Intelligent bonding silver wire equipment control method and system based on Internet of Things

Abstract

The invention discloses an intelligent bonding silver wire equipment control method and system based on the Internet of Things, and relates to the technical field of equipment control. An intelligent bonding silver wire equipment control system based on the Internet of Things comprises an equipment data acquisition module, an equipment abnormity judgment module and an equipment linkage control module. According to the method, the bonding external environment data, the bonding production process data and the bonding equipment state data are collected in real time, and the silver wire surface data are combined, so that the precise control on the bonding silver wire equipment can be realized, the production parameters can be optimized, and the bonding quality can be improved; in combination with dynamic equipment control analysis and dynamic silver wire state analysis, an environment control factor, a production process control factor, an equipment state control factor and a silver wire bonding process influence factor can be adjusted in real time, so that equipment can adapt to different production conditions, and the production flexibility is improved.

Description

Technical Field

[0001] The present invention relates to the field of equipment control technology, and in particular to an intelligent silver bonding wire equipment control method and system based on the Internet of Things. Background Art

[0002] In the semiconductor packaging process, bonding silver wire is the core material connecting the chip and the external circuit. Its antioxidant ability directly determines the long-term reliability of electronic devices. High-purity silver wire is widely used in high-frequency communications and power modules due to its excellent electrical and thermal conductivity. However, silver materials easily react chemically with sulfides, oxygen, etc. in the air, forming silver sulfide or oxide layers on the surface. These compounds will significantly increase contact resistance, weaken heat dissipation efficiency, and reduce the mechanical strength of the bonding point, ultimately leading to device performance degradation or even failure.

[0003] Traditional protection methods mainly rely on isolating oxygen in an inert gas environment, such as continuously introducing high-purity nitrogen. However, this method consumes a lot of energy and is difficult to accurately control the low oxygen state in local areas. This is especially prone to blind spots in high-speed production or complex structure packaging. Alternatively, an anti-oxidation coating is deposited on the surface of the silver wire. However, the coating process usually requires complex vacuum equipment, which not only increases production costs but may also affect bonding strength.

[0004] Therefore, a real-time monitoring and dynamic adjustment capability of multiple factors such as temperature, humidity, and gas concentration is required to intelligently control the intelligent silver wire bonding equipment and improve the antioxidant capacity of the silver wire during wire drawing, high-temperature bonding, and storage and transportation. Summary of the Invention

[0005] The present invention aims to provide an intelligent silver bonding wire equipment control method and system based on the Internet of Things to improve the oxidation resistance of the silver bonding wire during operation.

[0006] A method for controlling an intelligent silver bonding wire device based on the Internet of Things comprises the following steps:

[0007] For working silver wire bonding equipment, collect bonding external environment data, bonding production process data and bonding equipment status data; for silver wire in the bonding state, collect silver wire surface data;

[0008] Perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, output an emergency abnormal equipment strategy and control the silver wire bonding equipment until it returns to a normal oxidation state. Otherwise, perform dynamic equipment control analysis based on the bonding external environment data, bonding production process data, and bonding equipment status data to obtain the bonding external environment control factor, bonding production process control factor, and bonding equipment status control factor. Perform dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor.

[0009] Comprehensive analysis is performed based on the silver bonding wire equipment control model to obtain real-time silver bonding wire equipment control parameters;

[0010] Use real-time silver bonding wire equipment control parameters to control the silver bonding wire equipment and perform subsequent operations.

[0011] As a preferred technical solution of the present invention, the specific steps of diagnosing abnormalities in a silver bonding wire device include:

[0012] Perform oxidation detection based on the silver wire surface data to obtain an oxidation wavelength offset; wherein the silver wire surface data includes the thickness of the silver wire surface oxide layer, the refractive index of the silver wire surface oxide layer, the silver wire surface refractive index, and the silver wire surface temperature;

[0013] An oxidation degree index equation is established based on the oxidation wavelength offset to obtain an oxidation judgment index; the oxidation judgment index is input into a pre-trained oxidation anomaly recognition unit for judgment;

[0014] If the judgment result is that it is in an abnormal oxidation state, the emergency abnormal equipment strategy is output; otherwise, the output is no abnormality and the next step is performed.

[0015] As a preferred technical solution of the present invention, the specific steps of performing dynamic equipment control analysis include:

[0016] Among them, the bonding external environment data includes sulfide concentration, oxygen content, and volatile matter concentration;

[0017] The bonding production process data includes temperature and humidity data, drawing speed, annealing temperature, and lubricant flow rate;

[0018] The bonding equipment status data contains equipment vibration spectrum and abrasive wear parameters;

[0019] Constructing a dynamic equipment control analysis unit for performing dynamic equipment control analysis;

[0020] The dynamic equipment control analysis unit is constructed using a multi-level linkage convolution structure, which includes a spatiotemporal alignment layer, a first-level linkage convolution layer, a second-level adaptive convolution layer, and a third-level causal convolution layer. The spatiotemporal alignment layer is used to synchronize the time and space of the input data. The first-level linkage convolution layer is used to extract local features based on a three-dimensional variable convolution kernel. The second-level adaptive convolution layer is used to perform cross-scale feature fusion based on multi-resolution pyramid convolution. The third-level causal convolution layer is used to fuse temporal information based on dilated causal convolution.

[0021] The bonding external environment data, bonding production process data and bonding equipment status data are input into the dynamic equipment control analysis unit for feature recognition to obtain the bonding external environment control factor, bonding production process control factor and bonding equipment status control factor.

[0022] As a preferred technical solution of the present invention, the specific steps of performing dynamic silver wire state analysis include:

[0023] A three-layer modeling analysis was performed on the silver wire surface data, oxidation wavelength offset, and oxidation judgment index to obtain oxidation dynamics analysis values, thermal-mechanical coupling analysis values, and photo-electric response analysis values.

[0024] The oxidation dynamics analysis values, thermal-mechanical coupling analysis values ​​and photo-electric response analysis values ​​were feature merged using self-attention units to obtain the influencing factors of the silver wire bonding process.

[0025] As a preferred technical solution of the present invention, the silver wire bonding equipment control model includes an oxidation risk prediction layer and a dynamic protection output layer;

[0026] The oxidation risk prediction layer is used to perform multivariate data linkage modeling based on the bonding external environment control factors, the bonding production process control factors, the bonding equipment state control factors, and the silver wire bonding process influencing factors, update the basic silver wire bonding equipment model, and obtain the predicted silver wire bonding equipment model; based on the predicted silver wire bonding equipment model, the oxidation risk prediction factor is output;

[0027] The dynamic protection output layer is used to output strategies based on oxidation risk prediction factors and obtain real-time silver wire bonding equipment control parameters.

[0028] As a preferred technical solution of the present invention, the specific steps of outputting policies in the dynamic protection output layer include:

[0029] Genetic algorithm is used to output iterative strategies based on oxidation risk prediction factors and predicted silver wire bonding equipment models to obtain the optimal real-time silver wire bonding equipment control parameters.

[0030] An intelligent silver bonding wire equipment control system based on the Internet of Things, comprising:

[0031] The equipment data acquisition module includes a data acquisition unit for collecting bonding external environment data, bonding production process data and bonding equipment status data for the working silver wire bonding equipment; and collecting silver wire surface data for the silver wire in the bonding state;

[0032] The equipment abnormality judgment module includes a factor judgment unit, which is used to perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, it outputs an emergency abnormal equipment strategy and controls the silver wire bonding equipment until the silver wire bonding equipment returns to a normal oxidation state. Otherwise, it performs dynamic equipment control analysis based on the bonding external environment data, the bonding production process data, and the bonding equipment status data to obtain the bonding external environment control factor, the bonding production process control factor, and the bonding equipment status control factor. It also performs dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor.

[0033] The equipment linkage control module includes a linkage control unit, which is used to perform comprehensive analysis based on the bonding wire equipment control model to obtain real-time bonding wire equipment control parameters; and use the real-time bonding wire equipment control parameters to control the bonding wire equipment and perform subsequent operations.

[0034] The present invention has the following advantages:

[0035] 1. The present invention collects bonding external environment data, bonding production process data and bonding equipment status data in real time, and combines them with silver wire surface data to achieve precise control of silver wire bonding equipment, optimize production parameters and improve bonding quality. Combined with dynamic equipment control analysis and dynamic silver wire status analysis, the present invention can adjust the environmental control factors, production process control factors, equipment status control factors and silver wire bonding process influencing factors in real time, so that the equipment can adapt to different production conditions and improve production flexibility.

[0036] 2. The present invention realizes precise dynamic control of silver wire bonding equipment by constructing a dynamic equipment control analysis method based on a multi-level linkage convolution structure, combining multi-source data fusion, spatiotemporal feature extraction and causal reasoning, and realizing comprehensive fusion of multi-source data to provide more accurate dynamic control decisions; through the spatiotemporal alignment layer, data from different sources are synchronized in time and spatially aligned to ensure that data with different sampling frequencies can be jointly analyzed to improve data consistency. BRIEF DESCRIPTION OF THE DRAWINGS

[0037] Figure 1 This is a structural diagram of an Internet of Things-based intelligent silver bonding wire equipment control system adopted in an embodiment of the present invention. DETAILED DESCRIPTION

[0038] In order to enable those skilled in the art to better understand the technical solutions in the present invention, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the accompanying drawings in the embodiments of the present invention.

[0039] Example 1, a method for controlling an intelligent silver bonding wire device based on the Internet of Things, comprising the following steps:

[0040] For working silver wire bonding equipment, collect bonding external environment data, bonding production process data and bonding equipment status data; for silver wire in the bonding state, collect silver wire surface data;

[0041] Perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, output an emergency abnormal equipment strategy and control the silver wire bonding equipment until it returns to a normal oxidation state. Otherwise, perform dynamic equipment control analysis based on the bonding external environment data, bonding production process data, and bonding equipment status data to obtain the bonding external environment control factor, bonding production process control factor, and bonding equipment status control factor. Perform dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor.

[0042] Data collection sources can be divided into four aspects: bonding external environment data, bonding production process data, bonding equipment status data, and silver wire surface data. Each type of data is automatically collected and transmitted based on IoT technology.

[0043] The specific steps for abnormal diagnosis of silver bonding wire equipment include:

[0044] Oxidation detection is performed based on silver wire surface data to obtain an oxidation wavelength offset. The silver wire surface data includes the thickness of the silver wire surface oxide layer, the refractive index of the silver wire surface oxide layer, the refractive index of the silver wire surface oxide layer, and the silver wire surface temperature. The thickness, refractive index, and refractive index of the silver wire surface oxide layer can be obtained from equipment such as an ellipsometer and a spectroscopic reflectometer. The silver wire surface temperature can be obtained from equipment such as an infrared thermometer. Acquisition of this data relies on optical detection, X-ray analysis, thermal measurement, and other technologies, and is combined with an Internet of Things sensing system for real-time data collection and remote monitoring.

[0045] The specific formula is: Where Δγ represents the oxidation wavelength offset, K represents the fiber coupling coefficient, d is the thickness of the silver wire surface oxide layer, and Z o is the refractive index of the oxide layer on the surface of the silver wire, Z Ag is the surface refractive index of the silver wire;

[0046] An oxidation degree index equation is established based on the oxidation wavelength offset to obtain an oxidation judgment index; the oxidation judgment index is input into a pre-trained oxidation anomaly recognition unit for judgment;

[0047] The oxidation index equation is γ0 represents the basic wavelength in the unoxidized state, E is the activation energy of the oxidation reaction, R is the ideal gas constant, T is the surface temperature of the silver wire; T0 is the reference temperature, is the rate of change of oxygen concentration, α is the oxidation kinetic acceleration factor, β is the nonlinear diffusion index; ΔI is the oxidation judgment index;

[0048] If the judgment result is that it is in an abnormal oxidation state, the emergency abnormal equipment strategy is output; otherwise, the output is no abnormality and the next operation is carried out;

[0049] By combining silver wire surface data with the oxidation wavelength offset calculation formula, the degree of oxidation can be accurately calculated and the oxidation state can be quantified. The oxidation wavelength offset formula is combined with the oxidation degree index equation to construct a complete oxidation diagnosis model, improving the scientificity and accuracy of the diagnosis. The oxidation judgment index is used as a key parameter to achieve quantitative analysis of the oxidation state of the silver wire, avoiding the subjective errors of traditional reliance on experience-based judgment. The oxidation anomaly recognition unit automatically identifies oxidation anomalies, reduces human intervention, and improves diagnostic efficiency. Combined with real-time oxidation monitoring data, it can issue an early warning when the silver wire is slightly oxidized, preventing oxidation from worsening and reducing the risk of equipment failure. The emergency abnormal equipment strategy is adopted to take immediate measures when oxidation anomalies are detected to prevent further damage and improve the stability of the production line.

[0050] Through precise oxidation detection and intelligent control, bonding failures caused by silver wire oxidation can be effectively reduced, product consistency and reliability can be improved, and scrap rates can be reduced. The oxidation process can be predicted by combining the nonlinear diffusion index to make process parameters more precise, thereby improving bonding strength and product quality.

[0051] The pre-trained oxidation anomaly recognition unit uses machine learning or deep learning models to construct a high-precision oxidation anomaly detection unit through steps such as data collection, feature extraction, model training, and optimization. First, IoT devices such as spectrometers, high-precision cameras, and temperature sensors are used to collect silver wire surface data, including oxide layer thickness, refractive index, temperature, and oxidation wavelength offset, and annotate whether there is oxidation anomaly. Second, feature engineering is performed to extract key features such as oxidation wavelength offset, oxidation degree index, and temperature change rate. These features are input into the machine learning model or deep learning model to divide the data set. The model is trained using cross-validation, and hyperparameters are optimized to improve classification accuracy. The model performance is verified on a test data set, and the accuracy, recall rate, and F1 score are calculated to ensure the high robustness of the model. Ultimately, the oxidation anomaly recognition unit is obtained, realizing real-time oxidation anomaly detection and adaptive control.

[0052] The specific steps for dynamic equipment control analysis include:

[0053] Among them, the bonding external environment data includes sulfide concentration, oxygen content, and volatile matter concentration;

[0054] The bonding production process data includes temperature and humidity data, drawing speed, annealing temperature, and lubricant flow rate;

[0055] The bonding equipment status data contains equipment vibration spectrum and abrasive wear parameters;

[0056] The collection of bonding external environment data, bonding production process data, and bonding equipment status data relies on high-precision sensors, industrial testing equipment, and IoT technology to ensure real-time and accurate data feedback to the control system. Sulfide concentration, oxygen content, and volatile matter concentration are collected by gas sensors such as electrochemical sensors, photoionization detectors (PIDs), and infrared gas sensors to monitor the impact of environmental components on bonding quality. Temperature and humidity data are collected by industrial-grade temperature and humidity sensors such as SHT3x and DHT22. Wire drawing speed is detected by photoelectric encoders and laser velocimeters. Annealing temperature is obtained by infrared thermometers and thermocouples. Lubricant flow is measured by mass flow meters such as Coriolis flowmeters and electromagnetic flowmeters. Equipment vibration spectrum is monitored using MEMS accelerometers and piezoelectric vibration sensors, and mold wear parameters are measured in real time using high-precision laser rangefinders and 3D profile scanners. These data can be combined with wireless communication technology and IoT technology to achieve intelligent, automated real-time monitoring and dynamic control.

[0057] Constructing a dynamic equipment control analysis unit for performing dynamic equipment control analysis;

[0058] The dynamic equipment control analysis unit is constructed using a multi-level linkage convolution structure, which includes a spatiotemporal alignment layer, a first-level linkage convolution layer, a second-level adaptive convolution layer, and a third-level causal convolution layer. The spatiotemporal alignment layer is used to synchronize the time and space of the input data. The first-level linkage convolution layer is used to extract local features based on a three-dimensional variable convolution kernel. The second-level adaptive convolution layer is used to perform cross-scale feature fusion based on multi-resolution pyramid convolution. The third-level causal convolution layer is used to fuse temporal information based on dilated causal convolution.

[0059] Inputting bonding external environment data, bonding production process data and bonding equipment status data into a dynamic equipment control analysis unit for feature recognition to obtain bonding external environment control factors, bonding production process control factors and bonding equipment status control factors;

[0060] The dynamic equipment control analysis method based on the multi-level linkage convolution structure realizes precise dynamic control of the silver bonding wire equipment by constructing a dynamic equipment control analysis unit and combining multi-source data fusion, spatiotemporal feature extraction and causal reasoning. It realizes the comprehensive fusion of multi-source data and provides more accurate dynamic control decisions. Through the spatiotemporal alignment layer, data from different sources are synchronized in time and spatially aligned to ensure that data with different sampling frequencies can be jointly analyzed to improve data consistency. The first-level linkage convolution layer uses a three-dimensional variable convolution kernel to extract local features, effectively capture subtle changes in equipment status, and improve the perception of key process parameters. The second-level adaptive convolution layer uses multi-resolution pyramid convolution to achieve cross-scale feature fusion, which can adapt to the state of silver bonding wire equipment under different working conditions and enhance the robustness of the model. The third-level causal convolution layer is based on dilated causal convolution to enhance the capture ability of time series information, enabling the system to infer the current equipment status and future trends based on historical data, and realize time series dependency modeling.

[0061] The dynamic equipment control analysis unit accurately identifies factors controlling the bonding external environment, the bonding production process, and the bonding equipment status, and establishes a multi-level correlation mapping to ensure adaptive adjustment of control parameters. A deep learning-driven causal analysis method ensures that control decisions not only rely on current data but also consider historical data and future trends, enhancing the predictability and adaptability of the control system.

[0062] The specific steps for dynamic silver wire state analysis include:

[0063] A three-layer modeling analysis was performed on the silver wire surface data, oxidation wavelength offset, and oxidation judgment index to obtain oxidation dynamics analysis values, thermal-mechanical coupling analysis values, and photo-electric response analysis values.

[0064] The self-attention unit is used to merge the oxidation dynamics analysis value, the thermal-mechanical coupling analysis value and the photoelectric response analysis value to obtain the influencing factors of the silver wire bonding process.

[0065] A dynamic silver wire state analysis method based on three-layer modeling and self-attention feature fusion achieves precise dynamic monitoring and optimized control of the silver wire bonding process through multi-scale physical modeling, deep feature extraction, and intelligent fusion. The oxidation dynamics analysis value quantifies the oxidation process of the silver wire surface based on the oxidation reaction kinetics model and analyzes the growth rate of the oxide layer and its impact. The thermal-mechanical coupling analysis value combines finite element analysis to calculate the deformation and stress distribution of the silver wire under high temperature and stress environment, and optimize the bonding process parameters. The photoelectric response analysis value is based on spectral analysis and electron migration theory to evaluate the changes in the optical and electrical properties of the silver wire under different oxidation states to optimize the conductivity and bonding quality of the silver wire.

[0066] A self-attention unit is used to merge the features of oxidation dynamics analysis values, thermal-mechanical coupling analysis values, and photoelectric response analysis values, automatically focusing on key features, enhancing the correlation analysis of oxidation, thermal-mechanical effects, and photoelectric characteristics, and improving modeling accuracy. Adaptive weight allocation technology is used to make the model more robust and adaptable to state analysis under different silver wire materials and production conditions. By strengthening nonlinear relationship modeling, the ability of silver wire state to predict bonding quality is improved, and the process parameter adjustment strategy is optimized.

[0067] Comprehensive analysis is performed based on the silver bonding wire equipment control model to obtain real-time silver bonding wire equipment control parameters;

[0068] Use real-time silver bonding wire equipment control parameters to control the silver bonding wire equipment and perform subsequent operations;

[0069] The silver bonding wire equipment control model includes an oxidation risk prediction layer and a dynamic protection output layer;

[0070] The oxidation risk prediction layer is used to perform multivariate data linkage modeling based on the bonding external environment control factors, the bonding production process control factors, the bonding equipment state control factors, and the silver wire bonding process influencing factors, update the basic silver wire bonding equipment model, and obtain the predicted silver wire bonding equipment model; based on the predicted silver wire bonding equipment model, the oxidation risk prediction factor is output;

[0071] The dynamic protection output layer is used to output strategies based on oxidation risk prediction factors and obtain real-time silver wire bonding equipment control parameters;

[0072] The specific steps for policy output in the dynamic protection output layer include:

[0073] Genetic algorithms are used to iterate strategy output based on oxidation risk prediction factors and predicted silver wire bonding equipment models to obtain the optimal real-time silver wire bonding equipment control parameters;

[0074] Through the oxidation risk prediction layer, multivariate data fusion is performed based on the bonding external environment control factors, bonding production process control factors, bonding equipment status control factors and silver wire bonding process influencing factors to improve the global perception of the bonding process; the basic silver wire bonding equipment model is updated in real time using multivariate data linkage modeling, and a predictive silver wire bonding equipment model is constructed to improve the accuracy of equipment status and oxidation risk prediction; by predicting the silver wire bonding equipment model and calculating the oxidation risk prediction factor, early warning of possible oxidation anomalies, environmental fluctuations, and equipment status degradation can be provided, thereby reducing silver wire damage and improving bonding quality; through the dynamic protection output layer, based on the oxidation risk prediction factor, iterative optimization is performed using a genetic algorithm to obtain the optimal real-time silver wire bonding equipment control parameters, ensuring that the equipment can adaptively adjust its operating status under complex working conditions; the optimization mechanism of the genetic algorithm can search for the global optimal strategy, avoid local optimal problems, and enable the equipment control parameters to be adaptively optimized with environmental changes, thereby improving the stability of the production process;

[0075] The basic silver bonding wire equipment model is constructed using digital twin technology, employing a data-driven, physical modeling, and machine learning optimization approach to achieve real-time virtual mapping and predictive control of the equipment. Dynamic calibration based on deep learning and data assimilation algorithms enables the digital twin to accurately simulate the equipment status. A combination of edge computing and cloud computing is used to achieve real-time data fusion and predictive optimization of the equipment, and reinforcement learning is used to continuously optimize equipment parameters, giving the twin model adaptive adjustment capabilities. Ultimately, the basic silver bonding wire equipment model can be used for real-time monitoring, predictive maintenance, abnormality warnings, and intelligent optimization control, improving the stability, production efficiency, and product quality of the silver bonding wire equipment.

[0076] In this embodiment, the silver wire bonding equipment may encounter situations such as excessive oxygen concentration or contamination of the silver wire surface. The strategy is to reduce the oxygen concentration, adjust the protective gas ratio or increase the environmental filtration parameters to reduce the concentration of sulfides and volatiles. Or when the wire drawing speed is too fast or the annealing temperature is too low, resulting in increased brittleness of the silver wire, the wire drawing speed can be reduced to make the stress distribution of the silver wire uniform, the annealing temperature can be increased, the metallographic structure of the silver wire can be optimized, and the production parameters can be adjusted in combination with the machine learning prediction model to prevent breakage. Or when the lubricant flow control is abnormal and affects the bonding quality, the lubricant flow can be adjusted to ensure proper lubrication, the lubrication condition of the silver wire surface can be monitored, and the flow control can be optimized in combination with the visual inspection system. The specific control parameter values ​​are generated and output based on real-time data during the actual operation process.

[0077] Based on the real-time silver wire bonding equipment control parameters, comprehensive and intelligent optimization of the environment, production process, equipment operation, and equipment health can be carried out. For possible problems such as oxidation, wire drawing breakage, ultrasonic power fluctuations, lubrication abnormalities, equipment vibration abnormalities, temperature and humidity fluctuations, etc., automatic parameter adjustment, intelligent early warning, digital twin prediction, etc. can be used to improve production stability, equipment life and product quality, ultimately reducing production costs and improving manufacturing efficiency.

[0078] Example 2, an intelligent silver wire bonding equipment control system based on the Internet of Things, see Figure 1 Shown, including:

[0079] The equipment data acquisition module includes a data acquisition unit for collecting bonding external environment data, bonding production process data and bonding equipment status data for the working silver wire bonding equipment; and collecting silver wire surface data for the silver wire in the bonding state;

[0080] The equipment abnormality judgment module includes a factor judgment unit, which is used to perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, it outputs an emergency abnormal equipment strategy and controls the silver wire bonding equipment until the silver wire bonding equipment returns to a normal oxidation state. Otherwise, it performs dynamic equipment control analysis based on the bonding external environment data, the bonding production process data, and the bonding equipment status data to obtain the bonding external environment control factor, the bonding production process control factor, and the bonding equipment status control factor. It also performs dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor.

[0081] The equipment linkage control module includes a linkage control unit, which is used to perform comprehensive analysis based on the bonding wire equipment control model to obtain real-time bonding wire equipment control parameters; and use the real-time bonding wire equipment control parameters to control the bonding wire equipment and perform subsequent operations.

[0082] It should be understood that those skilled in the art may make improvements or modifications based on the above description, and all such improvements and modifications shall fall within the scope of protection of the appended claims. Any portion of this specification not described in detail is prior art known to those skilled in the art.

Claims

1. A method for controlling an intelligent silver bonding wire device based on the Internet of Things, characterized in that: The following steps are involved: For working silver wire bonding equipment, collect bonding external environment data, bonding production process data and bonding equipment status data; for silver wire in the bonding state, collect silver wire surface data; Perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, output an emergency abnormal equipment strategy and control the silver wire bonding equipment until it returns to a normal oxidation state. Otherwise, perform dynamic equipment control analysis based on the bonding external environment data, bonding production process data, and bonding equipment status data to obtain the bonding external environment control factor, bonding production process control factor, and bonding equipment status control factor. Perform dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor. Comprehensive analysis is performed based on the silver bonding wire equipment control model to obtain real-time silver bonding wire equipment control parameters; Use real-time silver bonding wire equipment control parameters to control the silver bonding wire equipment and perform subsequent operations.

2. The method for controlling an intelligent silver bonding wire device based on the Internet of Things according to claim 1, wherein: The specific steps for abnormal diagnosis of silver bonding wire equipment include: Perform oxidation detection based on the silver wire surface data to obtain an oxidation wavelength offset; wherein the silver wire surface data includes the thickness of the silver wire surface oxide layer, the refractive index of the silver wire surface oxide layer, the silver wire surface refractive index, and the silver wire surface temperature; An oxidation degree index equation is established based on the oxidation wavelength offset to obtain an oxidation judgment index; the oxidation judgment index is input into a pre-trained oxidation anomaly recognition unit for judgment; If the judgment result is that it is in an abnormal oxidation state, the emergency abnormal equipment strategy is output; otherwise, the output is no abnormality and the next step is performed.

3. The method for controlling an intelligent silver bonding wire device based on the Internet of Things according to claim 2, wherein: Specific steps for dynamic equipment control analysis, include: Among them, the bonding external environment data includes sulfide concentration, oxygen content, and volatile matter concentration; The bonding production process data includes temperature and humidity data, drawing speed, annealing temperature, and lubricant flow rate; The bonding equipment status data contains equipment vibration spectrum and abrasive wear parameters; Constructing a dynamic equipment control analysis unit for performing dynamic equipment control analysis; The dynamic equipment control analysis unit is constructed using a multi-level linkage convolution structure, which includes a spatiotemporal alignment layer, a first-level linkage convolution layer, a second-level adaptive convolution layer, and a third-level causal convolution layer. The spatiotemporal alignment layer is used to synchronize the time and space of the input data. The first-level linkage convolution layer is used to extract local features based on a three-dimensional variable convolution kernel. The second-level adaptive convolution layer is used to perform cross-scale feature fusion based on multi-resolution pyramid convolution. The third-level causal convolution layer is used to fuse temporal information based on dilated causal convolution. The bonding external environment data, bonding production process data and bonding equipment status data are input into the dynamic equipment control analysis unit for feature recognition to obtain the bonding external environment control factor, bonding production process control factor and bonding equipment status control factor.

4. The method for controlling an intelligent silver bonding wire device based on the Internet of Things according to claim 3, wherein: The specific steps for dynamic silver wire state analysis include: A three-layer modeling analysis was performed on the silver wire surface data, oxidation wavelength offset, and oxidation judgment index to obtain oxidation dynamics analysis values, thermal-mechanical coupling analysis values, and photo-electric response analysis values. The oxidation dynamics analysis values, thermal-mechanical coupling analysis values ​​and photo-electric response analysis values ​​were feature merged using self-attention units to obtain the influencing factors of the silver wire bonding process.

5. The method for controlling an intelligent silver bonding wire device based on the Internet of Things according to claim 4, wherein: The silver bonding wire equipment control model includes an oxidation risk prediction layer and a dynamic protection output layer; The oxidation risk prediction layer is used to perform multivariate data linkage modeling based on the bonding external environment control factors, the bonding production process control factors, the bonding equipment state control factors, and the silver wire bonding process influencing factors, update the basic silver wire bonding equipment model, and obtain the predicted silver wire bonding equipment model; based on the predicted silver wire bonding equipment model, the oxidation risk prediction factor is output; The dynamic protection output layer is used to output strategies based on oxidation risk prediction factors and obtain real-time silver wire bonding equipment control parameters.

6. The method for controlling an intelligent silver bonding wire device based on the Internet of Things according to claim 5, characterized in that: The specific steps for policy output in the dynamic protection output layer include: Genetic algorithm is used to output iterative strategies based on oxidation risk prediction factors and predicted silver wire bonding equipment models to obtain the optimal real-time silver wire bonding equipment control parameters.

7. An intelligent silver wire bonding equipment control system based on the Internet of Things, characterized in that: The system applies the method for controlling an intelligent silver bonding wire device based on the Internet of Things according to any one of claims 1 to 6, comprising: The equipment data acquisition module includes a data acquisition unit for collecting bonding external environment data, bonding production process data and bonding equipment status data for the working silver wire bonding equipment; and collecting silver wire surface data for the silver wire in the bonding state; The equipment abnormality judgment module includes a factor judgment unit, which is used to perform abnormal diagnosis on the silver wire bonding equipment. If it is in an abnormal oxidation state, it outputs an emergency abnormal equipment strategy and controls the silver wire bonding equipment until the silver wire bonding equipment returns to a normal oxidation state. Otherwise, it performs dynamic equipment control analysis based on the bonding external environment data, the bonding production process data, and the bonding equipment status data to obtain the bonding external environment control factor, the bonding production process control factor, and the bonding equipment status control factor. It also performs dynamic silver wire status analysis based on the silver wire surface data to obtain the silver wire bonding process influencing factor. The equipment linkage control module includes a linkage control unit, which is used to perform comprehensive analysis based on the bonding wire equipment control model to obtain real-time bonding wire equipment control parameters; and use the real-time bonding wire equipment control parameters to control the bonding wire equipment and perform subsequent operations.

Images