according to the actual welding situation, improve the stability of welding quality, reduce human intervention and welding defects caused by improper parameter settings;

[0076] In the welding of steel structure bridges, the thickness and material of the steel in different parts may vary. This module can automatically adjust parameters to ensure consistent welding quality;

[0077] Furthermore, the welding parameter adaptive adjustment module includes:

[0078] The welding parameter monitoring unit includes auxiliary sensors, and the auxiliary sensors include, but are not limited to, temperature sensors. The auxiliary sensors are used to monitor the welding parameters and the parameters of the workpiece to be welded in real time. The welding parameters include the molten pool morphology, arc state, and temperature information. The parameters of the workpiece to be welded include, but are not limited to, the material and thickness of the workpiece to be welded;

[0079] The welding parameter adjustment unit automatically adjusts the welding parameters according to the data feedback of the welding parameter monitoring unit. The welding parameters include current, voltage, welding speed, and wire feeding speed;

[0080] The intelligent control module, based on the data feedback of the welding parameter monitoring unit, adopts machine learning algorithms and fuzzy control algorithms, and adjusts the welding current and voltage according to the material and thickness of the workpiece to be welded, so as to realize the adaptive adjustment of welding parameters under different welding conditions.

[0081] The beneficial effect of adopting the above further solution is that this step solves the technical problem of accurately calculating the motion parameters in the motion control of the welding robot, so that the welding torch can accurately reach the specified position and posture. Compared with the prior art, the prominent beneficial effect lies in improving the accuracy and flexibility of the motion control of the welding robot, being able to adapt to more complex welding tasks, and ensuring the accurate welding of tiny solder joints in the welding of precision electronic equipment, thereby improving the yield rate of products.

[0082] Further, the intelligent control module includes:

[0083] The data fusion unit is used to receive in real time the molten pool morphology, arc state, temperature information, material and thickness data of the workpiece to be welded from the welding parameter monitoring unit, and perform normalization processing on them to generate a standardized input vector;

[0084] The machine learning model training unit uses supervised learning algorithms to train historical welding data, constructs a prediction model based on a deep neural network, and outputs optimized recommended values for welding current and voltage;

[0085] The fuzzy logic reasoning unit dynamically adjusts the welding speed and wire feeding speed based on a preset fuzzy rule base. The fuzzy rule base is defined by expert experience. The input variables include the molten pool width deviation, arc stability index, and material thermal conductivity, and the output variables are the welding speed correction coefficient and wire feeding speed increment;

[0086] The parameter collaborative optimization unit performs weighted fusion on the predicted values of the machine learning model and the fuzzy logic reasoning results to generate the final welding parameter combination, and online updates the model weights through reinforcement learning algorithms based on real-time feedback data;

[0087] The dynamic constraint module sets the safety threshold range of the current according to the material and thickness of the workpiece:

[0088] I min I ≤ I ≤ I max ;

[0089] Set the voltage fluctuation tolerance, ΔV ≤ 10%.

[0090] The beneficial effect of adopting the above further solution is that this step solves the technical problem of precisely controlling welding parameters under complex welding conditions and improving welding quality. Compared with the prior art, the prominent beneficial effect is that by integrating multiple algorithms and data processing methods, more precise control of welding parameters is achieved, enhancing the adaptive ability and stability of the system while improving welding quality. In the welding of new energy vehicle batteries, it can quickly adapt to the welding requirements of different batches of batteries, improving production efficiency and product quality.

[0091] Further, two fixtures and a welding torch bracket are respectively installed at the end of the six-degree-of-freedom robotic arm. A welding torch is assembled on the welding torch bracket. A sensor mounting seat and a probe are respectively clamped on the two fixtures. The laser sensor and the auxiliary sensor are installed on the sensor mounting seat. The probe is used to detect the thickness and material properties of the workpiece to be welded.

[0092] The beneficial effect of adopting the above further solution is that this step solves the technical problem of how to integrate multiple detection and execution components during the welding process to achieve comprehensive monitoring and precise welding of the welding process. Compared with the prior art, the prominent beneficial effect lies in the integration of multiple functions, improving the versatility and adaptability of the welding robot, facilitating the acquisition of more welding process information, and providing support for improving welding quality. In the precision welding of 3C products, it can simultaneously detect the material and thickness of the workpiece and use multiple sensors to monitor the welding process in real time to ensure welding accuracy and quality. BRIEF DESCRIPTION OF THE DRAWINGS

[0093] Figure 1 is the principle block diagram of an automated welding robot system applicable to different industries according to the present invention;

[0094] Figure 2 is the principle block diagram of the welding parameter adaptive adjustment module of the present invention;

[0095] Figure 3 is the principle block diagram of the path optimization module of the present invention;

[0096] Figure 4 is the structural schematic diagram of the six-degree-of-freedom robotic arm of the present invention.

[0097] In the drawings, the list of components represented by each reference numeral is as follows:

[0098] 1. Six-degree-of-freedom robotic arm; 2. Fixture; 3. Welding torch bracket. Detailed implementation manners

[0099] The principles and features of the present invention will be described below with reference to the accompanying drawings. The examples given are only used to explain the present invention and are not intended to limit the scope of the present invention.

[0100] The present invention provides the following preferred embodiments

[0101] As Figures 1-4 shown, an automated welding robot system applicable to different industries includes:

[0102] A weld seam recognition module for recognizing weld seams;

[0103] A welding path point extraction module for extracting welding path points;

[0104] A path optimization module for optimizing the welding path. The path optimization module obtains the three-dimensional contour data of the weld seam, processes the three-dimensional contour data and extracts the weld seam shape and position change parameters. Based on the three-dimensional contour data, it fits and smooths the weld seam shape and position change parameters, extracts the welding path, controls the welding robot torch to move along the optimal path, and uses an algorithm for path optimization;

[0105] A robot kinematics calculation module for calculating the kinematic parameters of the welding robot;

[0106] A welding parameter adjustment module for adaptively adjusting welding parameters;

[0107] A monitoring module for monitoring the quality of weld seams;

[0108] Each module communicates through an industrial bus.

[0109] An acquisition unit includes a laser sensor. The laser sensor scans the surface of the workpiece at a set frequency and scanning angle to obtain the three-dimensional contour data of the weld seam. First, median filtering is used to preliminarily process the collected raw data to remove discrete interference points such as pulse noise, and then Gaussian filtering is used to remove high-frequency noise to smooth the data. Then, the data is converted to a unified coordinate system through coordinate transformation, and the Canny algorithm is used to accurately extract the weld seam edge, and then the weld seam shape and position change are determined;

[0110] A point cloud data integration unit. The laser sensor emits a laser beam, and the reflected light is received by a CCD camera to form the three-dimensional point cloud data of the weld seam. The point cloud data is expressed as:

[0111] P = pi(xi, yi, zi), i = 1, 2,..., N

[0112] where pi(xi, yi, zi) represents the three-dimensional coordinates of the i-th point, and N is the total number of point cloud data;

[0113] The point cloud denoising unit uses Gaussian filtering to perform preliminary denoising on measurement errors and environmental noise, and then uses the RANSAC algorithm to fit the weld curve and remove outliers:

[0114] ax + by + cz + d = 0

[0115] where a, b, c, and d are fitting parameters, and x, y, and z represent the normal vector and intercept of the plane;

[0116] Calculate the distance from all points to the fitted plane:

[0117]

[0118] If D i <T threshold , then the point p i is regarded as an inlier, and outliers greater than the threshold are removed.

[0119] where T threshold is the distance threshold for determining whether a point is an inlier;

[0120] 0.5mm ≤ T threshold ≤ 2mm, and the specific value of T threshold is limited by the accuracy of the laser sensor and the shape of the weld.

[0121] This step solves the technical problem of accurately obtaining the shape and position information of the weld in a complex environment. Compared with the prior art, its beneficial effect is that it can more accurately identify the weld, reduce the influence of data noise interference on weld identification, and improve the accuracy and stability of weld identification;

[0122] For example, in an automobile manufacturing workshop, where the environmental light is complex and the mechanical vibration is large, this module can effectively extract the weld information accurately from the raw data with noise, providing a reliable basis for subsequent welding work.

[0123] In the welding scenario of an automobile engine block, the weld shape of the engine block is complex and it is in an environment of oil stain and vibration. The laser sensor scans the surface of the engine block at a frequency of 50 times per second and a scanning angle of ±45°;

[0124] The raw data is filtered by median filtering to remove the pulse noise caused by oil stain, and then Gaussian filtering is used to remove the noise generated by high-frequency vibration;

[0125] After unifying the coordinate system through coordinate transformation, the Canny algorithm is used to accurately extract the weld edge.

[0126] The point cloud data integration unit generates three-dimensional point cloud data. Through the point cloud denoising unit, abnormal points are removed according to the distance threshold (determined to be 1 mm based on the laser sensor accuracy and weld shape), and the weld seam is accurately identified, providing accurate weld position and shape information for subsequent welding work.

[0127] The welding path point extraction module includes the following units:

[0128] The weld shape analysis unit uses the least squares method to fit and calculate the weld center line from the three-dimensional point cloud data collected by the point cloud data integration unit;

[0129] y = ax 2 + bx + c

[0130] The coefficients a, b, and c are solved by minimizing the sum of squared errors:

[0131]

[0132] where x i , y i are the coordinates of the data points in the point cloud.

[0133] The spline curve fitting unit uses a B-spline curve for path fitting. The expression of the B-spline curve is:

[0134]

[0135] where P i is the control point, is the cubic Bernstein basis function:

[0136]

[0137] The least squares method is used to optimize the control points P i of the B-spline curve to minimize the fitting error. The control points P i correspond to the shape of the weld seam.

[0138] Through this step, the welding path can be more accurately fitted, improving the smoothness and accuracy of the welding path, reducing the path deviation during the welding process, and thus enhancing the welding quality;

[0139] In the welding scenario of large structural parts in shipbuilding, this module can accurately plan the welding path according to the complex weld shape, avoiding further welding defects caused by unreasonable welding paths. The path optimization module includes the following units:

[0140] In the welding scenario of wind turbine blades, the blade welds are complex curves. The weld shape analysis unit uses the least squares method to fit and calculate the weld center line from the collected three-dimensional point cloud data, determines the quadratic function coefficients, and accurately depicts the weld center line. The spline curve fitting unit uses B-spline curves for path fitting and optimizes the control points through the least squares method to minimize the fitting error. For example, in the fitting of a certain section of the weld, the optimized control points improve the smoothness of the welding path by 30%, effectively reducing the path deviation during welding and improving the welding quality.

[0141] Path calculation unit: The A-star algorithm is used for local path optimization. The cost function of the A-star algorithm is:

[0142] f(n) = g(n) + h(n)

[0143] where g(n) is the cumulative motion energy consumption from the starting point to the current point, usually the motion energy consumption from the current point to the next point;

[0144] Assume that the cost of each movement is related to the welding speed v and the welding torch attitude angle θ. Therefore:

[0145]

[0146] where t is time, C v and C θ are constant coefficients related to the welding speed and angle change.

[0147] h(n) is the heuristic estimate, the shortest path from the current point to the end point:

[0148]

[0149] where x, y are coordinate data.

[0150] Global optimization unit: The dynamic programming algorithm is used to search for the global optimal path. The constraint conditions are:

[0151] The speed v min ≤ v ≤ v max , and the welding torch attitude angle θ min ≤ θ ≤ θ max ;

[0152] where v min is the minimum value of the welding speed, v max is the maximum value of the welding speed, θ min is the minimum value of the angle change, θ max is the maximum value of the angle change;

[0153] The global optimal search of the path is carried out through dynamic programming.

[0154] This step solves the technical problem of finding the optimal welding path while meeting the requirements of the welding process. Compared with the prior art, the prominent beneficial effect is that it can significantly improve the welding efficiency, reduce the welding time and energy consumption, and at the same time ensure that the movement of the welding torch during the welding process is more reasonable, reducing the risk of welding deformation;

[0155] For example, in the welding of aerospace components, this module can improve the production efficiency and reduce the production cost while ensuring high-precision welding;

[0156] When welding aerospace engine components, extremely high requirements are placed on welding precision and efficiency;

[0157] The path calculation unit adopts the A-star algorithm to calculate the motion energy consumption according to the welding speed and the welding torch attitude angle;

[0158] When welding a certain key component, set the welding speed and the motion energy consumption coefficient C v = 0.5, and the welding torch attitude angle and the motion energy consumption coefficient C θ = 0.3, and combine the heuristic estimation with the coordinates to calculate the shortest path from the current point to the end point;

[0159] The global optimization unit adopts the dynamic programming algorithm;

[0160] Under the constraint conditions of the welding speed 10mm / s ≤ v ≤ 30mm / s and the welding torch attitude angle -15° ≤ θ ≤ 15°, search for the global optimal path;

[0161] After optimization, the welding time is shortened by 20%, the energy consumption is reduced by 15%, and the welding deformation is effectively controlled.

[0162] The robot kinematics calculation module includes:

[0163] The forward kinematics calculation unit: The forward kinematics calculates the motion parameters of the welding torch through the position and attitude of the end effector of the six-degree-of-freedom robotic arm (1), and uses the D-H parameter method to define the forward kinematics model;

[0164] The forward kinematics equation:

[0165]

[0166] Where:

[0167] θ i is the angle of the i-th joint, α i is the twist angle of the i-th joint, a i is the link length of the i-th joint, d i is the offset of the i-th joint;

[0168] Inverse Kinematics Calculation Unit: When inverse kinematics is used to calculate the desired position of the end of the welding torch and reverse-infer the angles of each joint, the reverse inference is iteratively solved by the Newton-Raphson method or the numerical optimization method, and finally the angles of each joint θ1, θ2,.., θ6 are solved. After the solution is completed, the end effector of the robot reaches the desired spatial position and attitude:

[0169]

[0170] Among them, T desired is the planned welding terminal position.

[0171] This step solves the technical problem of how to automatically adjust welding parameters under different welding conditions to ensure welding quality;

[0172] Compared with the prior art, the prominent beneficial effect is that it can adaptively adjust welding parameters in real time according to the actual welding situation, improve the stability of welding quality, reduce human intervention and welding defects caused by improper parameter settings;

[0173] In the welding of steel structure bridges, the thickness and material of steel in different parts may vary. This module can automatically adjust parameters to ensure consistent welding quality;

[0174] In the welding of building steel structures, the welding position and attitude are variable. The forward kinematics calculation unit uses the D-H parameter method to calculate the motion parameters of the welding torch according to the position and attitude of the end effector of the six-degree-of-freedom robotic arm 1;

[0175] When welding a certain high-rise steel structure node, by measuring the angles of each joint, the twist angle, the link length and the offset of the robotic arm, substituting them into the forward kinematics equation to calculate the motion parameters of the welding torch, and the inverse kinematics calculation unit uses the Newton-Raphson method to iteratively solve the angles of each joint according to the planned welding terminal position, so that the welding torch accurately reaches the specified position and attitude, ensuring stable welding quality;

[0176] The welding parameter adaptive adjustment module includes:

[0177] Welding Parameter Monitoring Unit, including auxiliary sensors. The auxiliary sensors include but are not limited to temperature sensors. The auxiliary sensors are used to monitor the welding parameters and the parameters of the workpiece to be welded in real time during the welding process. The welding parameters include the molten pool morphology, the arc state and the temperature information. The parameters of the workpiece to be welded include but are not limited to the material and thickness of the workpiece to be welded;

[0178] Welding Parameter Adjustment Unit, which automatically adjusts welding parameters according to the data feedback of the Welding Parameter Monitoring Unit. The welding parameters include current, voltage, welding speed and wire feeding speed;

[0179] The intelligent control module, based on the data feedback from the welding parameter monitoring unit, adopts machine learning algorithms and fuzzy control algorithms to adjust the welding current and voltage according to the material and thickness of the workpiece to be welded, so as to achieve the adaptive adjustment of welding parameters under different welding conditions.

[0180] This step solves the technical problem of accurately calculating the motion parameters in the motion control of the welding robot to make the welding torch accurately reach the specified position and posture. Compared with the existing technology, the prominent beneficial effect is that it improves the accuracy and flexibility of the motion control of the welding robot, can adapt to more complex welding tasks, and in the welding of precision electronic equipment, it can ensure the accurate welding of tiny solder joints and improve the yield rate of products.

[0181] The intelligent control module includes:

[0182] A data fusion unit, which is used to receive in real time the data of the molten pool shape, arc state, temperature information, material and thickness of the workpiece to be welded from the welding parameter monitoring unit, and perform normalization processing on it to generate a standardized input vector;

[0183] A machine learning model training unit, which uses supervised learning algorithms to train historical welding data, constructs a prediction model based on a deep neural network, and outputs the optimized recommended values of welding current and voltage;

[0184] A fuzzy logic inference unit, which dynamically adjusts the welding speed and wire feeding speed based on a preset fuzzy rule base. The fuzzy rule base is defined by expert experience. The input variables include the molten pool width deviation, arc stability index, and material thermal conductivity, and the output variables are the welding speed correction coefficient and wire feeding speed increment;

[0185] A parameter collaborative optimization unit, which performs weighted fusion on the prediction values of the machine learning model and the fuzzy logic inference results to generate the final welding parameter combination, and online updates the model weights through a reinforcement learning algorithm based on real-time feedback data;

[0186] A dynamic constraint module, which sets the safety threshold range of the current according to the material and thickness of the workpiece:

[0187] I min ≤I≤I max ;

[0188] Set the voltage fluctuation tolerance, ΔV≤10%.

[0189] This step solves the technical problem of precisely controlling welding parameters under complex welding conditions and improving welding quality. Compared with the prior art, the prominent beneficial effect is that by integrating multiple algorithms and data processing methods, more precise control of welding parameters is achieved, enhancing the adaptive capacity and stability of the system while improving welding quality. In the welding of new energy vehicle batteries, it can quickly adapt to the welding requirements of different batches of batteries, improving production efficiency and product quality.

[0190] In the welding of new energy vehicle battery modules, there are differences in the materials and thicknesses of different batches of battery modules.

[0191] The welding parameter monitoring unit uses auxiliary sensors such as temperature sensors and infrared sensors to monitor the molten pool shape, arc state, temperature information during the welding process, as well as the materials and thicknesses of the battery modules in real time.

[0192] The welding parameter adjustment unit automatically adjusts the current, voltage, welding speed, and wire feeding speed based on the monitored data.

[0193] The intelligent control module uses machine learning algorithms and fuzzy control algorithms to adjust the welding current and voltage according to the materials and thicknesses of the battery modules.

[0194] When welding a certain batch of aluminum alloy battery modules, the intelligent control module dynamically adjusts the welding speed and wire feeding speed based on the molten pool width deviation, arc stability index, and material thermal conductivity to ensure welding quality and improve production efficiency and product quality.

[0195] Two clamps 2 and a welding torch bracket 3 are respectively installed at the end of the six-degree-of-freedom robotic arm 1. A welding torch is assembled on the welding torch bracket 3. A sensor mounting seat and a probe are respectively clamped on the two clamps 2. A laser sensor and auxiliary sensors are installed on the sensor mounting seat. The probe is used to detect the thickness and material properties of the workpiece to be welded.

[0196] This step solves the technical problem of how to integrate multiple detection and execution components during the welding process to achieve comprehensive monitoring and precise welding of the welding process. Compared with the prior art, the prominent beneficial effect lies in the integration of multiple functions, improving the versatility and adaptability of the welding robot, facilitating the acquisition of more welding process information, providing support for improving welding quality. In the precision welding of 3C products, it can simultaneously detect the materials and thicknesses of workpieces and use multiple sensors to monitor the welding process in real time to ensure welding accuracy and quality.

[0197] In the welding of 3C products such as mobile phone motherboards, the welding accuracy requirements are extremely high and multi-faceted information needs to be obtained.

[0198] Two clamps 2 at the end of the six-degree-of-freedom robotic arm 1, one clamps the sensor mounting base, which integrates a laser sensor, a vision sensor and an auxiliary sensor, and the other clamps the probe. The laser sensor is used to detect the weld position, the vision sensor monitors the welding process in real time, and the auxiliary sensor monitors parameters such as welding temperature;

[0199] The probe detects the thickness and material of the mobile phone motherboard. During the welding process, each sensor works together to obtain welding process information in real time, providing support for precise welding, ensuring welding accuracy and quality, and improving the product yield rate;

[0200] The specific usage method steps of the present invention are as follows: In the manufacture of pressure vessels, the automated welding robot system of the present invention is implemented in the whole process as follows: First is weld seam recognition. The laser sensor of the acquisition unit is used to obtain the original three-dimensional contour data. After median filtering, Gaussian filtering for noise reduction and coordinate transformation, the Canny algorithm is used to extract the weld seam edge. The point cloud data integration unit forms point cloud data, and then it is processed by the point cloud denoising unit to accurately identify the weld seam according to the set threshold;

[0201] Next, the welding path point extraction module fits the weld seam center line through the weld seam shape analysis unit, and the spline curve fitting unit optimizes the path to improve the path smoothness and accuracy;

[0202] Then, the path calculation unit of the path optimization module performs local optimization using the A-star algorithm, and the global optimization unit searches for the global optimal path using the dynamic programming algorithm to reduce the risk of welding deformation;

[0203] After that, the forward kinematics calculation unit of the robot kinematics calculation module calculates the welding torch motion parameters using the D-H parameter method, and the inverse kinematics calculation unit controls the welding torch motion through iterative solution;

[0204] During the welding process, the monitoring unit of the welding parameter adaptive adjustment module monitors the welding and workpiece parameters in real time, the adjustment unit automatically adjusts the welding parameters, the intelligent control module uses machine learning and fuzzy control algorithms to adjust the current and voltage, and its internal units work together to generate the final parameter combination and update the model weights. The dynamic constraint module sets the parameter threshold;

[0205] Finally, the monitoring module monitors the weld quality in real time, and the feedback data is used to adjust the system in time to ensure that the welding quality meets the requirements of pressure vessel manufacturing.

[0206] In the description of the present invention, it should be understood that the orientation or positional relationship indicated by the terms "center", "longitudinal", "transverse", "length", "width", "thickness", "upper", "lower", "front", "rear", "left", "right", "vertical", "horizontal", "top", "bottom", "inner", "outer", "clockwise", "counterclockwise", "axial", "radial", "circumferential", etc. is based on the orientation or positional relationship shown in the drawings, and is only for the convenience of describing the present invention and simplifying the description, rather than indicating or implying that the device or element referred to must have a specific orientation, be constructed and operated in a specific orientation, and therefore should not be construed as a limitation on the present invention.

[0207] In addition, the terms "first" and "second" are only used for descriptive purposes and should not be construed as indicating or implying relative importance or implicitly specifying the quantity of the indicated technical features. Thus, features defined with "first" and "second" may explicitly or implicitly include at least one of such features. In the description of the present invention, the meaning of "a plurality" is at least two, such as two, three, etc., unless otherwise specifically and clearly defined.

[0208] The above are only the preferred embodiments of the present invention and are not intended to limit the present invention. Any modifications, equivalent replacements, improvements, etc. made within the spirit and principle of the present invention shall be included within the protection scope of the present invention.

Claims

1. An automated welding robot system suitable for different industries, characterized in that: include: A weld identification module, used for identifying welds; A welding path point extraction module is used to extract welding path points; A path optimization module is used to optimize the welding path. The path optimization module obtains the three-dimensional contour data of the weld, processes the three-dimensional contour data and extracts the weld shape and position change parameters, fits and smoothes the weld shape and position change parameters based on the three-dimensional contour data, extracts the welding path, controls the welding robot's welding gun to move along the optimal path, and uses an algorithm to optimize the path; Robot kinematics calculation module, used to calculate the kinematic parameters of the welding robot; Welding parameter adjustment module, used for adaptive adjustment of welding parameters; Monitoring module, used for weld quality monitoring; The modules communicate via an industrial bus.

2. The automatic welding robot system suitable for different industries according to claim 1 is characterized in that: The weld identification module includes the following units: The acquisition unit includes a laser sensor. The laser sensor scans the surface of the workpiece at a set frequency and scanning angle to obtain the three-dimensional contour data of the weld. The collected raw data is first processed by a median filter to remove discrete interference points such as pulse noise. The high-frequency noise is then removed by a Gaussian filter to smooth the data. The data is then converted to a unified coordinate system by coordinate transformation. The Canny algorithm is used to accurately extract the weld edge, thereby determining the shape and position changes of the weld. Point cloud data integration unit, the laser sensor emits a laser beam, and the CCD camera receives the reflected light to form the three-dimensional point cloud data of the weld. The point cloud data is expressed as: P = pi (xi, yi, zi), i = 1, 2, ..., N Among them, pi (xi, yi, zi) represents the three-dimensional coordinates of the i-th point, and N is the total number of point cloud data; The point cloud denoising unit uses Gaussian filtering to perform preliminary noise reduction on measurement errors and environmental noise, and then uses the RANSAC algorithm to fit the weld curve and remove abnormal points: ax+by+cz+d=0 Among them, a, b, c, d are fitting parameters, x, y and z represent the normal vector and intercept of the plane; Compute the distances of all points to the fitted plane: If D i <T threshold , then the point p i Treat them as inliers and remove outliers whose values ​​are larger than the threshold. Among them, T threshold is the distance threshold, used to determine whether a point is an interior point; 0.5mm≤T threshold ≤2mm, T threshold The specific value of is determined by the accuracy of the laser sensor and the shape of the weld.

3. The automatic welding robot system suitable for different industries according to claim 1 is characterized in that: The welding path point extraction module includes the following units: The weld shape analysis unit calculates the weld centerline by fitting the three-dimensional point cloud data collected by the point cloud data integration unit using the least square method; y=ax 2 +bx+c Solve for coefficients a, b, c by minimizing the sum of squared errors: Among them, x i ,y i is the coordinate of the data point in the point cloud; The spline curve fitting unit uses B-spline curve for path fitting. The expression of B-spline curve is: Among them, P i is the control point, is the cubic Bernstein basis function: The least square method is used to find the control point P of the B-spline curve. i Optimize to minimize the fitting error and control point P i Corresponding to the shape of the weld.

4. The automatic welding robot system applicable to different industries according to claim 1 is characterized in that: The path optimization module includes the following units: Path calculation unit: A-star algorithm is used for local path optimization. The cost function of A-star algorithm is: f(n)=g(n)+h(n) Among them, g(n) is the cumulative motion energy consumption from the starting point to the current point, usually the motion energy consumption of moving from the current point to the next point; Assume that the cost of each movement is related to the welding speed v and the welding gun posture angle θ, so: Where t is time, C v and C θ is a constant coefficient related to the change of welding speed and angle; h(n) is a heuristic estimate of the shortest path from the current point to the end point: Among them, x, y are coordinate data; The global optimization unit uses a dynamic programming algorithm to search for the global optimal path, and the constraints are: Speed min ≤v≤v max , welding gun posture angle θ min ≤θ≤θ max ; Among them, v min is the minimum welding speed, v max is the maximum welding speed, θ min is the minimum value of the angle change, θ max is the maximum value of the angle change; The global optimal path search is performed through dynamic programming.

5. The automatic welding robot system applicable to different industries according to claim 1 is characterized in that: The robot kinematics calculation module includes: Forward kinematics calculation unit: Forward kinematics calculates the motion parameters of the welding gun through the position and posture of the end effector of the six-degree-of-freedom robot (1), and uses the DH parameter method to define the forward kinematics model; Forward kinematics equation: in: θ i is the angle of the i-th joint, α i is the torsion angle of the i-th joint, a i is the connecting rod length of the ith joint, d i is the offset of the i-th joint; Inverse kinematics calculation unit: Inverse kinematics is used to calculate the desired position of the end of the welding gun, and the angle of each joint is inversely calculated. The inverse calculation is iteratively solved by the Newton-Raphson method or numerical optimization method, and finally the angles of each joint are solved θ1, θ2, .., θ6. After the solution is completed, the robot end effector reaches the desired spatial position and posture: Among them, T desired The planned welding terminal position.

6. The automatic welding robot system applicable to different industries according to claim 1, characterized in that: The welding parameter adaptive adjustment module includes: A welding parameter monitoring unit, comprising an auxiliary sensor, the auxiliary sensor including but not limited to a temperature sensor, the auxiliary sensor being used to monitor in real time the welding parameters during the welding process and the parameters of the workpiece to be welded, the welding parameters including the molten pool shape, arc state and temperature information, the parameters of the workpiece to be welded including but not limited to the material and thickness of the workpiece to be welded; A welding parameter adjustment unit, which automatically adjusts welding parameters according to data feedback from the welding parameter monitoring unit, wherein the welding parameters include current, voltage, welding speed and wire feeding speed; The intelligent control module, based on the data feedback from the welding parameter monitoring unit, adopts machine learning algorithm and fuzzy control algorithm to adjust the welding current and voltage according to the material and thickness of the workpiece to be welded, so as to achieve adaptive adjustment of welding parameters under different welding conditions.

7. The automatic welding robot system applicable to different industries according to claim 6 is characterized in that: The intelligent control module comprises: A data fusion unit is used to receive the molten pool morphology, arc state, temperature information, material and thickness data of the workpiece to be welded from the welding parameter monitoring unit in real time, and normalize them to generate a standardized input vector; The machine learning model training unit uses a supervised learning algorithm to train historical welding data, builds a prediction model based on a deep neural network, and outputs optimized recommended values ​​for welding current and voltage; A fuzzy logic reasoning unit dynamically adjusts the welding speed and wire feeding speed based on a preset fuzzy rule base, wherein the fuzzy rule base is defined by expert experience, and the input variables include the molten pool width deviation, arc stability index and material thermal conductivity, and the output variables are the welding speed correction coefficient and the wire feeding speed increment; The parameter collaborative optimization unit performs weighted fusion of the predicted value of the machine learning model and the result of fuzzy logic reasoning to generate the final welding parameter combination, and updates the model weights online through the reinforcement learning algorithm based on real-time feedback data; The dynamic constraint module sets the safe threshold range of the current according to the material and thickness of the workpiece: I min ≤I≤I max ; Set the voltage fluctuation tolerance, ΔV ≤ 10%.

8. The automatic welding robot system applicable to different industries according to claim 1 is characterized in that: Two clamps (2) and a welding gun bracket (3) are respectively mounted on the ends of the six-degree-of-freedom mechanical arm (1); a welding gun is mounted on the welding gun bracket (3); a sensor mounting seat and a probe are respectively mounted on the two clamps (2); the laser sensor and the auxiliary sensor are mounted on the sensor mounting seat; and the probe is used to detect the thickness and material properties of the workpiece to be welded.

Images