Control management system and method for silver wire bonding equipment based on Internet of Things

Through the Internet of Things-based control and management system, the welding process is monitored in real time and high-precision control is carried out, and the problem of lack of real-time monitoring and feedback in automated bonded silver wire equipment is solved, and the timely discovery and precise repair of welding defects is achieved, and production is improved. Efficiency and welding accuracy.

CN118502329BActive Publication Date: 2024-11-12SHENZHEN ZHONGBAO NEW MATERIAL TECH CO LTD
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202410962593.3
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2024-07-18
Publication Date
2024-11-12
Estimated Expiration
2044-07-18

AI Technical Summary

Technical Problem

The existing automated bonded silver wire equipment lacks real-time monitoring and feedback mechanisms, and cannot promptly detect and correct defects in the welding process, with limited control accuracy, difficult to meet high-precision welding requirements, and lack effective fault diagnosis and repair mechanisms, resulting in the equipment Maintenance and fault repair take time and affect production efficiency.

Method used

Using an Internet of Things control and management system, the welding process is monitored in real time through data acquisition equipment and sensors, and the welding defect detection is performed using the YOLO model, combining multiple linear regression and gradient descent methods to analyze and repair wrong data, and using genetic algorithms to find the optimal repair solution , and high-precision control of welding wire supply, welding temperature, pressure, position and operating speed through PID controller, dual input PID controller and fuzzy controller.

Benefits of technology

Real-time monitoring of the welding process and timely detection of defects, rapid adjustment of welding parameters, significantly improving the efficiency of fault diagnosis and repair, and ensuring welding accuracy and production efficiency.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN118502329B_ABST
    Figure CN118502329B_ABST
Patent Text Reader

Abstract

The present invention discloses a control management system and method for silver wire bonding equipment based on the Internet of Things, and belongs to the technical field of the Internet of Things. The present invention collects data generated during the operation of the silver wire bonding equipment, stores the collected data in a local database and uploads it to the cloud; performs data preprocessing, integrates the data to form a silver wire bonding equipment data set; uses historical data to train the YOLO model, detects image data, detects welding defects, forms a silver wire bonding equipment defect data set and adds specific welding defect content accordingly; extracts parameter features from the silver wire bonding equipment defect data set to form a feature matrix, finds error data related to specific welding defects, repairs the error data, and uses a genetic algorithm to find the optimal error data repair solution; repairs based on the optimal error data repair solution, and forms a silver wire bonding equipment data set again; and controls the equipment based on the Internet of Things according to the silver wire bonding equipment data set.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of Internet of Things, and in particular to a control management system and method for application bonding silver wire equipment based on the Internet of Things. Background Art

[0002] Silver wire bonding equipment is widely used in the electronics manufacturing industry to connect chips and substrates with silver wires to form electrical connections. Traditional silver wire bonding equipment usually relies on manual operation and experience, and is prone to welding defects, such as loose solder joints, offset welding positions, inappropriate welding temperature and pressure, etc. These defects can cause electronic products to perform poorly or even fail. In order to improve production efficiency and product quality, the industry has gradually introduced automation and intelligent technologies, but the existing technologies still have some shortcomings.

[0003] Most of the existing automated silver wire bonding equipment lacks real-time monitoring and feedback mechanisms, and is unable to promptly detect and correct defects that occur during the welding process. Although some equipment has data acquisition functions, data analysis and utilization are insufficient, making it difficult to optimize welding parameters based on historical data. The existing control system has limited accuracy in welding speed, temperature, pressure, and position control, making it difficult to meet high-precision welding requirements. The lack of an effective fault diagnosis and repair mechanism results in time-consuming equipment maintenance and fault repair, affecting production efficiency. Summary of the invention

[0004] The purpose of the present invention is to provide a control management system and method for application bonding silver wire equipment based on the Internet of Things to solve the problems raised in the above background technology.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions:

[0006] A control and management method for silver wire bonding equipment based on the Internet of Things, the method comprising the following steps:

[0007] S100, using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire device, storing the collected data in a local database and uploading it to the cloud; performing data cleaning, data smoothing and data normalization, and integrating the data to form a silver bonding wire device data set;

[0008] S200, using the YOLO model as a machine vision model, using normal and defective welding images as historical data, training the YOLO model, and after training, detecting the image data in the silver bonding wire equipment data set, detecting welding defects, forming a silver bonding wire equipment defect data set and adding specific welding defect content accordingly;

[0009] S300, analyzing each parameter in the silver bonding wire equipment defect data set, extracting parameter features to form a feature matrix, and using a multivariate linear regression algorithm to find error data related to specific welding defects; using gradient descent to repair the error data, considering the relationship between different parameters in the silver bonding wire equipment defect data set, and using a genetic algorithm to find the optimal error data repair solution;

[0010] S400. When a welding defect is detected in the silver wire bonding device, repair is performed based on the optimal error data repair solution to form a data set of the silver wire bonding device again. Based on the data set of the silver wire bonding device, a PID controller, a dual-input PID controller and a fuzzy controller are used based on the Internet of Things to control the welding wire supply, welding temperature and pressure, welding position, and welding action speed and accuracy of the silver wire bonding device.

[0011] According to step S100, the data generated during the operation of the bonding silver wire equipment include welding wire feeding speed, welding wire length, welding temperature, welding pressure, welding position coordinates, welding action speed, welding action accuracy and image data; using a welding wire feeding speed sensor to monitor the welding wire feeding speed in real time, using a welding wire length sensor to measure the welding wire length, using a welding temperature sensor to monitor the welding temperature during the welding process, using a welding pressure sensor to measure the pressure applied during welding as the welding pressure, using a position sensor to detect the position coordinates of the welding head as the welding position coordinates, using a motion speed sensor to measure the welding action speed, using a motion accuracy sensor to monitor the welding action accuracy, and using a camera to collect image data during the welding process;

[0012] Use MySQL database to store collected data locally, design database tables and store data, upload data to the cloud, use cloud storage service AWS S3 to store image data, use cloud database AWS RDS to synchronize local database data to the cloud, and use message queue service AWS SQS to ensure the reliability of data upload;

[0013] The interpolation method is used to handle missing values, the Z-score method is used to detect and handle outliers, the sliding average method is used to smooth the time series data to reduce the impact of noise, the data is scaled to the range of [0, 1] to ensure that different parameters have the same scale, and the data after data cleaning, data smoothing and data normalization are integrated to form the bonding silver wire equipment data set.

[0014] According to step S200, normal and defective welding image data are collected, the images are labeled using LabelImg, the locations and types of welding defects are marked, a label file in YOLO format is generated, and the data set is divided into a training set and a validation set;

[0015] Download and compile the code base of the YOLO model, modify the YOLO configuration file, adjust the network structure and training hyperparameters, create a list file of training and validation data, and list the paths of all training and validation images; use the pre-trained weight file to start training the YOLO model. During the training process, the model will generate a new weight file and save the training progress; load the trained YOLO model weight file and configure the detection parameters. Use the trained YOLO model to detect the newly collected welding images, identify and mark welding defects; process each image and output the detection results, which include the detected defect category and location;

[0016] Batch process all images in the silver wire bonding equipment data set, perform defect detection on each image, organize and save the detection results, form a silver wire bonding equipment defect data set and add specific welding defect content accordingly, the specific welding defect content includes defect category and defect location.

[0017] The YOLO model is an end-to-end object detection algorithm. Its core idea is to treat the object detection task as a regression problem and directly predict bounding boxes and category probabilities. The input image is divided into SxS grids, each of which is responsible for detecting objects in the central area of ​​the grid. The feature map of the image is extracted through a series of convolutional layers. Predictions are made on the feature map, and each grid outputs B bounding boxes, each containing 5 values ​​(center coordinates x, y, width w, height h, confidence) and C category probabilities. The loss function of YOLO consists of three parts: positioning loss, confidence loss, and category loss.

[0018] The localization loss is used to calculate the difference between the predicted bounding box and the true bounding box. The confidence loss is used to calculate the confidence of whether an object exists in the predicted bounding box. The category loss is used to calculate the difference between the predicted category probability and the true category.

[0019] According to step S300, various parameters are extracted from the bonding silver wire equipment defect data set to construct a feature matrix X, where each row represents a sample, each column represents a parameter feature, and the target vector y represents a specific welding defect category. The expressions of X and y are as follows:

[0020] ;

[0021] ;

[0022] Among them, from arrive Respectively represent the first parameter feature of the first sample to the nth parameter feature of the mth sample, from arrive Respectively represent the defect categories of the 1st sample to the mth sample;

[0023] The multivariate linear regression model is used to analyze the relationship between the feature matrix X and the target vector y. The formula of the multivariate linear regression model is as follows:

[0024] ;

[0025] in, is the regression coefficient vector, is the error term;

[0026] Estimate regression coefficients using the least squares method , the formula is as follows:

[0027] ;

[0028] is the estimated regression coefficient vector, represents the transpose of the feature matrix X, Represents the inverse matrix of the covariance matrix of the feature matrix X;

[0029] By analyzing the residuals of the regression model, abnormal data points related to specific welding defects are identified. The abnormal data points represent erroneous data. The calculation formula of the residual Residual is as follows:

[0030] ;

[0031] According to step S300, the detected erroneous data is repaired and the error function is minimized using the gradient descent method. :

[0032] ;

[0033] in, Represents the predicted value of the i-th sample, and the calculation formula is , is the feature vector of the i-th sample; the gradient descent update rule is as follows:

[0034] ;

[0035] in, is the learning rate, which controls the step size of each update.

[0036] According to step S300, the method of using the genetic algorithm to find the optimal error data repair solution is as follows:

[0037] A set of candidate solutions is randomly generated, the set of candidate solutions represents a parameter combination, the fitness of each candidate solution is evaluated, the parent individual is selected for reproduction according to the fitness, new candidate solutions are generated through crossover operation, some candidate solutions are mutated to increase the diversity of solutions, and the selection, crossover and mutation operations are repeated until the stopping condition is met; wherein the fitness function is defined as follows:

[0038] ;

[0039] in, is the fitness function, which indicates the quality of the candidate solution; in order to maximize the fitness function, it is defined as the negative value of the error function;

[0040] The fitness function is based on the error function The relationship between the error function and the parameters is analyzed, and the optimal solution is found by optimizing the correlation between the error function and the parameters to form the optimal error data repair plan.

[0041] According to step S400, the wire feeding speed is obtained from the silver bonding wire equipment data set. and wire length , use PID controller to control the wire feeding speed and wire length according to the wire feeding speed and wire length. The formula is as follows:

[0042] ;

[0043] ;

[0044] ;

[0045] ;

[0046] Where t represents time, is the current wire feed speed error, which indicates the difference between the expected feed speed and the actual feed speed. is the current actual wire feed speed, is the set wire feeding speed, Is the wire length error, which indicates the difference between the set length and the actual length. is the current actual welding wire length, is the set wire length, is the comprehensive error, combining the speed error and the length error, a is the first weight coefficient, b is the second weight coefficient, a and b are used to balance the influence of the speed error and the length error, is the PID controller output, used to adjust the wire feed equipment, is the proportional gain, which indicates the influence of adjusting the current error on the control quantity. is the integral gain, which indicates the influence of the adjustment cumulative error on the control quantity and is used to eliminate the static error. is the differential gain, which indicates the influence of the rate of change of the adjustment error on the control quantity, and is used to suppress oscillation and improve the system response speed;

[0047] Get the welding position coordinates from the silver bonding wire equipment dataset , use PID controller to control welding position according to welding position coordinates, the formula is as follows:

[0048] ;

[0049] ;

[0050] ;

[0051] ;

[0052] ;

[0053] ;

[0054] in, yes The position error, yes The position error, yes The position error, is the set welding position coordinate, is the current x-axis value, is the current y-axis value, is the current z-axis value, It is for The PID controller output, It is for The PID controller output, It is for The PID controller output, , and are the proportional, integral and differential gains of the x-axis position control, , and They are the proportional, integral and differential gains of the y-axis position control, , and They are the proportional, integral and differential gains of the z-axis position control respectively.

[0055] According to step S400, the welding temperature Temp and welding pressure Pres are obtained from the bonding silver wire equipment data set, and a dual-input PID controller is used to control the welding temperature and pressure according to the welding temperature and pressure. The temperature error and pressure error The definition is as follows:

[0056] ;

[0057] ;

[0058] in, is the current actual welding temperature, is the set welding temperature, is the current actual welding pressure, is the set welding pressure;

[0059] The formula for the PID controller output is as follows:

[0060] ;

[0061] ;

[0062] in, is the PID controller output of the welding temperature, is the PID controller output of the welding pressure, , and are the proportional, integral and derivative gains of temperature control, , and They are the proportional, integral and derivative gains of the pressure control respectively.

[0063] According to step S400, the welding action speed and welding action accuracy are obtained from the silver wire bonding equipment data set; a fuzzy controller is used to control the welding action speed and accuracy, and an input variable speed error is defined. and precision error , output control variable and , design the following fuzzy rules:

[0064] if is negative, indicating that the speed is too fast. Decrease, indicating a slowdown; if is positive, indicating insufficient precision, then Increasing means improving accuracy.

[0065] According to step S400, in the IoT architecture, sensors need to be deployed to collect data on wire feeding speed, wire length, welding temperature, welding pressure, welding position coordinates, welding action speed, and welding action accuracy. Use Wi-Fi to transmit data to edge computing devices or the cloud, deploy data processing and control algorithms on edge devices, process sensor data in real time, and generate control instructions, store, analyze, and optimize data on the cloud platform, provide support for control algorithms, and perform remote monitoring and maintenance.

[0066] The control and management system of the silver wire bonding equipment based on the Internet of Things includes:

[0067] Data acquisition and processing module: including: data acquisition unit and data processing unit; wherein the data acquisition unit is responsible for using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire equipment, storing the collected data in the local database and uploading it to the cloud, and the data processing unit is responsible for data cleaning, data smoothing and data normalization, and integrating the data to form a data set of the silver bonding wire equipment;

[0068] Welding defect detection module: including: YOLO model training unit and defect detection unit; the YOLO model training unit is responsible for using the YOLO model as a machine vision model, using normal and defective welding images as historical data to train the YOLO model, and the defect detection unit is responsible for using the trained YOLO model to detect image data in the silver bonding wire equipment data set, detect welding defects, form a silver bonding wire equipment defect data set and add specific welding defect content accordingly;

[0069] Error data analysis and repair module: including: feature extraction unit, multivariate linear regression unit, gradient descent repair unit and optimal repair solution unit; among them, the feature extraction unit is responsible for analyzing each parameter in the bonding silver wire equipment defect data set, extracting parameter features to form a feature matrix, the multivariate linear regression unit is responsible for using the multivariate linear regression algorithm to find the error data related to the specific welding defect, the gradient descent repair unit is responsible for using gradient descent to repair the error data, considering the relationship between different parameters in the bonding silver wire equipment defect data set, and the optimal repair solution unit is responsible for using the genetic algorithm to find the optimal error data repair solution;

[0070] Control management module: including: solution repair unit and parameter control unit; wherein, the solution repair unit is responsible for repairing the bonding silver wire equipment based on the optimal error data repair solution when welding defects are detected, and forming the bonding silver wire equipment data set again; the parameter control unit is responsible for using PID controller, dual-input PID controller and fuzzy controller based on the Internet of Things to control the welding wire supply, welding temperature and pressure, welding position, welding action speed and accuracy of the bonding silver wire equipment according to the bonding silver wire equipment data set.

[0071] Compared with the prior art, the beneficial effects achieved by the present invention are:

[0072] The present invention monitors the welding process in real time through data acquisition equipment and sensors, promptly discovers welding defects, and quickly adjusts welding parameters through a feedback mechanism.

[0073] The present invention realizes high-precision control of welding wire supply, welding temperature and pressure, welding position, and welding action speed and precision by using a PID controller, a dual-input PID controller and a fuzzy controller.

[0074] The present invention utilizes the YOLO model to detect welding defects, analyzes and repairs erroneous data through multiple linear regression and gradient descent method, and adopts genetic algorithm to find the optimal repair solution, thus significantly improving the efficiency of fault diagnosis and repair. BRIEF DESCRIPTION OF THE DRAWINGS

[0075] The accompanying drawings are used to provide a further understanding of the present invention and constitute a part of the specification. Together with the embodiments of the present invention, they are used to explain the present invention and do not constitute a limitation of the present invention. In the accompanying drawings:

[0076] Figure 1 It is a schematic diagram of the steps of the control and management method of the application bonding silver wire equipment based on the Internet of Things of the present invention;

[0077] Figure 2 It is a system structure diagram of the control and management system of the application bonding silver wire equipment based on the Internet of Things of the present invention. DETAILED DESCRIPTION

[0078] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0079] See also Figure 1 and Figure 2 , the present invention provides a technical solution:

[0080] According to one embodiment of the present invention, Figure 1 The control and management method of the application bonding silver wire device based on the Internet of Things is shown in the schematic diagram of the steps. The control and management method of the application bonding silver wire device based on the Internet of Things includes the following steps:

[0081] S100, using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire device, storing the collected data in a local database and uploading it to the cloud; performing data cleaning, data smoothing and data normalization, and integrating the data to form a silver bonding wire device data set;

[0082] S200, using the YOLO model as a machine vision model, using normal and defective welding images as historical data, training the YOLO model, and after training, detecting the image data in the silver bonding wire equipment data set, detecting welding defects, forming a silver bonding wire equipment defect data set and adding specific welding defect content accordingly;

[0083] S300, analyzing each parameter in the silver bonding wire equipment defect data set, extracting parameter features to form a feature matrix, and using a multivariate linear regression algorithm to find error data related to specific welding defects; using gradient descent to repair the error data, considering the relationship between different parameters in the silver bonding wire equipment defect data set, and using a genetic algorithm to find the optimal error data repair solution;

[0084] S400. When a welding defect is detected in the silver wire bonding device, repair is performed based on the optimal error data repair solution to form a data set of the silver wire bonding device again. Based on the data set of the silver wire bonding device, a PID controller, a dual-input PID controller and a fuzzy controller are used based on the Internet of Things to control the welding wire supply, welding temperature and pressure, welding position, and welding action speed and accuracy of the silver wire bonding device.

[0085] According to step S100, the data generated during the operation of the bonding silver wire equipment include welding wire feeding speed, welding wire length, welding temperature, welding pressure, welding position coordinates, welding action speed, welding action accuracy and image data; using a welding wire feeding speed sensor to monitor the welding wire feeding speed in real time, using a welding wire length sensor to measure the welding wire length, using a welding temperature sensor to monitor the welding temperature during the welding process, using a welding pressure sensor to measure the pressure applied during welding as the welding pressure, using a position sensor to detect the position coordinates of the welding head as the welding position coordinates, using a motion speed sensor to measure the welding action speed, using a motion accuracy sensor to monitor the welding action accuracy, and using a camera to collect image data during the welding process;

[0086] Use MySQL database to store collected data locally, design database tables and store data, upload data to the cloud, use cloud storage service AWS S3 to store image data, use cloud database AWS RDS to synchronize local database data to the cloud, and use message queue service AWS SQS to ensure the reliability of data upload;

[0087] The interpolation method is used to handle missing values, the Z-score method is used to detect and handle outliers, the sliding average method is used to smooth the time series data to reduce the impact of noise, the data is scaled to the range of [0, 1] to ensure that different parameters have the same scale, and the data after data cleaning, data smoothing and data normalization are integrated to form the bonding silver wire equipment data set.

[0088] According to step S200, normal and defective welding image data are collected, the images are labeled using LabelImg, the locations and types of welding defects are marked, a label file in YOLO format is generated, and the data set is divided into a training set and a validation set;

[0089] Download and compile the code base of the YOLO model, modify the YOLO configuration file, adjust the network structure and training hyperparameters, create a list file of training and validation data, and list the paths of all training and validation images; use the pre-trained weight file to start training the YOLO model. During the training process, the model will generate a new weight file and save the training progress; load the trained YOLO model weight file and configure the detection parameters. Use the trained YOLO model to detect the newly collected welding images, identify and mark welding defects; process each image and output the detection results, which include the detected defect category and location;

[0090] Batch process all images in the silver wire bonding equipment data set, perform defect detection on each image, organize and save the detection results, form a silver wire bonding equipment defect data set and add specific welding defect content accordingly, the specific welding defect content includes defect category and defect location.

[0091] According to step S300, various parameters are extracted from the bonding silver wire equipment defect data set to construct a feature matrix X, where each row represents a sample, each column represents a parameter feature, and the target vector y represents a specific welding defect category. The expressions of X and y are as follows:

[0092] ;

[0093] ;

[0094] Among them, from arrive Respectively represent the first parameter feature of the first sample to the nth parameter feature of the mth sample, from arrive They represent the defect categories of the 1st sample to the mth sample respectively; in y, 0 means no defect and 1 means defect.

[0095] The multivariate linear regression model is used to analyze the relationship between the feature matrix X and the target vector y. The formula of the multivariate linear regression model is as follows:

[0096] ;

[0097] in, is the regression coefficient vector, is the error term;

[0098] Estimate regression coefficients using the least squares method , the formula is as follows:

[0099] ;

[0100] is the estimated regression coefficient vector, represents the transpose of the feature matrix X, Represents the inverse matrix of the covariance matrix of the feature matrix X;

[0101] By analyzing the residuals of the regression model, abnormal data points related to specific welding defects are identified. The abnormal data points represent erroneous data. The calculation formula of the residual Residual is as follows:

[0102] ;

[0103] According to step S300, the detected erroneous data is repaired and the error function is minimized using the gradient descent method. :

[0104] ;

[0105] in, Represents the predicted value of the i-th sample, and the calculation formula is , is the feature vector of the i-th sample; the gradient descent update rule is as follows:

[0106] ;

[0107] in, is the learning rate, set to 0.1, which controls the step size of each update.

[0108] According to step S300, the method of using the genetic algorithm to find the optimal error data repair solution is as follows:

[0109] A set of candidate solutions is randomly generated, the set of candidate solutions represents a parameter combination, the fitness of each candidate solution is evaluated, the parent individual is selected for reproduction according to the fitness, new candidate solutions are generated through crossover operation, some candidate solutions are mutated to increase the diversity of solutions, and the selection, crossover and mutation operations are repeated until the stopping condition is met; wherein the fitness function is defined as follows:

[0110] ;

[0111] in, is the fitness function, which indicates the quality of the candidate solution; in order to maximize the fitness function, it is defined as the negative value of the error function;

[0112] The fitness function is based on the error function The relationship between the error function and the parameters is analyzed, and the optimal solution is found by optimizing the correlation between the error function and the parameters to form the optimal error data repair plan.

[0113] In this embodiment, abnormal data points related to welding defects are identified through multivariate linear regression analysis. These data points may be caused by sensor failure, data recording errors, etc. The identified abnormal data points are repaired using the gradient descent method, significantly reducing the impact of erroneous data in the data set on the overall analysis. Through genetic algorithm optimization, the optimal erroneous data repair solution is found to improve the quality of the repaired data and the prediction accuracy of the model.

[0114] The repaired dataset has higher quality and accuracy, with significantly fewer data anomalies.

[0115] According to step S400, the wire feeding speed is obtained from the silver bonding wire equipment data set. and wire length , use PID controller to control the wire feeding speed and wire length according to the wire feeding speed and wire length. The formula is as follows:

[0116] ;

[0117] ;

[0118] ;

[0119] ;

[0120] Where t represents time, is the current wire feed speed error, which indicates the difference between the expected feed speed and the actual feed speed. is the current actual wire feed speed, is the set wire feeding speed, Is the wire length error, which indicates the difference between the set length and the actual length. is the current actual welding wire length, is the set wire length, is the comprehensive error, combining the speed error and the length error, a is the first weight coefficient, b is the second weight coefficient, a and b are used to balance the influence of the speed error and the length error, is the PID controller output, used to adjust the wire feed equipment, is the proportional gain, which indicates the influence of adjusting the current error on the control quantity. is the integral gain, which indicates the influence of the adjustment cumulative error on the control quantity and is used to eliminate the static error. is the differential gain, which indicates the influence of the rate of change of the adjustment error on the control quantity, and is used to suppress oscillation and improve the system response speed;

[0121] Get the welding position coordinates from the silver bonding wire equipment dataset , use PID controller to control welding position according to welding position coordinates, the formula is as follows:

[0122] ;

[0123] ;

[0124] ;

[0125] ;

[0126] ;

[0127] ;

[0128] in, yes The position error, yes The position error, yes The position error, is the set welding position coordinate, is the current x-axis value, is the current y-axis value, is the current z-axis value, It is for The PID controller output, It is for The PID controller output, It is for The PID controller output, , and are the proportional, integral and differential gains of the x-axis position control, , and They are the proportional, integral and differential gains of the y-axis position control, , and They are the proportional, integral and differential gains of the z-axis position control respectively.

[0129] According to step S400, the welding temperature Temp and welding pressure Pres are obtained from the bonding silver wire equipment data set, and a dual-input PID controller is used to control the welding temperature and pressure according to the welding temperature and pressure. The temperature error and pressure error The definition is as follows:

[0130] ;

[0131] ;

[0132] in, is the current actual welding temperature, is the set welding temperature, is the current actual welding pressure, is the set welding pressure;

[0133] The formula for the PID controller output is as follows:

[0134] ;

[0135] ;

[0136] in, is the PID controller output of the welding temperature, is the PID controller output of the welding pressure, , and are the proportional, integral and derivative gains of temperature control, , and They are the proportional, integral and derivative gains of the pressure control respectively.

[0137] According to step S400, the welding action speed and welding action accuracy are obtained from the silver wire bonding equipment data set; a fuzzy controller is used to control the welding action speed and accuracy, and an input variable speed error is defined. and precision error , output control variable and , design the following fuzzy rules:

[0138] if is negative, indicating that the speed is too fast. Decrease, indicating a slowdown; if is positive, indicating insufficient precision, then Increasing means improving accuracy.

[0139] According to step S400, in the IoT architecture, sensors need to be deployed to collect data on wire feeding speed, wire length, welding temperature, welding pressure, welding position coordinates, welding action speed, and welding action accuracy. Use Wi-Fi to transmit data to edge computing devices or the cloud, deploy data processing and control algorithms on edge devices, process sensor data in real time, and generate control instructions, store, analyze, and optimize data on the cloud platform, provide support for control algorithms, and perform remote monitoring and maintenance.

[0140] According to another embodiment of the present invention, Figure 2 As shown in the system structure diagram of the control and management system of the application bonding silver wire equipment based on the Internet of Things, the control and management system of the application bonding silver wire equipment based on the Internet of Things includes:

[0141] Data acquisition and processing module: including: data acquisition unit and data processing unit; wherein the data acquisition unit is responsible for using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire equipment, storing the collected data in the local database and uploading it to the cloud, and the data processing unit is responsible for data cleaning, data smoothing and data normalization, and integrating the data to form a data set of the silver bonding wire equipment;

[0142] Welding defect detection module: including: YOLO model training unit and defect detection unit; the YOLO model training unit is responsible for using the YOLO model as a machine vision model, using normal and defective welding images as historical data to train the YOLO model, and the defect detection unit is responsible for using the trained YOLO model to detect image data in the silver bonding wire equipment data set, detect welding defects, form a silver bonding wire equipment defect data set and add specific welding defect content accordingly;

[0143] Error data analysis and repair module: including: feature extraction unit, multivariate linear regression unit, gradient descent repair unit and optimal repair solution unit; among them, the feature extraction unit is responsible for analyzing each parameter in the bonding silver wire equipment defect data set, extracting parameter features to form a feature matrix, the multivariate linear regression unit is responsible for using the multivariate linear regression algorithm to find the error data related to the specific welding defect, the gradient descent repair unit is responsible for using gradient descent to repair the error data, considering the relationship between different parameters in the bonding silver wire equipment defect data set, and the optimal repair solution unit is responsible for using the genetic algorithm to find the optimal error data repair solution;

[0144] Control management module: including: solution repair unit and parameter control unit; wherein, the solution repair unit is responsible for repairing the bonding silver wire equipment based on the optimal error data repair solution when welding defects are detected, and forming the bonding silver wire equipment data set again; the parameter control unit is responsible for using PID controller, dual-input PID controller and fuzzy controller based on the Internet of Things to control the welding wire supply, welding temperature and pressure, welding position, welding action speed and accuracy of the bonding silver wire equipment according to the bonding silver wire equipment data set.

[0145] In this embodiment, multiple linear regression and genetic algorithm are used to find the erroneous data points related to specific welding defects, and the identified abnormal data points are repaired according to the scheme optimized by the genetic algorithm. The repair process includes adjusting parameter values ​​or deleting abnormal data points to reduce their impact on the overall data set. The repaired silver bonding wire equipment data set has higher data quality and is suitable for further control parameter optimization and welding operations.

[0146] Based on the Internet of Things technology, the consumption of welding wire is monitored, and the PID controller is used to adjust the wire feeding rate to ensure stable welding quality. The temperature and pressure during the welding process are monitored and adjusted by the dual-input PID controller to ensure that the welding temperature is within a safe range while maintaining appropriate welding pressure. The fuzzy controller is used to adjust the position of the welding head according to real-time feedback to ensure that the welding position is accurate and stable. Combining the PID controller and the fuzzy controller, the speed and accuracy of the welding action are optimized to meet the changes in different welding requirements.

[0147] It should be noted that, in this article, relational terms such as first and second, etc. are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Moreover, the terms "include", "comprise" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device including a series of elements includes not only those elements, but also other elements not explicitly listed, or also includes elements inherent to such process, method, article or device.

[0148] Finally, it should be noted that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art can still modify the technical solutions described in the aforementioned embodiments or replace some of the technical features therein by equivalents. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A control and management method for silver wire bonding equipment based on the Internet of Things, characterized in that: The method comprises the following steps: S100, using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire device, storing the collected data in a local database and uploading it to the cloud; performing data cleaning, data smoothing and data normalization, and integrating the data to form a silver bonding wire device data set; S200, using the YOLO model as a machine vision model, using normal and defective welding images as historical data, training the YOLO model, and after training, detecting the image data in the silver bonding wire equipment data set, detecting welding defects, forming a silver bonding wire equipment defect data set and adding specific welding defect content accordingly; S300, analyzing each parameter in the silver bonding wire equipment defect data set, extracting parameter features to form a feature matrix, and using a multivariate linear regression algorithm to find error data related to specific welding defects; using gradient descent to repair the error data, considering the relationship between different parameters in the silver bonding wire equipment defect data set, and using a genetic algorithm to find the optimal error data repair solution; According to step S300, various parameters are extracted from the bonding silver wire equipment defect data set to construct a feature matrix X, where each row represents a sample, each column represents a parameter feature, and the target vector y represents a specific welding defect category. The expressions of X and y are as follows: , ; Among them, from arrive Respectively represent the first parameter feature of the first sample to the nth parameter feature of the mth sample, from arrive Represent the defect categories of the 1st sample to the mth sample respectively; The multivariate linear regression model is used to analyze the relationship between the feature matrix X and the target vector y. The formula of the multivariate linear regression model is as follows: ; in, is the regression coefficient vector, is the error term; Estimate regression coefficients using the least squares method , the formula is as follows: ; is the estimated regression coefficient vector, represents the transpose of the feature matrix X, Represents the inverse matrix of the covariance matrix of the feature matrix X; By analyzing the residuals of the regression model, abnormal data points related to specific welding defects are identified. The abnormal data points represent erroneous data. The calculation formula of the residual Residual is as follows: ; According to step S300, the detected erroneous data is repaired and the error function is minimized using the gradient descent method. : ; in, Represents the predicted value of the i-th sample, and the calculation formula is , is the feature vector of the i-th sample; the gradient descent update rule is as follows: ; in, is the learning rate, which is used to control the step size of each update; According to step S300, the method of using the genetic algorithm to find the optimal error data repair solution is as follows: A set of candidate solutions is randomly generated, the set of candidate solutions represents a parameter combination, the fitness of each candidate solution is evaluated, the parent individual is selected for reproduction according to the fitness, new candidate solutions are generated through crossover operation, some candidate solutions are mutated to increase the diversity of solutions, and the selection, crossover and mutation operations are repeated until the stopping condition is met; wherein the fitness function is defined as follows: ; in, is the fitness function, which indicates the quality of the candidate solution; in order to maximize the fitness function, it is defined as the negative value of the error function; The fitness function is based on the error function The relationship between the error function and the parameters is optimized to find the optimal solution and form the optimal error data repair solution; S400. When a welding defect is detected in the silver wire bonding device, repair is performed based on the optimal error data repair solution to form a data set of the silver wire bonding device again; based on the silver wire bonding device data set, a PID controller is used to control the welding wire supply and welding position of the silver wire bonding device based on the Internet of Things, a dual-input PID controller is used to control the welding temperature and pressure of the silver wire bonding device, and a fuzzy controller is used to control the welding action speed and accuracy of the silver wire bonding device.

2. The control and management method of the application bonding silver wire equipment based on the Internet of Things according to claim 1 is characterized in that: According to step S100, the data generated during the operation of the bonding silver wire equipment include welding wire feeding speed, welding wire length, welding temperature, welding pressure, welding position coordinates, welding action speed, welding action accuracy and image data; using a welding wire feeding speed sensor to monitor the welding wire feeding speed in real time, using a welding wire length sensor to measure the welding wire length, using a welding temperature sensor to monitor the welding temperature during the welding process, using a welding pressure sensor to measure the pressure applied during welding as the welding pressure, using a position sensor to detect the position coordinates of the welding head as the welding position coordinates, using a motion speed sensor to measure the welding action speed, using a motion accuracy sensor to monitor the welding action accuracy, and using a camera to collect image data during the welding process; Use MySQL database to store collected data locally, design database tables and store data, upload data to the cloud, use cloud storage service AWS S3 to store image data, use cloud database AWS RDS to synchronize local database data to the cloud, and use message queue service AWS SQS to ensure the reliability of data upload; The interpolation method is used to handle missing values, the Z-score method is used to detect and handle outliers, the sliding average method is used to smooth the time series data to reduce the impact of noise, the data is scaled to the range of [0, 1] to ensure that different parameters have the same scale, and the data after data cleaning, data smoothing and data normalization are integrated to form the bonding silver wire equipment data set.

3. The control and management method of the application bonding silver wire equipment based on the Internet of Things according to claim 1 is characterized in that: According to step S200, normal and defective welding image data are collected, the images are labeled using LabelImg, the locations and types of welding defects are marked, a label file in YOLO format is generated, and the data set is divided into a training set and a validation set; Download and compile the code base of the YOLO model, modify the YOLO configuration file, adjust the network structure and training hyperparameters, create a list file of training and validation data, and list the paths of all training and validation images; use the pre-trained weight file to start training the YOLO model. During the training process, the model will generate a new weight file and save the training progress; load the trained YOLO model weight file and configure the detection parameters. Use the trained YOLO model to detect the newly collected welding images, identify and mark welding defects; process each image and output the detection results, which include the detected defect category and location; Batch process all images in the silver wire bonding equipment data set, perform defect detection on each image, organize and save the detection results, form a silver wire bonding equipment defect data set and add specific welding defect content accordingly, the specific welding defect content includes defect category and defect location.

4. The control and management method of the application bonding silver wire equipment based on the Internet of Things according to claim 1 is characterized in that: According to step S400, the wire feeding speed is obtained from the silver bonding wire equipment data set. and wire length , use PID controller to control the wire feeding speed and wire length according to the wire feeding speed and wire length. The formula is as follows: ; ; ; ; Where t represents time, is the current wire feed speed error, which indicates the difference between the expected feed speed and the actual feed speed. is the current actual wire feed speed, is the set wire feeding speed, Is the wire length error, which indicates the difference between the set length and the actual length. is the current actual welding wire length, is the set wire length, is the comprehensive error, combining the speed error and the length error, a is the first weight coefficient, b is the second weight coefficient, a and b are used to balance the influence of the speed error and the length error, is the PID controller output, used to adjust the wire feed equipment, is the proportional gain, which indicates the influence of adjusting the current error on the control quantity. is the integral gain, which indicates the influence of the adjustment cumulative error on the control quantity and is used to eliminate the static error. is the differential gain, which indicates the influence of the rate of change of the adjustment error on the control quantity, and is used to suppress oscillation and improve the system response speed; Get the welding position coordinates from the silver bonding wire equipment dataset , use PID controller to control welding position according to welding position coordinates, the formula is as follows: ; ; ; ; ; ; in, yes The position error, yes The position error, yes The position error, is the set welding position coordinate, is the current x-axis value, is the current y-axis value, is the current z-axis value, It is for The PID controller output, It is for The PID controller output, It is for The PID controller output, , and are the proportional, integral and differential gains of the x-axis position control, , and They are the proportional, integral and differential gains of the y-axis position control, , and They are the proportional, integral and differential gains of the z-axis position control respectively.

5. The control and management method of the application bonding silver wire equipment based on the Internet of Things according to claim 1 is characterized in that: According to step S400, the welding temperature Temp and welding pressure Pres are obtained from the bonding silver wire equipment data set, and a dual-input PID controller is used to control the welding temperature and pressure according to the welding temperature and pressure. The temperature error and pressure error The definition is as follows: ; ; in, is the current actual welding temperature, is the set welding temperature, is the current actual welding pressure, is the set welding pressure; The formula for the PID controller output is as follows: ; ; in, is the PID controller output of the welding temperature, is the PID controller output of the welding pressure, , and are the proportional, integral and derivative gains of temperature control, , and They are the proportional, integral and derivative gains of the pressure control respectively.

6. The control and management method of the application bonding silver wire equipment based on the Internet of Things according to claim 1 is characterized in that: According to step S400, the welding action speed and welding action accuracy are obtained from the silver wire bonding equipment data set; a fuzzy controller is used to control the welding action speed and accuracy, and an input variable speed error is defined. and precision error , output control variable and , design the following fuzzy rules: if is negative, indicating that the speed is too fast. Decrease, indicating a slowdown; if is positive, indicating insufficient precision, then Increasing means improving accuracy.

7. A control and management system for application bonding silver wire equipment based on the Internet of Things, using the control and management method for application bonding silver wire equipment based on the Internet of Things according to any one of claims 1 to 6, characterized in that: include: Data acquisition and processing module: including: data acquisition unit and data processing unit; wherein the data acquisition unit is responsible for using data acquisition equipment and sensors to collect data generated during the operation of the silver bonding wire equipment, storing the collected data in the local database and uploading it to the cloud, and the data processing unit is responsible for data cleaning, data smoothing and data normalization, and integrating the data to form a data set of the silver bonding wire equipment; Welding defect detection module: including: YOLO model training unit and defect detection unit; the YOLO model training unit is responsible for using the YOLO model as a machine vision model, using normal and defective welding images as historical data to train the YOLO model, and the defect detection unit is responsible for using the trained YOLO model to detect image data in the silver bonding wire equipment data set, detect welding defects, form a silver bonding wire equipment defect data set and add specific welding defect content accordingly; Error data analysis and repair module: including: feature extraction unit, multivariate linear regression unit, gradient descent repair unit and optimal repair solution unit; among them, the feature extraction unit is responsible for analyzing each parameter in the bonding silver wire equipment defect data set, extracting parameter features to form a feature matrix, the multivariate linear regression unit is responsible for using the multivariate linear regression algorithm to find the error data related to the specific welding defect, the gradient descent repair unit is responsible for using gradient descent to repair the error data, considering the relationship between different parameters in the bonding silver wire equipment defect data set, and the optimal repair solution unit is responsible for using the genetic algorithm to find the optimal error data repair solution; Control management module: including: solution repair unit and parameter control unit; wherein, the solution repair unit is responsible for repairing the bonding silver wire equipment based on the optimal error data repair solution when welding defects are detected, and forming the bonding silver wire equipment data set again; the parameter control unit is responsible for using the PID controller based on the Internet of Things to control the welding wire supply and welding position of the bonding silver wire equipment according to the bonding silver wire equipment data set, using a dual-input PID controller to control the welding temperature and pressure of the bonding silver wire equipment, and using a fuzzy controller to control the welding action speed and accuracy of the bonding silver wire equipment.