Method for evaluating health state of bonding packaging equipment based on twin data
By using a twin-data-based approach to perform hierarchical evaluation of bonding packaging equipment, the problem of accurately assessing equipment health status is solved, enabling more efficient equipment maintenance and operation management.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- THE 724TH RESEARCH INSTITUTE OF CHINA STATE SHIPBUILDING CORP LTD
- Filing Date
- 2026-02-12
- Publication Date
- 2026-07-10
AI Technical Summary
Existing methods for assessing the health status of bonding packaging equipment are difficult to accurately assess equipment failure thresholds due to their high complexity and uncertainty. This leads to over-maintenance or untimely maintenance, which affects production efficiency and costs.
Using a twin-data-based approach, a hierarchical structural model and fault tree are established by classifying the ultrasonic welding head system, motion control system, and machine vision system of the bonding packaging equipment. Combining fuzzy mathematics and entropy weight method, a multi-channel neural network and gray clustering model are used to evaluate health indicators and determine the health status of the equipment.
It improves the reliability and accuracy of equipment health status assessment, shortens maintenance time, reduces production costs, and improves equipment operating efficiency.
Smart Images
Figure CN122365028A_ABST
Abstract
Description
Technical Field
[0001] This invention relates to micro-nano electronic device technology, and more specifically to a method for assessing the health status of bonding packaging equipment based on twin data. Background Technology
[0002] Wire bonding is a special and critical semiconductor packaging process that uses a specialized wedge to guide metal leads through complex, high-speed movements in three-dimensional space to form various special arc shapes required for different packaging forms. It employs the principle of thermo-pressurized ultrasonic welding to create a reliable connection between the bonding wires and the package shell or chip pads. Wire bonding equipment is the carrier of wire bonding technology, representing the concentration of extremely small feature sizes and high throughput in semiconductor production. It is the most critical, technologically dense, and reliability-demanding equipment among the key equipment in integrated circuit post-packaging. The operational stability of wire bonding packaging equipment and its health status assessment have a significant impact on product quality, production costs, and production efficiency. Therefore, developing a health status assessment method for wire bonding packaging equipment is crucial.
[0003] Currently, there are three main types of methods used for equipment health status assessment: model-based health status assessment methods, knowledge-based health status assessment methods, and data-driven health status assessment methods. In practical industrial applications, model-based health status assessment methods have significant limitations in terms of universality. For complex equipment, high-precision physical models need to be built, which is technically challenging and requires substantial investment. Knowledge-based health status assessment methods rely on extensive verification by domain experts to improve the accuracy of equipment assessments, but the collection and representation of expert knowledge present considerable difficulties. Data-driven health status assessment methods are currently the most widely used due to their ease of implementation, high accuracy, and good assessment results.
[0004] However, traditional data-driven health assessments require a large amount of reliable implementation data and various efficient and intelligent algorithms to be truly effective. With technological advancements and societal development, the objects being evaluated are becoming increasingly complex and uncertain, with data types that cannot be directly read or quantitatively analyzed. Different data processed using different methods yield significantly different results, potentially leading to either high equipment failure thresholds causing over-maintenance or low failure thresholds resulting in delayed maintenance and production disruptions. Summary of the Invention
[0005] The purpose of this invention is to provide a health status assessment method for bonding packaging equipment based on twin data, so as to solve the problems of difficult evaluation of faults and difficult processing of health indicators of bonding packaging equipment.
[0006] The technical solution to achieve the purpose of this invention is: a method for assessing the health status of bonding packaging devices based on twin data, comprising the following steps:
[0007] Step 1: Classify the three major functional components of the bonding packaging equipment: the ultrasonic welding head system, the motion control system, and the machine vision system; sort out the common failure modes of each component; and establish a hierarchical structural model and fault tree based on failure mode and effect analysis.
[0008] Step 2: For the fault modes in the fault tree, extract the corresponding monitorable parameters as key health indicators, and construct a three-level health evaluation model that includes the component layer, functional component layer and whole machine layer.
[0009] Step 3: Divide the equipment health status into four qualitative levels: excellent, good, average, and observation. Combine fuzzy mathematics theory to introduce health degree for quantitative characterization. The health degree is calculated through a membership function. The four qualitative levels correspond to the numerical ranges of 0.75–1.0, 0.5–0.75, 0.25–0.5, and 0–0.25, respectively.
[0010] Step 4: Classify and process key health indicators, including: for vibration signals, use a multi-channel neural network model based on attention mechanism to identify health status; for non-vibration health indicators that are difficult to directly quantify, use a gray clustering model to calculate the clustering coefficients of each health status based on the whitening weight function to determine the health status.
[0011] Step 5: Determine the weight of each health indicator using the entropy weight method driven by twin data, including: standardizing the original health indicator data, calculating the proportion of each indicator, calculating the information entropy of each indicator, and finally determining the weight; then perform gray-class weighted fusion with the various health states obtained in Step 4 to output the real-time health status and health status level of the device.
[0012] Furthermore, in step 1, the common failure modes of the three main functional components—the ultrasonic welding head system, the motion control system, and the machine vision system—are as follows:
[0013] Faults in the ultrasonic welding head system include: transducer faults, cutting tool faults, wire clamp control faults, identification camera serial port faults, wire feeding sensor faults, fastener faults, anti-collision faults, welding faults, and encoder faults.
[0014] Faults in the motion control system include: X / Y / Z axis motor failure, pressure sensor failure, encoder failure, timing failure, communication failure, COM port failure caused by controller drive failure, and keyway wear of the steering motor shaft.
[0015] Faults in machine vision systems include: camera aging, light source failure, vision failure, communication failure, display failure, and interface failure.
[0016] Furthermore, in step 2, the key health indicators for each of the three major functional components—the ultrasonic welding head system, the motion control system, and the machine vision system—are as follows:
[0017] The health indicators of a motion control system include: vibration signal, motor status, controller signal, shock absorber status, encoder status, frame status, and conductive ring status.
[0018] The health indicators of a machine vision system include: image processing board position signal, camera clock output signal, camera current, voltage status, image memory card status, image recording status, and data interface status.
[0019] The health indicators of the welding head system include: drive power excitation signal, piezoelectric motor status, fastener status, amplitude transformer status, PLC status, wire feeder status, and consumable status.
[0020] Furthermore, in step 2, the three-level health evaluation model adopts a hierarchical approach from the local to the overall and from the bottom up. First, the health status of key components is evaluated, then the health status of functional components is obtained from the health status of key components, and finally the overall health status of the bonding packaging equipment is obtained.
[0021] Furthermore, in step 4, the multi-channel neural network model based on the attention mechanism is the MC-IWDCNN-LSTM model, which focuses on extracting time-domain features through one channel and extracting frequency-domain features through another channel. It also performs feature fusion and health status classification by assigning weights to the features of different channels through the attention layer, and outputs a four-dimensional vector to represent the membership degree of the vibration signal to the four health states.
[0022] Furthermore, in step 4, the specific implementation of the gray clustering model includes: setting gray classes according to the number of health status levels, constructing a triangular whitening weight function for each gray class, calculating the membership degree of each health indicator relative to each gray class, i.e., the clustering coefficient, and classifying the health indicator as belonging to the health status corresponding to the gray class with the largest clustering coefficient.
[0023] Furthermore, in step 5, the specific process of determining the weights using the entropy weight method is as follows:
[0024] Step 5-1: Standardization Processing
[0025] Let there be m indicators X1, X2, ..., X m Each indicator has n observations X i ={x i1 ,xi2 ,…,x in The raw data of each health indicator are standardized.
[0026] Step 5-2: Determine the weight of each indicator
[0027] Calculate the weight of each standardized indicator parameter in all indicators, as shown in Equation 1;
[0028] (1);
[0029] in The standardized values of the original data.
[0030] in The proportion of standardized indicator parameters.
[0031] Step 5-3: Determine the information entropy of each indicator
[0032] The information entropy of each indicator is calculated, as shown in Equation 2.
[0033] (2);
[0034] Step 5-4: Determine the weight of each indicator
[0035] Calculate the weight of each indicator As shown in Equation 3;
[0036] (3).
[0037] A health status assessment system for bonding packaging equipment based on twin data includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned health status assessment method for bonding packaging equipment based on twin data.
[0038] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the aforementioned twin-data-based bonding packaging device health status assessment method.
[0039] A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the aforementioned twin-data-based bonding packaging device health status assessment method.
[0040] Compared with the prior art, the significant advantages of this invention are: it divides the packaging equipment into hierarchical models, uses different methods to effectively evaluate the health status of component health indicators, determines the weight of each component health indicator through entropy weight method and completes the health status assessment of the equipment by weighted synthesis, which effectively improves the reliability and accuracy of equipment health status assessment and equipment operating efficiency, shortens maintenance time and reduces production costs. Attached Figure Description
[0041] Figure 1 This is a diagram illustrating the health status assessment scheme for the bonding packaging equipment of the present invention.
[0042] Figure 2 This is a health-related data graph of the bonding packaging equipment used in this invention.
[0043] Figure 3 This is a diagram of the health status evaluation model for the bonding packaging equipment of the present invention. Detailed Implementation
[0044] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.
[0045] To effectively ensure the reliable and stable operation of bonding and packaging equipment and avoid situations where potential faults are not identified in a timely manner, this invention proposes a health status assessment method for bonding and packaging equipment based on twin data, such as... Figure 1 As shown, perform a health assessment of the equipment following these steps:
[0046] Step 1: Classify the main functional components of the bonding packaging equipment, summarize common equipment faults, and establish the structure and fault tree of each functional component of the bonding packaging equipment;
[0047] Figure 2 This is a health-related data chart of the bonding and packaging equipment of the present invention. The main functional components of the bonding and packaging equipment include: an ultrasonic welding head system, a motion control system, and a machine vision system. Potential faults that may occur during equipment operation are summarized below:
[0048] (1) Ultrasonic welding head system. The ultrasonic welding head system is the "heart" of the bonding packaging equipment. It is responsible for automatic wire feeding, image recognition, ultrasonic welding, etc. It is the core component of the equipment and also one of the systems most prone to failure. Common failures mainly include: transducer failure, cleaving blade failure, wire clamp control failure, recognition camera serial port failure, wire feeding sensor failure, fastener failure, anti-collision failure, welding failure, encoder failure, etc.
[0049] (2) Motion Control System. The motion control system is the "brain" and "limbs" of the bonding packaging equipment. It executes various electrical control actions, precisely controls the timing of wire bonding actions, and ensures the accuracy of equipment execution. Common faults mainly include: X / Y / Z axis motor failure, pressure sensor failure, encoder failure, timing failure, communication failure, COM port failure caused by controller drive failure, and keyway wear of steering motor shaft, etc.
[0050] (3) Machine Vision System. The machine vision system is the "eye" of the bonding packaging equipment. It is responsible for equipment system analysis, visual recognition, image processing, decision-making, feedback, and storage. Common faults in this system mainly include: camera aging, light source failure, vision failure, communication failure, display failure, and interface failure.
[0051] Step 2: Determine the key health indicators of the bonding packaging equipment and construct a device-level health evaluation model;
[0052] Figure 3 This is a diagram illustrating the health status evaluation model for the bonding packaging equipment of this invention. Key health indicators for the bonding packaging equipment are determined, and a hierarchical health evaluation model is constructed. The functional modules of the bonding packaging equipment are divided, and the fault characteristics and indicator parameters of each functional component are analyzed to rationally select various health indicators.
[0053] (1) Motion control system. Health indicators include vibration signal, motor status, controller signal, shock absorber status, encoder status, frame status, and conductive ring status;
[0054] (2) Machine vision system. Health indicators include image processing board position signal, camera clock output signal, camera current, voltage status, image memory card status, image recording status, and data interface status;
[0055] (3) Ultrasonic welding head system. Health indicators include drive power excitation signal, piezoelectric motor status, fastener status, amplitude transformer status, PLC status, wire feeder status, and consumable status.
[0056] The health status of Category C indicators in the health status evaluation model for bonding packaging equipment needs to be determined according to one or more specific health indicator parameters and an evaluation method. Category D indicators can be determined directly through software or communication status. For example, the health status of C1 needs to be determined based on twin data parameters such as motor temperature, current, and vibration signals collected in real time by sensors, through further processing and quantitative analysis. The health status of D1 can be determined directly based on the controller feedback status in the twin data signal. The health status of D2 can be determined based on the communication quality of the corresponding sensor in the twin data signal.
[0057] A hierarchical approach, from the local to the overall and from the bottom up, is adopted. First, the health status of key components is assessed. Then, the health status of functional components is obtained from the health status of key components. Finally, the overall health status of the bonding packaging equipment is obtained, thus establishing a hierarchical health status evaluation model for the bonding packaging equipment.
[0058] Step 3: Determine the health status of the bonding packaging equipment and define the equipment health level;
[0059] After establishing the health status evaluation model for bonding packaging equipment, the health status of the bonding packaging equipment is classified and the equipment health degree is defined. The equipment health degree is expressed in both qualitative and quantitative ways.
[0060] Quantitatively, equipment status is categorized into four levels: "Excellent," "Good," "Average," and "Observant." Qualitatively, a health metric (HD) is introduced using fuzzy mathematics theory. This represents the membership degree of a certain health status level. A relationship between equipment health score and membership degree function is established, and finally, the correspondence between health status level and health degree is realized based on the calculation results.
[0061] Table 1 Health Status Level and Health Score
[0062]
[0063]
[0064] Table 2 Correspondence between Health Status Level and Health Score
[0065]
[0066] Step 4: Classify the health indicators of the equipment's functional components and establish corresponding health status evaluation models. For vibration signals that are easily affected by noise in actual working conditions, a deep learning model is used to classify their health status. For health status signal types that cannot be directly extracted, a gray clustering model is used to classify their health status.
[0067] (1) For vibration signal health indicators such as motors, amplitude rods, frames, and fasteners, the collected operating data may deviate significantly from the actual situation due to noise interference in actual working conditions. Therefore, a deep learning model with strong noise resistance, namely the multi-channel processing neural network model MC-IWDCNN-LSTM based on the attention mechanism, is used to determine their health status. The model takes a noisy vibration signal under actual working conditions as input and uses two channels for feature extraction. Each channel learns different features. One channel focuses on the extraction of time domain features, and the other channel focuses on the extraction of frequency domain features. Finally, the attention mechanism is used to adaptively allocate weights to the two processing channels for feature fusion and health status classification. After the model is trained, a real-time vibration signal is input, and a four-dimensional vector can be output to represent the membership degree of the vibration signal to the four health statuses of "excellent", "good", "average" and "observation", which is the clustering coefficient.
[0068] (2) For non-vibration-related health indicators that are difficult to quantify directly, such as signal current, voltage, ultrasound, and pressure, a grey clustering model is used to determine their health status. Four grey classes are defined based on the four health status levels identified in step 3, and a triangular whitening weight function is constructed for each grey class. The standardized values of the health indicators are denoted as... Then, the clustering coefficients of each gray class are calculated using the following whitening weight function:
[0069] (1)
[0070] (2)
[0071] (3)
[0072] (4)
[0073] This yields the clustering coefficient vector of the health indicator, and the gray class corresponding to the maximum value is taken as the health status judgment result of the indicator.
[0074] Step 5: Use the entropy weight method driven by twin data to determine the weight of health indicators, and use gray class weighted comprehensive evaluation to evaluate the real-time health status and health degree of the bonding packaging device.
[0075] To objectively reflect the importance of different health indicators in the comprehensive assessment, the entropy weight method based on twin data is used to dynamically calculate the weight of each indicator. The specific calculation process is as follows:
[0076] (1) Standardization process
[0077] Suppose there are m indicators x1, x2, ..., xn m , where x i={x i1 ,x i2 ,…,x in Based on the characteristics of the indicators, a data standardization method is selected to perform data standardization processing on the raw data of each health indicator.
[0078] Among these, a larger value represents a better performance indicator, and its standardization method is as follows:
[0079] (5)
[0080] A smaller value indicates better performance; its standardization method is as follows:
[0081] (6)
[0082] The closer an indicator is to the standard value, the better its performance. The standardization method is as follows:
[0083] (7)
[0084] (2) Determine the weight of each indicator
[0085] Calculate the weight of each standardized indicator parameter in all indicators;
[0086] (8)
[0087] (3) Determine the information entropy of each indicator
[0088] Calculate the information entropy of each indicator;
[0089] (9)
[0090] (4) Determine the weight of each indicator
[0091] Calculate the weight of each indicator ;
[0092] (10)
[0093] The weights calculated in this step will be used for subsequent weighted fusion.
[0094] Step 6: Collect real-time health indicator twin data from both the digital twin virtual entity and the physical entity; standardize the parameters of each health indicator according to different categories; evaluate the real-time health status of the current part using different methods based on the level to be evaluated; combine the weights determined in Step 5, and repeat the real-time health status evaluation for each level of components to finally obtain the overall health status and health degree of the bonding packaging equipment system. The specific steps are as follows:
[0095] (61) Data collection and standardization: Use the twin data collected in real time as the parameter values for each health indicator. Different formulas are used to standardize the data for different types of health indicator parameter values to obtain dimensionless data between 0 and 1.
[0096] (62) Calculation of health indicator state vector:
[0097] For vibration-related indicators: input the standardized vibration signal data into the trained MC-IWDCNN-LSTM model to directly obtain its health status membership vector;
[0098] For non-vibration indicators: substitute the standardized values of the health indicators into the gray clustering whitening weight functions (1) to (4) in step 4 to calculate their clustering coefficient vector;
[0099] Alarm judgment: If the membership value corresponding to the observed state is the largest in the state vector of any health indicator, an alarm will be triggered immediately and the subsequent comprehensive evaluation process will be suspended.
[0100] (63) Dynamic weight calculation: Based on the standardized data matrix within the current collection window, dynamically calculate the weight vector of each health indicator according to formulas (8) to (10) in step 5;
[0101] (64) Layer-by-layer gray category weighted comprehensive evaluation:
[0102] Component-level evaluation: For a component consisting of k health indicators, the real-time health status of each health indicator is used. For vibration indicators, the health status membership vector is used, and for other indicators, the clustering coefficient vector is used. The weights of each indicator are determined by the entropy weight method. The real-time health status of the current component is evaluated by gray class weighted comprehensive evaluation to obtain its health status vector.
[0103] Functional component layer evaluation: Based on the health status vector of the next layer (such as components), the weights of each component are determined by the entropy weight method, and the real-time health status of the current functional component is obtained through gray class weighted comprehensive evaluation, thus obtaining its health status vector.
[0104] Overall system evaluation: Using the health status vectors of each functional component and the weights of each functional component determined by the entropy weight method, the real-time health status of the current equipment system is obtained through gray class weighted comprehensive evaluation, and its health status vector is obtained.
[0105] (65) Calculate the equipment health status based on the overall health status vector of the whole machine, and output the final health status level of "excellent", "good", "average" or "observation" according to the range in which the HD value falls.
[0106] In summary, this invention innovatively establishes a hierarchical health status evaluation model for bonding packaging equipment and a method for evaluating and processing health indicators. It proposes two different health indicator processing methods—deep learning and grey clustering—to assess the health status of equipment components and functional parts. Finally, it uses the entropy weight method to determine the weights of the health indicators for each component and completes the reliability prediction of the health status of the bonding packaging equipment through weighted comprehensive evaluation. This improves equipment operating efficiency, shortens maintenance time, and reduces production costs.
[0107] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.
[0108] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.
Claims
1. A method for assessing the health status of bonding packaging devices based on twin data, characterized in that, Includes the following steps: Step 1: Classify the three major functional components of the bonding packaging equipment: the ultrasonic welding head system, the motion control system, and the machine vision system; sort out the common failure modes of each component; and establish a hierarchical structural model and fault tree based on failure mode and effect analysis. Step 2: For the fault modes in the fault tree, extract the corresponding monitorable parameters as key health indicators, and construct a three-level health evaluation model that includes the component layer, functional component layer and whole machine layer. Step 3: Divide the equipment health status into four qualitative levels: excellent, good, average, and observation. Combine fuzzy mathematics theory to introduce health degree for quantitative characterization. The health degree is calculated through a membership function. The four qualitative levels correspond to the numerical ranges of 0.75–1.0, 0.5–0.75, 0.25–0.5, and 0–0.25, respectively. Step 4: Classify and process key health indicators, including: for vibration signals, use a multi-channel neural network model based on attention mechanism to identify health status; for non-vibration health indicators that are difficult to directly quantify, use a gray clustering model to calculate the clustering coefficients of each health status based on the whitening weight function to determine the health status. Step 5: Determine the weight of each health indicator using the entropy weight method driven by twin data, including: standardizing the original health indicator data, calculating the proportion of each indicator, calculating the information entropy of each indicator, and finally determining the weight; then perform gray-class weighted fusion with the various health states obtained in Step 4 to output the real-time health status and health status level of the device.
2. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 1, the common failure modes of the three main functional components—the ultrasonic welding head system, the motion control system, and the machine vision system—are as follows: Faults in the ultrasonic welding head system include: transducer faults, cutting tool faults, wire clamp control faults, identification camera serial port faults, wire feeding sensor faults, fastener faults, anti-collision faults, welding faults, and encoder faults. Faults in the motion control system include: X / Y / Z axis motor failure, pressure sensor failure, encoder failure, timing failure, communication failure, COM port failure caused by controller drive failure, and keyway wear of the steering motor shaft. Faults in machine vision systems include: camera aging, light source failure, vision failure, communication failure, display failure, and interface failure.
3. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 2, the key health indicators for each of the three major functional components—the ultrasonic welding head system, the motion control system, and the machine vision system—are as follows: The health indicators of a motion control system include: vibration signal, motor status, controller signal, shock absorber status, encoder status, frame status, and conductive ring status. The health indicators of a machine vision system include: image processing board position signal, camera clock output signal, camera current, voltage status, image memory card status, image recording status, and data interface status. The health indicators of the welding head system include: drive power excitation signal, piezoelectric motor status, fastener status, amplitude transformer status, PLC status, wire feeder status, and consumable status.
4. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 2, the three-level health assessment model adopts a hierarchical approach from the local to the overall and from the bottom up. First, the health status of key components is assessed, then the health status of functional components is obtained from the health status of key components, and finally the overall health status of the bonding packaging equipment is obtained.
5. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 4, the attention-based multi-channel neural network model is the MC-IWDCNN-LSTM model, which focuses on extracting time-domain features through one channel and extracting frequency-domain features through another channel. It also uses an attention layer to assign weights to the features of different channels for feature fusion and health status classification, and outputs a four-dimensional health status vector to represent the membership degree of the vibration signal to the four health states.
6. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 4, the specific implementation of the gray clustering model includes: setting gray classes according to the number of health status levels, constructing a triangular whitening weight function for each gray class, calculating the membership degree of each health indicator relative to each gray class, i.e., the clustering coefficient, and classifying the health indicators as belonging to the health status corresponding to the gray class with the largest clustering coefficient.
7. The method for assessing the health status of bonding packaging devices based on twin data according to claim 1, characterized in that, In step 5, the specific process of determining the weights using the entropy weight method is as follows: Step 5-1: Standardization Processing Let there be m indicators X1, X2, ..., X m Each indicator has n observations X i ={x i1 ,x i2 ,…,x in The raw data of each health indicator are standardized. Step 5-2: Determine the weight of each indicator Calculate the weight of each standardized indicator parameter in all indicators, as shown in Equation 1; （1）； in The standardized values of the original data. in The proportion of standardized indicator parameters. Step 5-3: Determine the information entropy of each indicator The information entropy of each indicator is calculated, as shown in Equation 2. （2）； Step 5-4: Determine the weight of each indicator Calculate the weight of each indicator As shown in Equation 3; （3）。 8. A health status assessment system for bonding packaging equipment based on twin data, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the health status assessment method for bonding packaging equipment based on twin data as described in any one of claims 1-7.
9. A computer device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein when the processor executes the computer program, it implements the twin-data-based bonding packaging device health status assessment method according to any one of claims 1-7.
10. A computer-readable storage medium having a computer program stored thereon, wherein when executed by a processor, the computer program implements the twin-data-based bonding packaging device health status assessment method according to any one of claims 1-7.