Intelligent welding system based on automatic control robot

Abstract

The invention discloses an intelligent welding system based on an automatic control robot. Belongs to the fields of deep learning and artificial intelligence application, robot welding technology and the like. The system comprises a welding robot, a control unit, a welding unit, a monitoring unit and a dust adsorption device; welding parameters are monitored in real time through a voltage sensor and a current sensor, and the stability of the welding process is ensured; the visibility sensor can be used for monitoring pollutants, and dust and harmful gas in the welding process are effectively reduced; the precision sensor is used for improving the welding precision and the precision of a welding path; the speed sensor is used for ensuring the consistency of the welding speed and assisting in weld joint optimal selection and a welding path model, the model adopts an ant colony algorithm to optimize a welding path, and the welding efficiency is improved; the welding finished product passes through the quality detection unit, automatic detection of the quality of the welding finished product is achieved through an intelligent judgment mechanism, the automation level of welding operation is improved, and the welding quality and the production efficiency are remarkably improved.

Description

Technical Field

[0001] The present invention belongs to the fields of deep learning and artificial intelligence applications, robot welding technology, etc., and relates to a new type of intelligent welding technology based on an automatic control robot; specifically, it relates to an intelligent welding system based on an automatic control robot. Background Art

[0002] At present, traditional welding technology mainly relies on manual operation and manual experience, and has problems such as harsh environment, high labor intensity, low production efficiency and unstable product quality; these problems limit the further development of welding technology and also pose a threat to the health and safety of operators; in contrast, intelligent welding technology, with its high precision, high efficiency and high stability, provides solutions to many shortcomings of traditional welding technology. Summary of the invention

[0003] In view of the above problems, the purpose of the present invention is to propose a new quality intelligent welding technology based on an automatic control robot. The automation and intelligent means of this technology reduce manual intervention, improve the consistency and stability of the welding process, thereby improving the welding quality and production efficiency; at the same time, the intelligent welding robot can work in dangerous environments such as high temperature and high pressure, reducing the risk of workers being exposed to harmful gases and thermal radiation, and improving work safety.

[0004] The technical solution of the present invention is: the intelligent welding technology based on the automatic control robot described in the present invention includes an automatic control robot system, and the system includes a welding robot, a control unit, a welding unit, a monitoring unit and a dust adsorption device connected to each other;

[0005] The mechanical arm of the welding robot is controlled by a control unit. The mechanical arm is composed of multiple joints. The execution tasks of different units are run by setting the parameters of the range of motion, load capacity and accuracy of the mechanical arm.

[0006] The control unit is used to process sensor data, execute welding strategies and control the movements of the robotic arm, and control the welding unit and the monitoring unit through the robotic arm;

[0007] The welding unit includes a current sensor, a voltage sensor, a speed sensor and an accuracy sensor;

[0008] The monitoring unit includes a visibility sensor;

[0009] The dust adsorption device adsorbs the dust and harmful gases monitored in the monitoring unit, and is processed by the connected pollutant treatment device.

[0010] Furthermore, the current sensor is used to detect the welding current, optimize the welding process and reduce welding defects by accurately controlling the current;

[0011] The voltage sensor is used to detect the output voltage of the welding power supply to ensure the stability of the voltage during the welding process;

[0012] The speed sensor is used to detect the moving speed of the robot arm and the welding tool to ensure the consistency of the welding speed;

[0013] The precision sensor is used to detect and correct the position accuracy of the robot arm to ensure the accuracy of the welding path;

[0014] The visibility sensor is used to monitor contaminants in the welding area.

[0015] Furthermore, a welding path model is created, and welding information is imported into the welding path model through a welding information receiving device. The path that the welding robot must plan before performing a welding task. The model mainly plans the path of the weld and adopts an ant colony algorithm to ensure the efficiency and quality of the welding process;

[0016] Finished welding quality inspection is used to assess whether the welding quality meets the standards. This usually includes non-destructive testing of welded joints, such as radiographic testing, ultrasonic testing, etc., as well as inspection of the welding appearance;

[0017] Finished welded products are classified as "passed" or "failed" based on the quality inspection results; finished welded products that pass can enter the next production link, while finished welded products that fail need to be reworked or scrapped.

[0018] Furthermore, the load capacity calculation formula of the mechanical arm is:

[0019]

[0020] Where, L is the load capacity of the robot arm, A m is the robot arm torque, L m is the welding load moment, C is the moment coefficient;

[0021] In addition, the accuracy calculation formula of the robotic arm is:

[0022]

[0023] Where AT p is the robot arm accuracy, Δx i , Δy i , Δz i is the deviation of the ith measurement, Δx c , Δy c , Δz c are the coordinates of the cluster center.

[0024] Furthermore, the calculation formula for detecting the welding current is:

[0025] I=K×d

[0026] In the formula, I is the welding current, K is the empirical coefficient, and d is the electrode diameter;

[0027] The output voltage calculation formula of the welding power supply is:

[0028]

[0029] Where U is the output voltage of the welding power supply;

[0030] The calculation formula of the welding speed is:

[0031]

[0032] Where v is the welding speed, k is the proportional constant, and t is the thickness of the welding material;

[0033] The position accuracy calculation formula of the robotic arm is:

[0034] ΔP=N a Δa+N α Δα+N d Δd+N θ Δθ

[0035] Where ΔP is the error between the actual position and the theoretical position coordinates, N a 、N α 、N d 、N θ are the coefficients corresponding to the errors of the connecting rod length a, the connecting rod rotation angle α, the connecting rod offset d, and the joint angle θ, respectively. Δa, Δα, Δd, and Δθ are the small deviations of the connecting rod length, the connecting rod rotation angle, the connecting rod offset, and the joint angle, respectively;

[0036] The monitoring calculation formula of the visibility sensor is:

[0037]

[0038] Where FFR is the welding dust generation rate, M is the mass of the filter paper before welding, M′ is the mass of the filter paper after welding, and Δt is the welding time.

[0039] Furthermore, the information processing calculation formula of the welding information receiving and processing device is as follows:

[0040] (1) After the device receives the data, it needs to preprocess the data. The preprocessing formula is as follows:

[0041]

[0042] Where, I', U', and v' are the welding current intensity value, welding voltage intensity value, and welding speed intensity value, respectively. 1 , U 1 , v 1 are the average values ​​of the readings of each sensor, σI, σU, and σv are the standard deviations of the readings of each sensor;

[0043] (2) Extract key features from the preprocessed data, such as the peak value of current and voltage, speed stability, etc. The feature extraction formula is as follows:

[0044] F I =max(I′)

[0045]

[0046] In the formula, F I , F U , F V represents variance, which is an indicator of the dispersion of each data. max(I′) is the maximum value of welding current, max(U′) is the maximum value of welding voltage, N is the total number of data points in the speed data set, and v i is the i-th data point in the velocity data set;

[0047] Importing key information features into the welding path model can more effectively complete the selection of welds.

[0048] Furthermore, the construction process of the welding path model is as follows:

[0049] (1) Construct a mathematical model of the weld. The starting and ending points of the weld are represented by coordinates. The starting point of the weld is P start =(x start ,y start ,z start ), the end point is P end =(x end ,y end ,z end );weld directional mark w se It is expressed as:

[0050]

[0051] (2) Construct a mathematical model for welding path planning, and the welding path optimization sequence is expressed as:

[0052] i=(i 1 ,i 2 ,...,i m )

[0053] (3) The no-load path length and weld length in the welding path are expressed as:

[0054]

[0055] In the formula, m represents the number of welds, i and j represent the welding sequence of welds, and d ij Indicates the no-load path length in the welding path, l i Indicates the length of the weld;

[0056] (3) Time cost function t during welding 1 The calculation formula is as follows:

[0057]

[0058] In the formula, v 1 is the moving speed of the robot on the empty path, v 2 is the speed at which the robot welds the weld;

[0059] (4) The formula of the path planning objective function is as follows:

[0060] f=μ·l p +λ·E c

[0061] In the formula, l p is the path length, E c is the energy consumption, μ and λ are weight factors.

[0062] Furthermore, the specific steps of the ant colony algorithm for path optimization are as follows:

[0063] Constructing the mathematical model of the ant colony algorithm, the probability of the nth ant moving from weld i to weld j is expressed as:

[0064]

[0065] In the formula, represents the probability that the nth ant moves from weld i to weld j at time t; β represents the pheromone heuristic factor, which represents the relative importance of the remaining pheromone after the ant colony walks; γ represents the expected heuristic factor, which represents the relative importance of the path length; τ ij represents the residual pheromone concentration on path (i, j) at time t; δ ij Indicates the visibility of the path from weld i to weld j; allow n represents the weld selected by the nth ant in the next path selection;

[0066] Specifically, when the ants are selecting paths, the shorter the path distance between the current weld i and weld j, the higher the pheromone concentration on path ij, and the higher the probability that weld i will be selected for welding first. To prevent the pheromone concentration on all paths from being too high and interfering with the actions of the following ants, the pheromone concentration information on all paths is updated after each ant takes a step. For path ij, the pheromone concentration at the next moment can be updated based on the current pheromone:

[0067]

[0068] In the conventional ant colony algorithm, the pheromone inspiration factor β and the expected inspiration factor γ are constants. The values ​​of β and γ corresponding to all ants are the same. Therefore, the path planning effect of the algorithm depends on the values ​​of β and γ. Therefore, it is necessary to improve and optimize the ant colony algorithm. The improved algorithm reassigns the β, γ, and ρ parameters after each iteration. At the same time, it sorts the paths selected by all ants to increase the residual amount of pheromones of ants with shorter paths and avoid too many ants falling into the local optimal value. In order to give full play to the function of each ant in searching the route, different β and γ parameter values ​​are set for each ant. The calculation formula is as follows:

[0069]

[0070] Where rand(nums_ants,1) is a randomly generated matrix of size nums_ants*1, all values ​​are [0,1], and nums_ants is the number of ants set at the beginning;

[0071] When updating the global pheromone, the improved ant colony algorithm will normalize the pheromone and add 0.01 to all pheromones. The calculation formula for updating the pheromone is as follows:

[0072] τ ij (t+1)=(1-ρ)τ ij (t)+ρΔτ ij

[0073] In the formula, iter is the iteration coefficient. With the number of iterations, the volatility coefficient of pheromone will also change, and the minimum value is 0.1.

[0074] Furthermore, the calculation formula for the welding finished product quality detection is as follows:

[0075] Calculation formula for radiographic detection of welded joints:

[0076]

[0077] Attenuation coefficient, ΔT is the thickness difference in the direction of irradiation, and s is the scattering ratio, which represents the ratio of the scattered ray intensity to the transmitted ray intensity;

[0079] Calculation formula for ultrasonic testing of welded joints:

[0080]

[0081] In the formula, Δ 12 is the decibel difference in sound pressure of the reflected echo from two long horizontal holes of different diameters, D f1 and D f2 are the diameters of the two long horizontal holes respectively;

[0082] Finally, the welding quality is defined by inspecting the welding appearance of the finished welding product. The calculation formula for welding appearance inspection is as follows:

[0083]

[0084] Where Q is the welding appearance quality score, S d is the total score of surface defects, quantified according to defect type and severity; S p It is the total score of geometric dimension deviation, which is quantified according to the deviation between the actual dimension and the standard dimension; S i It is a score for the overall appearance, quantified based on the smoothness and glossiness of the weld joint; T is a preset threshold used to standardize the score, which is set according to specific welding standards and requirements.

[0085] Furthermore, the calculation formula of the welding finished product quality inspection result is as follows:

[0086]

[0087] Where R is the quality inspection result of the finished welding product, which is "pass" or "fail".

[0088] The beneficial effects of the present invention are: 1. Compared with traditional welding technology, the present invention realizes the automation of the welding process by integrating an automatic control robot, reduces the need for manual operation, reduces labor intensity, and improves the continuity and stability of welding operations; 2. Compared with traditional welding technology, the present invention uses high-precision sensors and advanced algorithms, such as current sensors, voltage sensors, speed sensors and precision sensors. This technology can accurately control welding parameters, optimize the welding process, reduce welding defects, and thus improve welding quality; the ant colony algorithm is used to optimize the welding path, improve welding efficiency, shorten the production cycle, and improve production efficiency; 3. Compared with traditional welding technology, the present invention effectively reduces the dust and harmful gases generated during the welding process, protects the health of operators, reduces pollution to the environment, and meets the requirements of green manufacturing and sustainable development. BRIEF DESCRIPTION OF THE DRAWINGS

[0089] Figure 1 It is a schematic diagram of the structure of the present invention;

[0090] Figure 2 It is the flow chart of the ant colony algorithm of the present invention;

[0091] Figure 3 It is the optimization flow chart of the ant colony algorithm of the present invention. DETAILED DESCRIPTION

[0092] The specific technical scheme of the present invention is further described in detail below with reference to specific examples.

[0093] As shown in the figure, the intelligent welding system based on the automatic control robot described in the present invention includes a welding robot, a control unit, a welding unit, a monitoring unit and a dust adsorption device; the welding parameters are monitored in real time by a voltage sensor and a current sensor to ensure the stability of the welding process;

[0094] The robotic arm of the welding robot is controlled by a control unit. The robotic arm is the main execution component of the welding robot, composed of multiple joints, and can achieve multi-degree-of-freedom movement. The execution tasks of different units are run by setting parameters such as the motion range, load capacity and accuracy of the robotic arm.

[0095] The control unit is the brain of the welding robot, responsible for processing sensor data, executing welding strategies and controlling the movements of the robotic arm. The robotic arm mainly controls the welding unit and the monitoring unit. The welding unit mainly includes a current sensor, a voltage sensor, a speed sensor and an accuracy sensor. The monitoring unit mainly includes a visibility sensor.

[0096] Current sensors are used to detect welding current, which is a key parameter affecting welding depth and welding heat input; by accurately controlling the current, the welding process can be optimized and welding defects such as lack of fusion or overheating can be reduced;

[0097] The voltage sensor is used to detect the output voltage of the welding power supply to ensure the stability of the voltage during the welding process; the fluctuation of welding voltage will affect the welding quality, so the voltage sensor is essential to maintain the consistency and reliability of the welding process;

[0098] The speed sensor is used to detect the moving speed of the robot arm and welding tools to ensure the consistency of the welding speed; the control of welding speed has an important influence on the formation of welding joints and welding quality;

[0099] Precision sensors are used to detect and correct the position accuracy of the robot arm to ensure the accuracy of the welding path; this is crucial for the production of high-quality welding products, especially in applications with high precision requirements;

[0100] Visibility sensors are used to monitor contaminants in the welding area, especially dust and harmful gases generated during automated welding processes.

[0101] The dust adsorption device is used to adsorb pollutants such as dust and harmful gases monitored in the monitoring unit, and the pollutants are treated by the pollutant treatment device.

[0102] Create a welding path model and import the welding information into the welding path model through the welding information receiving device. The welding robot must plan the path before performing the welding task. The model mainly plans the path of the weld and uses the ant colony algorithm to ensure the efficiency and quality of the welding process.

[0103] Finished welding quality inspection is used to assess whether the welding quality meets the standards; this usually includes non-destructive testing of welded joints, such as radiographic testing, ultrasonic testing, etc., as well as inspection of the weld appearance.

[0104] Finished welded products are classified as "passed" or "failed" based on the quality inspection results; finished welded products that pass can enter the next production link, while finished welded products that fail need to be reworked or scrapped.

[0105] Furthermore, the calculation formula of the load capacity of the robot arm is:

[0106]

[0107] Where, L is the load capacity of the robot arm, A m is the robot arm torque, L m is the welding load moment, C is the moment coefficient;

[0108] The calculation formula of the robot arm accuracy is:

[0109]

[0110] Where AT p is the robot arm accuracy, Δx i , Δy i , Δz i is the deviation of the ith measurement, Δx c , Δy c , Δz c are the coordinates of the cluster center point;

[0111] The calculation formula for detecting welding current is:

[0112] I=K×d

[0113] In the formula, I is the welding current, K is the empirical coefficient, and d is the electrode diameter;

[0114] The output voltage calculation formula of welding power supply is:

[0115]

[0116] Where U is the output voltage of the welding power supply;

[0117] The calculation formula for welding speed is:

[0118]

[0119] Where v is the welding speed, k is the proportional constant, and t is the thickness of the welding material;

[0120] The calculation formula of the position accuracy of the robot arm is:

[0121] ΔP=N a Δa+N α Δα+N d Δd+N θ Δθ

[0122] Where ΔP is the error between the actual position and the theoretical position coordinates, N a 、N α 、N d 、N θ are the coefficients corresponding to the errors of the connecting rod length a, the connecting rod rotation angle α, the connecting rod offset d, and the joint angle θ, respectively. Δa, Δα, Δd, and Δθ are the small deviations of the connecting rod length, the connecting rod rotation angle, the connecting rod offset, and the joint angle, respectively;

[0123] The monitoring calculation formula of the visibility sensor is:

[0124]

[0125] Where FFR is the welding dust generation rate, M is the mass of the filter paper before welding, M′ is the mass of the filter paper after welding, and Δt is the welding time.

[0126] The construction process of the welding path model is as follows:

[0127] (1) Construct a mathematical model of the weld. The starting point and end point of the weld can be represented by coordinates. Let the starting point of the weld be P start =(x start ,y start ,z start ), the end point is P end =(x end ,y end ,z end );weld directional mark w se It is expressed as:

[0128]

[0129] (2) Construct a mathematical model for welding path planning. The welding path optimization sequence can be expressed as:

[0130] i=(i 1 ,i 2 ,...,i m )

[0131] The no-load path length and weld length in the welding path can be expressed as:

[0132]

[0133] In the formula, m represents the number of welds, i and j represent the welding sequence of welds, and d ij Indicates the no-load path length in the welding path, l i Indicates the length of the weld;

[0134] (3) Time cost function t during welding 1 The calculation formula is as follows:

[0135]

[0136] In the formula, v 1 is the moving speed of the robot on the empty path, v 2 is the speed at which the robot welds the weld;

[0137] (4) The formula of the path planning objective function is as follows:

[0138] f=μ·l p +λ·E c

[0139] In the formula, l pis the path length, E c is the energy consumption, μ and λ are weight factors;

[0140] The specific steps of the ant colony algorithm in path optimization are as follows:

[0141] Constructing the mathematical model of the ant colony algorithm, the probability of the nth ant moving from weld i to weld j can be expressed as:

[0142]

[0143] In the formula, represents the probability that the nth ant moves from weld i to weld j at time t; β represents the pheromone heuristic factor, which represents the relative importance of the remaining pheromone after the ant colony walks; the larger β is, the more likely the ants will choose the path they have walked before, and the randomness of the search path will decrease; γ represents the expected heuristic factor, which represents the relative importance of the path length. The larger γ is, the easier it is for the ant colony to choose the local shortest path, and the model is easy to obtain the local optimal result; τ ij represents the residual pheromone concentration on path (i, j) at time t; δ ij Indicates the visibility of the path from weld i to weld j; allow n represents the welds that the nth ant can choose in the next path selection;

[0144] When the ants are selecting paths, the shorter the path distance between the current weld i and weld j, the higher the pheromone concentration on path ij will be, and the higher the probability that weld i will be selected for welding first. To prevent the pheromone concentration on all paths from being too high and interfering with the actions of the following ants, the pheromone concentration information on all paths needs to be updated after each ant takes a step. For path ij, the pheromone at the next moment can be updated based on the current pheromone:

[0145]

[0146] In the conventional ant colony algorithm, the two parameters of pheromone inspiration factor β and expected inspiration factor γ are constants. The values ​​of β and γ for all ants are the same. Therefore, the path planning effect of the algorithm depends largely on the values ​​of β and γ. Inappropriate parameter values ​​will cause pheromones to spread all over the paths, making the later ants lose their way.

[0147] Therefore, it is necessary to improve and optimize the ant colony algorithm. The improved algorithm reassigns the β, γ, and ρ parameters after each iteration. At the same time, it sorts the paths selected by all ants to increase the residual amount of pheromones of ants with shorter paths and avoid too many ants falling into the local optimal value.

[0148] In order to give full play to the function of each ant in searching routes, different β and γ parameter values ​​are set for each ant. The calculation formula is as follows:

[0149]

[0150] Where rand(nums_ants,1) is a randomly generated matrix of size nums_ants*1, all values ​​are [0,1], and nums_ants is the number of ants set at the beginning;

[0151] When updating the global pheromone, the improved ant colony algorithm will normalize the pheromone and add 0.01 to all pheromones to prevent some ants from being misled by a small amount of pheromone. The calculation formula for updating pheromones is as follows:

[0152] τ ij (t+1)=(1-ρ)τ ij (t)+ρΔτ ij

[0153] In the formula, iter is the iteration coefficient. With the number of iterations, the volatility coefficient of pheromone will also change, and the minimum value is 0.1;

[0154] The improved ant colony algorithm allows more ants to leave pheromones as much as possible in the initial period of iteration, and increases the pheromone volatility coefficient to accumulate more pheromones as soon as possible. In the middle and late selection process, only the top 20%-30% of ants are allowed to leave pheromones, and the pheromone volatility coefficient is gradually reduced to 0.1 to ensure that the influence of the ants that take the optimal path is large enough. Finally, the ant path with the highest pheromone concentration is selected to obtain the optimal path τ ij(best) .

[0155] The calculation formula for the quality inspection of the finished welding product is as follows:

[0156] Calculation formula for radiographic detection of welded joints:

[0157]

[0158] In the formula, is the object contrast, which represents the ratio of the change in ray intensity to the original intensity, ΔB is the increment of ray intensity, B is the original intensity of the ray, ζ is the linear attenuation coefficient of the material, ζ' is the linear attenuation coefficient of the defect, ΔT is the thickness difference in the irradiation direction of the ray, and s is the scattering ratio, which represents the ratio of the scattered ray intensity to the transmitted ray intensity;

[0159] Calculation formula for ultrasonic testing of welded joints:

[0160]

[0161] In the formula, Δ 12 is the decibel difference in sound pressure of the reflected echo from two long horizontal holes of different diameters, D f1 and D f2 are the diameters of the two long horizontal holes respectively;

[0162] Finally, the welding quality is defined by inspecting the welding appearance of the finished welding product. The calculation formula for welding appearance inspection is as follows:

[0163]

[0164] Where, Q is the welding appearance quality score (percentage); S d is the total score of surface defects, which can be quantified according to defect type and severity; S p It is the total score of geometric dimension deviation, which can be quantified according to the deviation between the actual dimension and the standard dimension; S i It is a score for the overall appearance, which can be quantified based on the smoothness, glossiness, etc. of the weld joint; T is a preset threshold used to standardize the score, which can be set according to specific welding standards and requirements.

[0165] The calculation formula for the welding finished product quality inspection result is as follows:

[0166]

[0167] Where R is the quality inspection result of the finished welding product, which can be "pass" or "fail".

Claims

1. An intelligent welding system based on an automatic control robot, characterized in that: The invention comprises an automatic control robot system, wherein the system comprises a welding robot, a control unit, a welding unit, a monitoring unit and a dust adsorption device connected to each other; The mechanical arm of the welding robot is controlled by a control unit. The mechanical arm is composed of multiple joints. The execution tasks of different units are run by setting the parameters of the range of motion, load capacity and accuracy of the mechanical arm. The control unit is used to process sensor data, execute welding strategies and control the movements of the robotic arm, and control the welding unit and the monitoring unit through the robotic arm; The welding unit includes a current sensor, a voltage sensor, a speed sensor and an accuracy sensor; The monitoring unit includes a visibility sensor; The dust adsorption device adsorbs the dust and harmful gases monitored in the monitoring unit, and is processed by the connected pollutant treatment device.

2. The intelligent welding system based on automatic control robot according to claim 1 is characterized in that: The current sensor is used to detect the welding current, optimize the welding process by accurately controlling the current, and reduce welding defects; The voltage sensor is used to detect the output voltage of the welding power supply to ensure the stability of the voltage during the welding process; The speed sensor is used to detect the moving speed of the robot arm and the welding tool to ensure the consistency of the welding speed; The precision sensor is used to detect and correct the position accuracy of the robot arm to ensure the accuracy of the welding path; The visibility sensor is used to monitor contaminants in the welding area.

3. The intelligent welding system based on automatic control robot according to claim 1 is characterized in that: A welding information receiving device is connected to the welding unit, and welding information is imported into the created welding path model through the welding information receiving device; The welding robot plans the path before performing the welding task. The welding path model uses the ant colony algorithm to plan the weld path to ensure the quality inspection of the welding process; The quality inspection results are classified as "pass" or "fail". The welded products that pass enter the next production link, while the welded products that fail are reworked or scrapped.

4. The intelligent welding system based on automatic control robot according to claim 1 is characterized in that: The calculation formula of the load capacity of the robotic arm is: Where L is the load capacity of the robot arm, A m is the robot arm torque, L m is the welding load moment, C is the moment coefficient; In addition, the accuracy calculation formula of the robotic arm is: Where AT p is the robot arm accuracy, Δx i , Δy i , Δz i is the deviation of the ith measurement, Δx c , Δy c , Δz c are the coordinates of the cluster center.

5. The intelligent welding system based on automatic control robot according to claim 2 is characterized in that: The calculation formula for detecting the welding current is: I=K×d In the formula, I is the welding current, K is the empirical coefficient, and d is the electrode diameter; The output voltage calculation formula of the welding power supply is: Where U is the output voltage of the welding power supply; The calculation formula of the welding speed is: Where v is the welding speed, k is the proportional constant, and t is the thickness of the welding material; The position accuracy calculation formula of the robotic arm is: ΔP=N a Δa+N α D+N d Δd+N θ Dth Where ΔP is the error between the actual position and the theoretical position coordinates, N a 、N α 、N d 、N θ are the coefficients corresponding to the errors of the connecting rod length a, the connecting rod rotation angle α, the connecting rod offset d, and the joint angle θ, respectively. Δa, Δα, Δd, and Δθ are the small deviations of the connecting rod length, the connecting rod rotation angle, the connecting rod offset, and the joint angle, respectively; The monitoring calculation formula of the visibility sensor is: Where FFR is the welding dust generation rate, M is the mass of the filter paper before welding, M′ is the mass of the filter paper after welding, and Δt is the welding time.

6. The intelligent welding system based on automatic control robot according to claim 3 is characterized in that: The information processing calculation formula of the welding information receiving and processing device is as follows: (1) After the device receives the data, it needs to preprocess the data. The preprocessing formula is as follows: Where, I', U', v' are the welding current intensity value, welding voltage intensity value, welding speed intensity value, I1, U1, v1 are the average readings of each sensor, σI, σU, σv are the standard deviations of the readings of each sensor; (2) Extract key features from the preprocessed data; the feature extraction formula is as follows: F I =max(I′) In the formula, F I , F U , F V represents variance, which is an indicator of the dispersion of each data. max(I′) is the maximum value of welding current, max(U′) is the maximum value of welding voltage, N is the total number of data points in the speed data set, and v i is the i-th data point in the velocity data set.

7. The intelligent welding system based on automatic control robot according to claim 3 is characterized in that: The construction process of the welding path model is as follows: (1) Construct a mathematical model of the weld. The starting and ending points of the weld are represented by coordinates. The starting point of the weld is P. start =(x start ,y start ,z start ), the end point is P end =(x end ,y end ,z end );weld directional mark w se It is expressed as: (2) Construct a mathematical model for welding path planning, and the welding path optimization sequence is expressed as: i=(i1,i2,...,i m ) (3) The no-load path length and weld length in the welding path are expressed as: In the formula, m represents the number of welds, i and j represent the welding sequence of welds, and d ij Indicates the no-load path length in the welding path, l i Indicates the length of the weld; (4) The calculation formula of the time cost function t1 in the welding process is as follows: Where v1 is the moving speed of the robot in the unloaded path, and v2 is the speed of the robot welding the weld; (5) The formula of the path planning objective function is as follows: f=μ·l p +λ·E c In the formula, l p is the path length, E c is the energy consumption, μ and λ are weight factors.

8. The intelligent welding system based on automatic control robot according to claim 3 is characterized in that: The specific steps of the ant colony algorithm for path optimization are as follows: Constructing the mathematical model of the ant colony algorithm, the probability of the nth ant moving from weld i to weld j is expressed as: In the formula, represents the probability that the nth ant moves from weld i to weld j at time t; β represents the pheromone heuristic factor, which represents the relative importance of the remaining pheromone after the ant colony walks; γ represents the expected heuristic factor, which represents the relative importance of the path length; τ ij represents the residual pheromone concentration on path (i, j) at time t; δ ij Indicates the visibility of the path from weld i to weld j; allow n represents the weld selected by the nth ant in the next path selection; Specifically, when the ants are selecting paths, the shorter the path distance between the current weld i and weld j, the higher the pheromone concentration on path ij, and the higher the probability that weld i will be selected for welding first. To prevent the pheromone concentration on all paths from being too high and interfering with the actions of the following ants, the pheromone concentration information on all paths is updated after each ant takes a step. For path ij, the pheromone concentration at the next moment can be updated based on the current pheromone: In the conventional ant colony algorithm, the pheromone inspiration factor β and the expected inspiration factor γ are constants. The values ​​of β and γ corresponding to all ants are the same. Therefore, the path planning effect of the algorithm depends on the values ​​of β and γ. Therefore, it is necessary to improve and optimize the ant colony algorithm. The improved algorithm reassigns the β, γ, and ρ parameters after each iteration. At the same time, it sorts the paths selected by all ants to increase the residual amount of pheromones of ants with shorter paths and avoid too many ants falling into the local optimal value. In order to give full play to the function of each ant in searching the route, different β and γ parameter values ​​are set for each ant. The calculation formula is as follows: Where rand(nums_ants,1) is a randomly generated matrix of size nums_ants*1, all values ​​are [0,1], and nums_ants is the number of ants set at the beginning; When updating the global pheromone, the improved ant colony algorithm will normalize the pheromone and add 0.01 to all pheromones. The calculation formula for updating the pheromone is as follows: t ij (t+1)=(1-ρ)τ ij (t)+ρΔτ ij In the formula, iter is the iteration coefficient. With the number of iterations, the volatility coefficient of pheromone will also change, and the minimum value is 0.

1.

9. The intelligent welding system based on automatic control robot according to claim 3 is characterized in that: The calculation formula for the welding finished product quality inspection is as follows: Calculation formula for radiographic detection of welded joints: In the formula, is the object contrast, which represents the ratio of the change in ray intensity to the original intensity, ΔB is the increment of ray intensity, B is the original intensity of the ray, ζ is the linear attenuation coefficient of the material, ζ' is the linear attenuation coefficient of the defect, ΔT is the thickness difference in the irradiation direction of the ray, and s is the scattering ratio, which represents the ratio of the scattered ray intensity to the transmitted ray intensity; Calculation formula for ultrasonic testing of welded joints: In the formula, Δ 12 is the decibel difference in sound pressure of the reflected echo from two long horizontal holes of different diameters, D f1 and D f2 are the diameters of the two long horizontal holes respectively; Finally, the welding quality is defined by inspecting the welding appearance of the finished welding product. The calculation formula for welding appearance inspection is as follows: Where Q is the welding appearance quality score, S d is the total score of surface defects, quantified according to defect type and severity; S p It is the total score of geometric dimension deviation, which is quantified according to the deviation between the actual dimension and the standard dimension; S i It is a score for the overall appearance, quantified based on the smoothness and glossiness of the weld joint; T is a preset threshold used to standardize the score, which is set according to specific welding standards and requirements.

10. The intelligent welding system based on automatic control robot according to claim 3, characterized in that: The calculation formula of the welding finished product quality test result is as follows: Where R is the quality inspection result of the finished welding product, which is "pass" or "fail".

Images