Intelligent bonding copper wire power control method and system based on clustering analysis model

Abstract

The invention relates to an intelligent bonding copper wire power control method and system based on a clustering analysis model, and the system comprises a data collection module which is used for obtaining multi-dimensional parameters related in a copper wire bonding process in real time; the data preprocessing module is used for carrying out normalization, noise reduction and feature extraction processing on the collected data; the clustering analysis module is used for clustering the feature data by using a clustering algorithm to generate a plurality of feature clusters; the power prediction module is used for predicting the required optimal power by using a deep learning algorithm based on the clustering result and historical data; the dynamic power adjusting module is used for predicting power in real time and adjusting power output; through the technical means of clustering analysis, deep learning and PID control, the problem of power control in the copper wire bonding process is effectively solved, the process stability and the production efficiency are remarkably improved, and the method has important application value.

Description

Technical Field

[0001] The present invention belongs to the field of power control and automation technology, and specifically relates to an intelligent power control method and system based on a cluster analysis model, which is particularly suitable for dynamic adjustment and optimization of precision power output in intelligent equipment, and can be widely used in copper wire bonding, semiconductor manufacturing, precision welding and other industrial fields that require high-precision power control. The present invention comprehensively uses cluster analysis, big data processing, deep learning and real-time control algorithms to provide an efficient and stable solution for power management under complex working conditions. Background Art

[0002] With the widespread application of copper wire bonding technology in semiconductor manufacturing and precision electronic assembly, the demand for power control technology continues to increase. However, existing technologies have obvious deficiencies in dynamic adaptability, data mining and utilization, control accuracy, multi-parameter optimization and quality assessment, and are difficult to meet the requirements of high-precision power control under complex working conditions. This is specifically reflected in the following aspects: Complexity of copper wire bonding power requirements: The power output requirements of the copper wire bonding process are highly dependent on specific working conditions, such as surface roughness, operating temperature, and bonding pressure. Fluctuations in these parameters directly affect power requirements. However, traditional processes usually use fixed power or simple adjustment methods, which lack dynamic adaptability and can easily lead to energy waste due to overpower or bonding failure due to insufficient power.

[0003] Lack of working condition classification and power matching: Under complex working conditions, different operating conditions require different power outputs. However, traditional power control methods lack the ability to intelligently classify working conditions and can only rely on manual experience or preset rules. As working conditions become more diverse and random, manual rules are difficult to adapt to dynamic changes, resulting in insufficient accuracy and adaptability of power regulation.

[0004] Limitations of data processing capabilities: The large amount of operating data generated during the copper wire bonding process, including real-time collected temperature, current, voltage, and pressure information, has not been fully utilized. Existing technologies usually rely on fixed parameters and simple real-time feedback for adjustment, ignoring the mining of historical data and the analysis of complex correlations, and are unable to extract deep operating characteristics from big data.

[0005] Application bottlenecks of deep learning algorithms: The potential of deep learning in power prediction has gradually emerged. It can generate power prediction models through large amounts of data training. However, current technology lacks a complete framework that combines deep learning with actual processes, especially in the coordinated control between power demand prediction and actual power output. There is no mature solution yet.

[0006] Insufficient dynamic control accuracy: The dynamic adjustment of power output usually relies on a simple PID control algorithm, but this algorithm has limited error compensation capabilities for complex working conditions. In actual processes, when working conditions change drastically, traditional PID controllers are prone to over-adjustment or under-adjustment, resulting in an increase in the deviation between power output and actual demand, further affecting bonding quality.

[0007] The disconnect between power control and quality assessment: Existing technologies usually treat power control and quality assessment as two independent links, lacking consistency and closed-loop optimization capabilities. During the copper wire bonding process, the power control effect is not combined with the bonding quality feedback in real time, making it difficult to adaptively adjust the power control strategy, thus affecting production efficiency and final quality.

[0008] Insufficient comprehensive optimization of multi-dimensional parameters: The copper wire bonding process involves a complex relationship between multiple parameters (such as temperature, pressure, and roughness) and power. However, the existing technology lacks a comprehensive optimization mechanism for the relationship between multi-dimensional parameters and usually adopts a single variable adjustment method, resulting in limited optimization of the overall performance of the system.

[0009] Singularity of power calculation: Traditional power calculations are mostly based on empirical formulas, which usually only consider linear combinations of a small number of parameters and ignore nonlinear factors and complex relationships. In actual working conditions, due to the existence of multiple nonlinear factors, it is difficult to accurately describe power requirements using only traditional formulas, which leads to inaccurate regulation results. Summary of the invention

[0010] The present invention provides a power control method and system for intelligent copper wire bonding based on a cluster analysis model, which combines a data acquisition module, a data preprocessing module, a cluster analysis module, a power prediction module, a dynamic power adjustment module, a power calculation module, and a quality assessment and optimization module, and solves the problems of low power control accuracy, poor dynamic adaptability, and unstable power regulation in the existing copper wire bonding process by combining cluster analysis, big data processing, deep learning algorithms, and PID control technology, thereby improving bonding quality, reducing energy consumption, and improving production efficiency. The following is a detailed introduction to each module: 1. Data acquisition module: Function: The data acquisition module is responsible for real-time acquisition of key parameters that affect power requirements during copper wire bonding, including temperature, pressure, and surface roughness; Implementation: By installing embedded sensors on the bonding equipment, various data in the process are monitored in real time, and the data is transmitted to the central control unit via wired (RS485) or wireless (Wi-Fi). In order to ensure high reliability and real-time data, the sensor collection cycle is milliseconds to ensure that changes in key working conditions can be accurately captured; step: After the system starts, initialize each sensor to ensure calibration; Temperature collection: The sensor detects the ambient temperature in real time; Pressure acquisition: The pressure sensor is installed on the bonding machine robot arm to record the applied pressure in real time and synchronize the processing with the robot arm movement; Roughness acquisition: The laser displacement sensor scans the surface of the copper wire, obtains the three-dimensional profile data and calculates the roughness value; The collected multidimensional data is stored in the local cache and uploaded to the central processing unit.

[0011] 2. Data preprocessing module Function: The data preprocessing module is responsible for denoising, normalizing, and feature extraction of the raw data transmitted by the data acquisition module to ensure the accuracy and consistency of the data; Implementation: Noise filtering: high-frequency noise in the data is removed by a low-pass filter (the cutoff frequency of the filter is set to 100 Hz to ensure that the noise is removed and the useful signal is retained); Data normalization: By unifying data of different dimensions into the same range to avoid the influence of different dimensions, the normalization formula is used: ; in: is the original data, and are the maximum and minimum values ​​of the data respectively; Feature extraction: extract the most relevant features from multiple sensor data through principal component analysis (PCA) dimensionality reduction technology; The calculation formula for PCA dimensionality reduction is: ; in: is the eigenvector matrix, is the feature data before dimensionality reduction, is the feature data after dimensionality reduction; step: Read real-time data from the data acquisition module; Noise filtering: high-frequency noise in the data is removed by a low-pass filter (the cutoff frequency of the filter is set to 100 Hz to ensure that the noise is removed and the useful signal is retained); Data normalization: Unify data of different dimensions into the same range to avoid the influence of different dimensions; Feature extraction: The most relevant features are extracted from multiple sensor data through principal component analysis (PCA) dimensionality reduction technique.

[0012] 3. Cluster analysis module Function: Group data through clustering algorithms to identify different operating modes and help with subsequent power prediction and adjustment; Implementation: K-means clustering algorithm: Use the K-means algorithm to perform cluster analysis on the data and form multiple categories based on different feature combinations; K-means clustering cluster error minimization goal: ; in: is the number of clusters, Cluster , Cluster The cluster center, is a data point; step: receiving processed data from a data preprocessing module; Initially set the K value (such as K=3) and use the K-means algorithm to randomly select the initial cluster center; Calculate the distance from the data point to the center of each cluster and assign the data point to the nearest cluster; Update the cluster center position until the cluster center no longer changes; Output the final clustering results as input for subsequent power prediction; Example: The system performs cluster analysis based on the copper wire surface roughness, applied pressure and temperature data, and obtains three clusters, corresponding to normal operating conditions, overheating conditions and low power conditions.

[0013] 4. Power prediction module Function: Use cluster analysis results, deep learning algorithms and minimized loss functions to optimize the prediction accuracy of the model to achieve accurate power and predict the optimal power requirements under different working conditions; Implementation: Deep neural network (DNN) model: The input is the feature data obtained by cluster analysis, and the output is the predicted power value; Training method: Use the minimization loss function to optimize the prediction accuracy of the model to achieve accurate power demand prediction; Minimize the loss function: ; in: To predict power, is the actual power, is the number of data points; For example, for a specific copper wire surface roughness, applied pressure and temperature conditions, the model predicts that the required power is 8W, and this power value will be used for subsequent power adjustments.

[0014] 5. Dynamic power adjustment module Function: According to the prediction value provided by the power prediction module, the working power of the equipment is adjusted in real time to ensure that the optimal power requirement is maintained during the copper wire bonding process; Implementation: PID controller: dynamically adjusts control parameters according to power error to optimize power output; step: The output of the receiving power prediction module is used to determine the target power value; Adjustments are made using a PID controller, which calculates the new power setting and adjusts the device power output in real time; Continuously monitor actual power and adjust control parameters based on feedback information; For example, if the power prediction module predicts a power of 8W and the current power is 7.5W, the PID controller will calculate the power difference and adjust the device power to the target value.

[0015] 6. Power calculation module Function: Based on real-time temperature, pressure, and surface roughness data, an empirical model of power demand is generated. The coefficients in the empirical model are determined through experiments, and the physical quantities under different working conditions are mapped to the actual power, which serves as the basic input of the dynamic power adjustment module; Implementation: Real-time measurement of temperature, pressure, and surface roughness data through sensors; The empirical coefficient is determined through the empirical model, and the physical quantity under different working conditions is mapped to the actual power, which is used as the basic input of the dynamic power adjustment module. The power calculation formula is: ; in: and is the empirical coefficient, For pressure, is the surface roughness, is the ambient temperature; step: Obtain temperature, pressure, and surface roughness data from sensors in real time; Calculate the actual power according to the power calculation formula; The calculation result is passed to the dynamic power adjustment module for adjusting the power output; 7.Quality assessment and optimization module Function: Based on multiple quality evaluation indicators (such as bonding strength, thermal stability, surface smoothness), the quality score is calculated by weighted average method, and the system parameters are adjusted according to the evaluation results to further optimize the cluster analysis and power prediction model, thereby improving the stability and production efficiency of the system; Implementation: Monitoring of bonding quality parameters (bond strength, thermal stability, surface smoothness); The quality score was calculated using the weighted average method; Optimize system parameters based on evaluation feedback, adjust cluster analysis and power prediction models to continuously improve final product quality and process stability; Example: If the quality inspection finds that the bonding strength of a batch of copper wires is low, the system will optimize the power setting based on historical data to ensure that the quality of the next bonding is up to standard; The quality score formula is as follows: ; in: is the bonding strength, For thermal stability, For surface smoothness, , , is the weight coefficient.

[0016] A power control method for an intelligent copper bonding wire based on a cluster analysis model comprises the following steps: Step 1: Data collection: The data collection module collects working condition data related to the copper wire bonding process in real time, including temperature, pressure, surface roughness environmental factors, obtains these data through sensors and monitoring equipment, and transmits them to the data preprocessing module; Step 2: Data preprocessing: Remove noise and normalize the raw data collected from the data acquisition module. Through the standardization method, the feature data of different dimensions are converted into a unified scale to ensure that the cluster analysis and power prediction modules can process the data correctly. Step 3: Cluster analysis: Use the K-means clustering algorithm to analyze the data and divide the collected operating data into several clusters according to similarity. Each cluster represents a specific working state or operating mode. In this way, the cluster analysis module provides classification features for subsequent power prediction and dynamic adjustment, and provides input based on actual working conditions for power prediction. Step 4: Power prediction: Based on the working condition cluster characteristics generated by the cluster analysis module, a deep neural network is used to predict power. This module uses the prediction model trained based on historical data and the cluster analysis results as input. The power demand during the bonding process is predicted through a regression algorithm. The power prediction results provide a reference for the system to adjust the actual power. Step 5: Dynamic power adjustment: The actual power is adjusted through the dynamic power adjustment module. The module uses the PID control algorithm to calculate the error between the target power and the actual power in real time, and adjusts the power output to ensure that the actual power is close to the predicted power, avoiding bonding quality problems caused by power fluctuations. Step 6. Power calculation: Calculate the reference power value based on environmental parameters, pressure, temperature and roughness factors. This module uses empirical formulas and combines process data in actual production to determine the power value required under different working conditions and provide input for the dynamic power adjustment module. Step 7. Quality assessment and optimization: Comprehensively evaluate the copper wire bonding quality based on multiple quality assessment indicators (such as bonding strength, thermal stability, and surface smoothness), calculate the quality score by weighted average method, and adjust the system parameters based on the evaluation results. Further optimize the cluster analysis and power prediction model to improve the system stability and production efficiency.

[0017] Beneficial Effects Compared with the prior art, the present invention provides a power control method and system for intelligent bonding copper wire based on cluster analysis model, which has the following beneficial effects: Improve power accuracy and stability in copper wire bonding process: The present invention realizes precise adjustment of power control by integrating data acquisition, preprocessing, cluster analysis and deep learning technology. By real-time monitoring of environmental parameters (such as temperature, pressure and surface roughness) during copper wire bonding, and predicting the required power value based on historical data and real-time calculation results, the system can dynamically adjust the power output to ensure the stability of power during each bonding process, which effectively avoids poor bonding caused by power fluctuations and improves the stability and consistency of copper wire bonding. Enhance the system's adaptability and intelligence: The system uses a cluster analysis algorithm to divide the operating data into multiple feature clusters. Based on different operating conditions, the system can adaptively predict and adjust power. The cluster analysis module can identify different production conditions and predict power through a deep learning model to ensure that under each specific condition, the system can accurately calculate the required power value and make dynamic adjustments, thereby improving the intelligence and flexibility of the system. Improve production efficiency and resource utilization: The system avoids excessive or insufficient power output and reduces energy waste through accurate power prediction and dynamic adjustment. By optimizing the power control of copper wire bonding, the system improves production efficiency and resource utilization, and reduces energy consumption and production costs. Ensure process repeatability and stability: The system can not only adjust the power output in real time, but also continuously optimize the power control strategy according to the working condition changes and quality feedback in each production process. By learning from historical data, the system can repeatedly generate consistent power control solutions under the same or similar working conditions, ensuring the repeatability and stability of the copper wire bonding process, and providing reliable guarantee for large-scale production. Simplify power control in production: By integrating a variety of complex algorithms and technologies (such as data preprocessing, cluster analysis, deep learning, and PID control) into a unified system, the present invention simplifies the power control process in the production process. Operators only need to use the real-time feedback and adjustment suggestions provided by the system to accurately grasp the power demand in the copper wire bonding process, reducing the complexity and error probability of manual operation and improving the convenience and safety of operation. Provide comprehensive fault warning and optimization feedback: By real-time monitoring of key parameters in the copper wire bonding process and comparing them with historical data, the system can promptly detect situations where process parameters deviate from the predetermined range and give corresponding adjustment suggestions. Through continuous feedback and optimization of the quality assessment module, the system can warn of potential failures or quality problems during the production process and make timely adjustments and corrections, thereby reducing the defect rate in production and improving the product qualification rate. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] Figure 1 It is the flow chart of data acquisition module; Figure 2 It is the flow chart of data preprocessing module; Figure 3 It is the flow chart of cluster analysis module; Figure 4 It is the flow chart of power prediction module; Figure 5 It is the flow chart of dynamic power adjustment module; Figure 6 It is the flow chart of power calculation module; Figure 7 It is the flow chart of quality assessment and optimization module. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the accompanying drawings in the embodiments of the present invention: Data collection: By installing embedded sensors on the bonding equipment, various data in the process are monitored in real time, and the data is transmitted to the central control unit via wired (RS485) or wireless (Wi-Fi). In order to ensure high reliability and real-time data, the sensor collection cycle is milliseconds to ensure that changes in key working conditions can be accurately captured; step: After the system starts, initialize each sensor to ensure calibration; Temperature collection: The sensor detects the ambient temperature in real time; Pressure acquisition: The pressure sensor is installed on the bonding machine robot arm to record the applied pressure in real time and synchronize the processing with the robot arm movement; Roughness acquisition: The laser displacement sensor scans the surface of the copper wire, obtains the three-dimensional profile data and calculates the roughness value; The collected multidimensional data is stored in a local cache and uploaded to a central processing unit; Data preprocessing: The high-frequency noise in the data is removed by a low-pass filter (the cutoff frequency of the filter is set to 100 Hz to ensure that the noise is removed and the useful signal is retained). The data of various dimensions are unified into the same range to avoid the influence of different dimensions. The normalization formula is used: ; in: is the original data, and are the maximum and minimum values ​​of the data, respectively. The most relevant features are extracted from multiple sensor data through the principal component analysis (PCA) dimensionality reduction technique; The calculation formula for PCA dimensionality reduction is: ; in: is the eigenvector matrix, is the feature data before dimensionality reduction, is the feature data after dimensionality reduction; step: Read real-time data from the data acquisition module; Noise filtering: high-frequency noise in the data is removed by a low-pass filter (the cutoff frequency of the filter is set to 100 Hz to ensure that the noise is removed and the useful signal is retained); Data normalization: Unify data of different dimensions into the same range to avoid the influence of different dimensions; Feature extraction: extract the most relevant features from multiple sensor data through principal component analysis (PCA) dimensionality reduction technology; Cluster analysis: The K-means algorithm is used to cluster the data and form multiple categories based on different feature combinations, which are used as input features in subsequent power prediction. K-means clustering cluster error minimization goal: ; in: is the number of clusters, Cluster , Cluster The cluster center, is a data point; step: receiving processed data from a data preprocessing module; Initially set the K value (such as K=3) and use the K-means algorithm to randomly select the initial cluster center; Calculate the distance from the data point to the center of each cluster and assign the data point to the nearest cluster; Update the cluster center position until the cluster center no longer changes; Output the final clustering results as input for subsequent power prediction; Power prediction: Using the cluster analysis results, deep learning algorithms and minimizing loss functions, the prediction accuracy of the model is optimized to achieve accurate power, predict the optimal power requirements under different working conditions, and minimize the loss function: ; in: To predict power, is the actual power, is the number of data points.

[0020] Dynamic Power Scaling: According to the prediction value provided by the power prediction module, the working power of the equipment is adjusted in real time according to the PID control to ensure that the optimal power demand is maintained during the copper wire bonding process; step: The output of the receiving power prediction module is used to determine the target power value; Adjustments are made using a PID controller, which calculates the new power setting and adjusts the device power output in real time; Continuously monitor actual power and adjust control parameters based on feedback information; Power calculation: Based on real-time temperature, pressure, and surface roughness data, an empirical model of power demand is generated. The coefficients in the empirical model are determined through experiments, and the physical quantities under different working conditions are mapped to the actual power, which serves as the basic input of the dynamic power adjustment module. Power calculation formula: ; in: and is the empirical coefficient, For pressure, is the surface roughness, is the ambient temperature; step: Obtain temperature, pressure, and surface roughness data from sensors in real time; Calculate the actual power according to the power calculation formula; The calculation results are passed to the dynamic power adjustment module for adjusting the power output.

[0021] Quality Assessment and Optimization Module The quality score is calculated by weighted average method based on multiple quality evaluation indicators (such as bonding strength, thermal stability, and surface smoothness), and the system parameters are adjusted according to the evaluation results to further optimize the cluster analysis and power prediction model, thereby improving the stability and production efficiency of the system; step: Monitoring of bonding quality parameters (bond strength, thermal stability, surface smoothness); The quality score was calculated using the weighted average method; The quality score formula is as follows: ; in: is the bonding strength, For thermal stability, For surface smoothness, , , is the weight coefficient; Optimize system parameters based on evaluation feedback, adjust cluster analysis and power prediction models to continuously improve final product quality and process stability; A power control method for an intelligent copper bonding wire based on a cluster analysis model comprises the following steps: Step 1: Data collection: The data collection module collects working condition data related to the copper wire bonding process in real time, including temperature, pressure, and surface roughness environmental factors; these data are obtained through sensors and monitoring equipment and transmitted to the data preprocessing module; Step 2: Data preprocessing: remove noise and normalize the raw data collected from the data acquisition module; convert feature data of different dimensions into a unified scale through standardization methods to ensure that the cluster analysis and power prediction modules can process the data correctly; Step 3: Cluster analysis: Use the K-means clustering algorithm to analyze the data and divide the collected operating data into several clusters according to similarity; each cluster represents a specific working state or operating mode; in this way, the cluster analysis module provides classification features for subsequent power prediction and dynamic adjustment, and provides input based on actual operating conditions for power prediction; Step 4: Power prediction: Based on the working condition cluster characteristics generated by the cluster analysis module, a deep neural network is used to predict power. This module uses the prediction model trained based on historical data and the cluster analysis results as input to predict the power demand during the bonding process through a regression algorithm. The power prediction results provide a reference for the system to adjust the actual power. Step 5: Dynamic power adjustment: The actual power is adjusted through the dynamic power adjustment module; this module uses the PID control algorithm to calculate the error between the target power and the actual power in real time, and adjusts the power output to ensure that the actual power is close to the predicted power, avoiding bonding quality problems caused by power fluctuations; Step 6: Power calculation: Calculate the reference power value based on environmental parameters, pressure, temperature and roughness factors. This module uses empirical formulas and combines process data from actual production to determine the power values ​​required under different working conditions and provide input for the dynamic power adjustment module. Step 7. Quality assessment and optimization: Comprehensively evaluate the copper wire bonding quality based on multiple quality assessment indicators (such as bonding strength, thermal stability, and surface smoothness); calculate the quality score through the weighted average method, and adjust the system parameters based on the evaluation results, and further optimize the cluster analysis and power prediction model to improve the system stability and production efficiency.

Claims

1. A power control system for intelligent copper bonding wire based on a cluster analysis model, characterized in that: include: Data acquisition module, used to obtain multi-dimensional parameters involved in the copper wire bonding process in real time; A data preprocessing module is used to normalize, reduce noise and extract features from the collected data; The clustering analysis module is used to cluster the feature data using the clustering algorithm to generate multiple feature clusters; The power prediction module uses a deep learning algorithm to predict the required optimal power based on clustering results and historical data; Dynamic power adjustment module, used to predict power in real time and adjust power output; Power calculation module, used to collect data and calculate reference power value using empirical formula; The quality assessment and optimization module is used to comprehensively evaluate the bonding quality and optimize the system parameters based on the evaluation results.

2. The system according to claim 1, characterized in that The data acquisition module specifically includes: the main function of the data acquisition module is to obtain the environment and process parameters related to the copper wire bonding process in real time, and these parameters include temperature, pressure, and surface roughness, wherein: Temperature: Use temperature sensors to monitor the real-time temperature of the working environment , this temperature affects the thermal expansion and electrical conductivity of the copper wire, thus affecting the power demand; Pressure: Pressure sensors measure the pressure applied during the bonding process ,Excessive pressure will cause damage to the copper wire, and too little pressure will result in loose bonding, so precise control is required; Surface roughness: Use laser displacement sensor to obtain the surface roughness of bonding copper wire , surface roughness affects the electrical contact effect, thereby affecting power consumption.

3. The system according to claim 1, characterized in that The data preprocessing module specifically includes: the data preprocessing module cleans, normalizes and performs noise filtering on the raw data obtained from the acquisition module to remove irrelevant information and extract valuable features, wherein: Normalization: Since different sensors have different output data dimensions, normalization is part of the standardization process to ensure that each data feature is within the same range, thereby preventing a certain feature from having too much influence on the result; Noise filtering: Use a low-pass filter to remove high-frequency noise in the data. The source of the noise is sensor failure, environmental interference, or unstable operation of electronic equipment. The cutoff frequency of the filter is set to 100Hz to ensure that the noise is removed and the useful signal is retained; Feature extraction: The most relevant features are extracted from multiple sensor data through the principal component analysis (PCA) dimensionality reduction technology. PCA projects the original data into a new orthogonal space, retaining most of the data variance and reducing computational complexity. The calculation formula for PCA dimensionality reduction is: ； in: is the eigenvector matrix, is the feature data before dimensionality reduction, is the feature data after dimensionality reduction.

4. The system according to claim 1, characterized in that The cluster analysis module specifically includes using an unsupervised learning method to classify the feature data obtained from the data acquisition module and the data preprocessing module, and dividing the data into several clusters through a clustering algorithm to identify different operating modes. This module uses a K-means clustering algorithm to group the feature data points according to the similarity between the feature data points. Each cluster represents a specific operating mode, which will be used as an input feature in subsequent power prediction; the K-means clustering cluster error minimization objective is: ； in: is the number of clusters, Cluster , Cluster The cluster center, is the data point.

5. The system according to claim 1, characterized in that The power prediction module specifically includes: the power prediction module performs power prediction through a deep learning algorithm based on the cluster features provided by the cluster analysis module. The module adopts a deep neural network model, and the power prediction model obtained by training according to historical data. The input is the output features of the cluster analysis module, and the output is the power prediction value. By minimizing the loss function, the prediction accuracy of the model is optimized to achieve accurate power demand prediction; Minimize the loss function: ； in: To predict power, is the actual power, is the number of data points.

6. The system according to claim 1, characterized in that The dynamic power adjustment module specifically includes: the dynamic power adjustment module dynamically adjusts the actual power output through the PID control algorithm according to the output of the power prediction module to ensure the precise control and stability of the power during the copper wire bonding process. The module calculates the error between the target power and the actual power, and adjusts the power output through the proportional, integral and differential feedback regulation controller, so that the actual power is consistent with the predicted power, avoiding bonding quality problems caused by unstable power.

7. The system according to claim 1, characterized in that The power calculation module specifically includes: the power calculation module calculates the reference power value according to historical data and empirical formulas; the module generates an empirical model of power demand by combining and analyzing three working condition parameters, namely, temperature, pressure and surface roughness; the coefficients in the empirical model are determined by experiments, and the physical quantities under different working conditions are mapped to the actual power, which is used as the basic input of the dynamic power adjustment module; Reference power calculation formula: ； in: and is the empirical coefficient, For pressure, is the surface roughness, is the ambient temperature.

8. The system according to claim 1, characterized in that The quality assessment and optimization module specifically includes: the quality assessment and optimization module comprehensively evaluates the copper wire bonding quality based on multiple quality assessment indicators, evaluates the bonding quality in combination with process parameters and operating conditions, calculates the quality score using a weighted average method, and optimizes system parameters based on the evaluation result feedback, adjusts the cluster analysis and power prediction models, so as to continuously improve the final product quality and process stability; The quality score formula is as follows: ； in: is the bonding strength, For thermal stability, For surface smoothness, , , is the weight coefficient.

9. A power control method for intelligent bonding copper wire based on a cluster analysis model, characterized in that: The following steps are involved: Step 1: Data collection: The data collection module collects working condition data related to the copper wire bonding process in real time, including three environmental factors: temperature, pressure, and surface roughness. These data are obtained through sensors and monitoring equipment and transmitted to the data preprocessing module. Step 2: Data preprocessing: Remove noise and normalize the raw data collected from the data acquisition module. Through the standardization method, the feature data of different dimensions are converted into a unified scale to ensure that the cluster analysis and power prediction modules can process the data correctly. Step 3: Cluster analysis: Use the K-means clustering algorithm to analyze the data and divide the collected operating data into several clusters according to similarity. Each cluster represents a specific working state or operating mode. In this way, the cluster analysis module provides classification features for subsequent power prediction and dynamic adjustment, and provides input based on actual working conditions for power prediction. Step 4: Power prediction: Based on the working condition cluster characteristics generated by the cluster analysis module, a deep neural network is used to predict power. This module uses the prediction model trained based on historical data and the cluster analysis results as input. The power demand during the bonding process is predicted through a regression algorithm. The power prediction results provide a reference for the system to adjust the actual power. Step 5: Dynamic power adjustment: The actual power is adjusted through the dynamic power adjustment module. The module uses the PID control algorithm to calculate the error between the target power and the actual power in real time, and adjusts the power output to ensure that the actual power is close to the predicted power, avoiding bonding quality problems caused by power fluctuations. Step 6. Power calculation: Calculate the reference power value based on environmental parameters, pressure, temperature and roughness factors. This module uses empirical formulas and combines process data in actual production to determine the power value required under different working conditions and provide input for the dynamic power adjustment module. Step 7: Quality assessment and optimization: Based on multiple quality assessment indicators, the copper wire bonding quality is comprehensively evaluated, the quality score is calculated by the weighted average method, and the system parameters are adjusted according to the evaluation results. The cluster analysis and power prediction models are further optimized to improve the system stability and production efficiency.

Images