System and method for judging health degree of manufacturing equipment

Abstract

The invention discloses a system and a method for judging the health degree of manufacturing equipment. The system comprises a code book database, a historical data database, a modeling server and an edge operation server. The codebook database stores a plurality of monitoring parameters of the manufacturing equipment in each process. The historical data database stores real-time high-frequency data of all the parameters of the manufacturing equipment so as to provide historical high-frequency data of all the parameters of the manufacturing equipment. The modeling server is used for extracting the historical high-frequency data of the plurality of monitoring parameters from the historical database according to the codebook data of the codebook database, and extracting the characteristics of the historical high-frequency data of the plurality of monitoring parameters so as to establish a health degree model of the manufacturing equipment according to the characteristics of the historical high-frequency data of the plurality of monitoring parameters. The edge operation server analyzes real-time high-frequency data and a health degree model which are thrown upwards by the manufacturing equipment in the current station manufacturing process in real time, and then health index scores of the manufacturing equipment in the current station manufacturing process are generated.

Description

Technical Field

[0001] The present invention relates to a system and method for determining the health of manufacturing equipment, and more particularly to a system and method for determining the health of manufacturing equipment by monitoring the current process status in real time. Background Art

[0002] Various variations exist in semiconductor manufacturing processes. These variations can be caused by material differences, deviations in machine parameter settings, disturbances in the manufacturing environment, or inconsistent operating techniques. These variations can collectively lead to post-process product variability. The goal of process control is to control inter-process variation to reduce the degree of variation between products and maintain consistent product quality.

[0003] In existing process control technologies, abnormalities occurring at a given process station often require waiting until the next inspection station to detect defective products. Sometimes, problems with the final product are only detected after the final inspection station has completed the inspection. Therefore, a system and method are needed to monitor the status of the current process station in real time to determine the health of manufacturing equipment. Summary of the Invention

[0004] The object of the present invention is to provide a system and method for determining the health of manufacturing equipment.

[0005] The present invention provides a system for determining the health of manufacturing equipment, which includes a code book database, a historical data database, a modeling server, and an edge computing server. The code book database is used to store code book data, which is a plurality of monitoring parameters of a manufacturing equipment in each process. The historical data database is used to store real-time high-frequency data of all parameters of the manufacturing equipment to provide historical high-frequency data of all parameters of the manufacturing equipment. The modeling server is used to extract the historical high-frequency data of the plurality of monitoring parameters from the historical data database based on the code book data of the code book database, and extract the features of the historical high-frequency data of the plurality of monitoring parameters to establish a health model of the manufacturing equipment accordingly. The edge computing server is used to analyze the real-time high-frequency data and the health model of the manufacturing equipment in a current process in real time, and thereby generate a health index score of the manufacturing equipment in the current process.

[0006] The present invention further provides a method for determining the health of manufacturing equipment, which includes setting a code book data for multiple monitoring parameters of a manufacturing equipment in each process; storing real-time high-frequency data of all parameters of the manufacturing equipment to provide historical high-frequency data of all parameters of the manufacturing equipment; extracting the historical high-frequency data of the multiple monitoring parameters from the historical high-frequency data of all parameters of the manufacturing equipment based on the code book data, and extracting features of the historical high-frequency data of the multiple monitoring parameters to establish a health model of the manufacturing equipment accordingly; and real-time analysis of the real-time high-frequency data and the health model of the manufacturing equipment in a current process, thereby generating a health index score for the manufacturing equipment in the current process. BRIEF DESCRIPTION OF THE DRAWINGS

[0007] Figure 1 This is a functional block diagram of a system for determining the health of manufacturing equipment in an embodiment of the present invention.

[0008] Figure 2 This is a functional block diagram of a system for determining the health of manufacturing equipment in another embodiment of the present invention.

[0009] Figure 3 This is a functional block diagram of a system for determining the health of manufacturing equipment in another embodiment of the present invention.

[0010] Figure 4 The figure is a flow chart of a method for determining the health of manufacturing equipment in an embodiment of the present invention.

[0011] Explanation of reference numerals: 10 - modeling server; 20 - edge computing server; 30 - analysis server; 40 - terminal device; 100, 200, 300 - system; 410-470 - steps; DB1 - codebook database; DB2 - historical data database; DB3 - model database; DB4 - health indicator database; ME1-ME N ~ manufacturing equipment. DETAILED DESCRIPTION

[0012] Figure 1 FIG. 1 is a functional block diagram of a system 100 for determining the health of manufacturing equipment according to an embodiment of the present invention. Figure 2 FIG. 2 is a functional block diagram of a system 200 for determining the health of manufacturing equipment according to another embodiment of the present invention. Figure 3 This is a functional block diagram of a system 300 for determining the health of manufacturing equipment in another embodiment of the present invention. Systems 100, 200, and 300 each include a modeling server 10, an edge computing server 20, an analysis server 30, a codebook database DB1, a historical data database DB2, a model database DB3, and a health indicator database DB4, which can monitor the health of manufacturing equipment ME1-ME NWhere N is a positive integer. The codebook database DB1 is used to store the health of the manufacturing equipment ME1-ME N The parameters that need to be monitored in each process (ie, the code book data below) are stored in the historical data database DB2 for the manufacturing equipment ME1-ME N The real-time high-frequency data of all parameters provide the historical high-frequency data of the corresponding parameters, and the model database DB3 is used to store the manufacturing equipment ME1-ME N The modeling server 10 can extract the historical high-frequency data of the corresponding parameters from the historical data database DB2 based on the code book data of the code book database DB1, and extract the characteristics of the historical high-frequency data to establish a health model for each manufacturing equipment. The edge computing server 20 can analyze the real-time high-frequency data and health model of each manufacturing equipment in real time, and then generate a health index score for each manufacturing equipment. The health index database DB4 is used to store the health index score of each manufacturing equipment. The analysis server 30 is used to determine whether the health index score of each manufacturing equipment is qualified, and analyze the cause of the abnormality when the health index score is determined to be unqualified. Then, the analysis server 30 can provide information on the analysis results to the terminal device 40 (such as a computer, mobile phone, tablet, etc.) so that the user can deal with the abnormal conditions of the manufacturing equipment accordingly.

[0013] exist Figure 1 In the system 100 shown, the modeling server 10, the analysis server 30, the code book database DB1 and the health index database DB4 are set in the cloud, while the edge computing server 20, the historical data database DB2, the model database DB3 and the manufacturing equipment ME1-ME N Set up at the factory end, wherein the factory end may include one or more factories. In one embodiment, the manufacturing equipment ME1-ME N Set up in the same factory, they share the edge computing server 20, historical data database DB2 and model database DB3, and share the modeling and analysis services of the cloud. For example, each factory sets up manufacturing equipment ME1-ME N In addition, an edge computing server 20, a historical data database DB2 and a model database DB3 are also provided. N The manufacturing equipment in each factory is set up in multiple factories, and the manufacturing equipment in each factory uses the edge computing server 20, the historical data database DB2 and the model database DB3, as well as the modeling and analysis services in the cloud. For example, each factory sets up manufacturing equipment ME1-ME N , and one of the factories can also set up an edge computing server 20, a historical data database DB2 and a model database DB3 for common use by other factories.

[0014] exist Figure 2 In the system 200 shown, the modeling server 10, the analysis server 30, the code book database DB1, the historical data database DB2 and the health index database DB4 are set in the cloud, while the edge computing server 20, the model database DB3 and the manufacturing equipment ME1-ME N Set up at the factory end, wherein the factory end may include one or more factories. In one embodiment, the manufacturing equipment ME1-ME N Set up in the same factory, they share the edge computing server 20 and the model database DB3, as well as the cloud's real-time / historical high-frequency data storage, modeling and analysis services. For example, each factory is equipped with manufacturing equipment ME1-ME N In addition, an edge computing server 20 and a model database DB3 are also provided. In another embodiment, the manufacturing equipment ME1-ME N The manufacturing equipment in each factory is set up in multiple factories, and the manufacturing equipment in each factory uses the edge computing server 20 and the model database DB3, as well as the real-time / historical high-frequency data storage, modeling and analysis services in the cloud. For example, each factory sets up manufacturing equipment ME1-ME N , and one of the factories can also set up an edge computing server 20 and a model database DB3 for shared use by other factories.

[0015] exist Figure 3 In the system 300 shown, the modeling server 10, the edge computing server 20, the analysis server 30, the codebook database DB1, the historical data database DB2, the model database DB3 and the health indicator database DB4 are all located in the same factory. The factory independently manages its manufacturing equipment ME1-ME N All real-time / historical high-frequency data storage, modeling and analysis services are completed in the factory, which is conducive to improving data confidentiality and security.

[0016] In the embodiment of the present invention, the modeling server 10, edge computing server 20, and analysis server 30 may include devices such as a motherboard, a central processing unit (CPU), memory, a chipset, a hard drive, a network card, and a power supply, capable of storing, processing, and analyzing data. However, the implementation of each server does not limit the scope of the present invention.

[0017] Figure 4 The figure is a flow chart of a method for determining the health of manufacturing equipment in an embodiment of the present invention. Figure 4 The method flow chart shown comprises the following steps:

[0018] Step 410: Set corresponding codebook data for each manufacturing equipment to determine monitoring parameters and monitoring methods for each process, and store the codebook data in the codebook database DB1.

[0019] Step 420 : Each manufacturing device sends real-time high-frequency data of all parameters to the historical data database DB2 and the edge computing server 20 .

[0020] Step 430 : Based on the codebook data in the codebook database DB1 , the modeling server 10 retrieves the historical high-frequency data of the monitoring parameters of each manufacturing equipment from the historical data database DB2 .

[0021] Step 440 : The modeling server 10 processes and analyzes the captured historical high-frequency data to establish a health model for each manufacturing equipment.

[0022] Step 450: The modeling server 10 uploads the codebook data and health model of each manufacturing equipment to the model database DB3.

[0023] Step 460: The edge computing server 20 compares the health model of each manufacturing device with the real-time high-frequency data thrown by the manufacturing device to generate a health index score for each manufacturing device and stores it in the health index database DB4.

[0024] Step 470: The analysis server 30 retrieves the health index score of each manufacturing device from the health index database DB4, determines whether the health index score of each manufacturing device is qualified, and analyzes the cause of the abnormality if the health index score is determined to be unqualified.

[0025] As is well known to those skilled in the art, manufacturing equipment typically requires multiple different parameter settings for each process, and each process may take a different amount of time. If different parameters and time durations in different processes are monitored separately, the required time dimension may be very large. Furthermore, monitoring the correlation between different parameters is difficult when each parameter is monitored independently, making it difficult for users to quickly identify parameters that affect the health of manufacturing equipment. Therefore, in step 410 of the present invention, users can set the codebook data for each manufacturing equipment based on personal or historical experience to determine the monitoring parameters and monitoring methods for each process, but this is not limited to this.

[0026] In an embodiment of the present invention, the parameters that need to be monitored for each manufacturing equipment during each process may include one or more physical parameters (e.g., pressure, temperature, flow rate, etc.) and / or one or more chemical parameters (e.g., concentration, pH value, etc.). The parameters that need to be monitored may vary for each process. In an embodiment of the present invention, a user may use a terminal device 40 (e.g., a computer, mobile phone, tablet, etc.) to program the codebook data for each manufacturing equipment, but the present invention is not limited thereto.

[0027] Table 1 below shows a schematic diagram of the codebook data set in step 410 according to an embodiment of the present invention. The SVID column displays all parameters of a specific manufacturing equipment in the current process station. The Mon column is used to determine whether a parameter is monitored (e.g., 1 indicates a monitored parameter, 0 indicates a non-monitored parameter). The FtrTp1 column is used to set the number of pre-filter seconds, the FtrTp2 column is used to set the number of post-filter seconds, the LSL column is used to set the lower limit of the monitoring interval, the Target column is used to set the target value of the monitoring interval, and the USL column is used to set the upper limit of the monitoring interval. In the embodiment shown in Table 1, the monitored parameters of the specific manufacturing equipment in the current process station are N2, SIH4, and H2, while the parameters NF3, AR, NH3, PH3, and PRESSU do not need to be monitored. Furthermore, assuming the entire process duration is P seconds, according to Table 1, the monitoring interval for parameter N2 is 3 seconds after the start of the process to 6 seconds before the end (e.g., P - 9 seconds), with the lower limit, target value, and upper limit values ​​of the monitoring interval being 400, 600, and 800, respectively. The monitoring interval for parameter SIH4 is 7 seconds after the start of the process to 5 seconds before the end (e.g., P - 12 seconds), with the lower limit, target value, and upper limit values ​​of the monitoring interval being 2375, 2500, and 2625, respectively. The monitoring interval for parameter H2 is 10 seconds after the start of the process to 10 seconds before the end (e.g., P - 20 seconds), with the lower limit, target value, and upper limit values ​​of the monitoring interval being 6650, 7000, and 7350, respectively. The settings for the monitored parameters, monitoring interval, and the lower limit, target value, and upper limit values ​​of the monitoring interval can be based on user process experience, thereby reducing the probability of false alarms. In some embodiments, the codebook data may also set weight values ​​for monitoring parameters.

[0028] Table 1

[0029] SVID Mon FtrTp1 FtrTp2 LSL Target USL NF3 0 0 0 AR 0 0 0 NH3 0 0 0 N2 1 3 6 400 600 800 SIH4 1 7 5 2375 2500 2625 PH3 0 0 0 H2 1 10 10 6650 7000 7350 PRESSU 0 0 0

[0030] In step 420, each manufacturing device will upload high-frequency data of all parameters of the current process to the historical data database DB2 and the edge computing server 20 in real time during operation. In the present invention, high-frequency data refers to multiple data in time units in a process, such as multiple data measured at intervals of 5 seconds, 2 seconds, 1 second, 500 milliseconds, 100 milliseconds, or 10 milliseconds, but is not limited to these. In one embodiment, each manufacturing device will periodically upload real-time high-frequency data of all parameters to the historical data database DB2 and the edge computing server 20. In another embodiment, each manufacturing device can upload real-time high-frequency data of all parameters to the historical data database DB2 and the edge computing server 20 at a pre-set time point.

[0031] In step 430, the modeling server 10 retrieves historical high-frequency data for each manufacturing equipment monitoring parameter from the historical data database DB2 based on the codebook data. For example, based on the codebook data shown in Table 1, the modeling server 10 retrieves every real-time data entry for parameters N2, SIH4, and H2 during the entire process at the current station. However, according to the codebook data, parameters NF3 / AR / NH3 / PH3 / PRESSU do not need to be monitored, so the modeling server 10 does not retrieve every real-time data entry for NF3 / AR / NH3 / PH3 / PRESSU.

[0032] In step 440, the modeling server 10 processes and analyzes the captured historical high-frequency data to establish a health model for each manufacturing equipment. In more detail, the above-mentioned data processing and analysis process includes stages such as data pre-processing, data feature extraction, and data modeling. In the data pre-processing stage, the modeling server 10 first performs steps such as data alignment, filtering, value addition, function smoothing, and calculation of the data median curve on the captured historical high-frequency data. Different manufacturing equipment may take different times to execute the same process due to slight differences in hardware or operation, so the historical high-frequency data captured by the modeling server 10 may also have different lengths. The present invention can set the data with a median length among all the captured historical high-frequency data as standard historical high-frequency data, and then perform data alignment accordingly to align the head and tail of each historical high-frequency data with the head and tail of the standard historical high-frequency data.

[0033] Assuming the length L1 of the original historical high-frequency data is less than the standard length L0 of the standard historical high-frequency data, the length L1' of the historical high-frequency data after data alignment will be stretched to the standard length L0. This will result in missing data values ​​at one or more locations. In this case, the modeling server 10 can use interpolation to fill in the missing data values. Assuming the length L2 of the original historical high-frequency data is greater than the standard length L0 of the standard historical high-frequency data, the length L2' of the historical high-frequency data after data alignment will be shortened to the standard length L0. This will result in multiple data values ​​at the same location. In this case, the modeling server 10 can average the multiple data values ​​corresponding to the same location.

[0034] Next, the modeling server 10 can perform second filtering and value interpolation based on the corresponding codebook data. As shown in Table 1, for the codebook data example, users can set different monitoring intervals and lower, target, and upper limits for each monitoring parameter. For example, since the monitoring interval for parameter N2 is 3 seconds after the start of the current station process to 6 seconds before the end, the modeling server 10 will filter out the data values ​​for the first 3 seconds and last 6 seconds of the corresponding historical high-frequency data. Since the monitoring interval for parameter SIH4 is 7 seconds after the start of the current station process to 5 seconds before the end, the modeling server 10 will filter out the data values ​​for the first 7 seconds and last 5 seconds of the corresponding historical high-frequency data. Since the monitoring interval for parameter H2 is 10 seconds after the start of the current station process to 10 seconds before the end, the modeling server 10 will filter out the data values ​​for the first and last 10 seconds of the corresponding historical high-frequency data. After filtering, there may be missing values ​​at one or more locations in the historical high-frequency data. In these cases, the modeling server 10 can interpolate and interpolate. Next, the modeling server 10 uses a functional data analysis method to smooth the curve of each pre-processed historical high-frequency data to reduce the random variation of the high-frequency data and retain the curve trend, and then calculates the data median curve of each monitoring parameter.

[0035] The data feature extraction stage includes calculating statistical feature values ​​and relational feature values. The statistical feature values ​​for each data item may include at least one of the median, maximum, minimum, average, or other statistical values. The relational feature values ​​for each historical high-frequency data item may include at least one of the mean directional outlyingness (MO), the variation of directional outlyingness (VO), the distance from the data median curve, or other appropriate values. The mean directional outlyingness and variation of directional outlyingness for each historical high-frequency data item can be calculated by the modeling server 10 for each data item's projection depth, and then calculated based on the projection depth. Finally, the modeling server 10 can establish a health model for each manufacturing equipment item based on the characteristics (statistical features and relational features) of the historical high-frequency data item's monitored parameters. Since statistical features represent the characteristics of each data item itself, and relational features represent the degree of difference between each data item and the data median curve, models established using statistical and relational features help identify relationships between multiple monitoring parameters and improve model accuracy. In some embodiments, the modeling server 40 also simulates the upper and lower high-frequency data for each process, and uses the characteristics of the upper specification limit high-frequency data, the characteristics of the lower specification limit high-frequency data, and the characteristics of the historical high-frequency data of the monitoring parameters to establish a health model for each manufacturing equipment. Specifically, the modeling server 40 can shift the maximum and minimum values ​​of the data median curve to the upper and lower limits of the monitoring interval, respectively, and then use a Gaussian process to simulate the upper and lower specification limit high-frequency data. Next, the modeling server 40 can calculate the statistical characteristics and relationship characteristics of the upper specification limit high-frequency data and the statistical characteristics and relationship characteristics of the lower specification limit high-frequency data.

[0036] In step 450, the modeling server 10 uploads the code book data and health model of each manufacturing device to the model database DB3. In step 460, the edge computing server 20 compares the health model of each manufacturing device with the real-time high-frequency data uploaded by the manufacturing device in real time to generate a health index score for each manufacturing device and stores it in the health index database DB4. In more detail, after each manufacturing device uploads the real-time high-frequency data of all parameters to the edge computing server 20 in step 420, the edge computing server 20 can use principal component analysis (PCA) to group the parameters and summarize the parameters with high correlation to solve the collinearity problem. Grouping the parameters can reduce the impact of random variation. In addition to making it easier for users to observe the main changes in the sample, it can also reduce the dimensionality of the variables. Next, based on the mean and covariance of the data matrix obtained after PCA of the modeled data, the edge computing server 20 calculates the Mahalanobis distance between the real-time high-frequency data of each parameter group in each manufacturing equipment and the corresponding health model, and then establishes a Hotelling's T-square test to calculate the T-square distribution value (T-square). The edge computing server 20 can use a nonlinear transformation (e.g., ) The score is expressed as 0-100 points (where 0 is the lowest score and 100 is the highest score) as the health index score of each manufacturing equipment. In the present invention, the health index score after nonlinear conversion is not limited to the expression of 0-100. In some embodiments, the edge server 20 may also add the weight value of each monitoring parameter to perform principal component analysis, but is not limited to this. The weight value can be obtained according to the weight value set by the user in the coding book data, but is not limited to this. In some embodiments, the weight value can be automatically calculated based on the ratio of the difference between the maximum and minimum values ​​in the historical high-frequency data and the difference between the monitoring upper limit value USL and the lower limit value LSL of the coding book data. For example, in all historical high-frequency data of the monitoring parameter N2, the maximum value is 750 and the minimum value is 430. The monitoring upper limit value in the coding book data is 800 and the lower limit value is 400. Then the weight value of the monitoring parameter N2 is automatically set to (750-430) / (800-400).

[0037] In step 470, the analysis server 30 retrieves the health index score for each manufacturing equipment from the health index database DB4, determines whether the health index score for each manufacturing equipment is qualified, and analyzes the cause of the anomaly if the health index score is determined to be unqualified. For example, the present invention may set the health index score control threshold at 90 points. If the health index score of a specific parameter group in a specific manufacturing equipment is less than 90 points, the analysis server 30 will determine that the health status is unqualified. In one embodiment, when the analysis server 30 determines that the health index score of a specific parameter group in a specific manufacturing equipment is unqualified, it attributes the abnormal parameter to the cause and then provides the analysis results to the terminal device 40, allowing the user to maintain and / or adjust the specific manufacturing equipment accordingly. In another embodiment, the analysis server 30 further concatenates multiple maintenance records to determine the applicability of the health model for each manufacturing equipment. If the health index score of the specific parameter group in a specific manufacturing equipment is still unqualified after maintenance, the analysis server 30 will notify the modeling server 10 to re-model the model.

[0038] In summary, the present invention provides a system and method that can monitor the current process status in real time to determine the health of manufacturing equipment. The user's process experience can be incorporated into the modeling process to set the monitoring parameters, monitoring intervals, and the lower limit values, target values, and upper limit values ​​of the monitoring intervals to reduce the probability of false alarms. In addition, the present invention uses functional data analysis to automatically extract the characteristics of the process data of each manufacturing equipment, and establishes a health model for each manufacturing equipment based on this. On the other hand, the present invention uses statistical multivariate methods combined with machine learning technology to integrate the characteristics of multiple monitoring parameters of each manufacturing equipment into a single health index score to facilitate user management and quickly query parameters that cause process instability.

[0039] The foregoing description is merely an embodiment of the present invention and is not intended to limit the present invention. Persons skilled in the art will readily appreciate that the present invention is susceptible to various modifications and variations. Any modifications, equivalent substitutions, or improvements made within the spirit and principles of the present invention shall be included within the scope of protection of the present invention.

Claims

1. A system for determining the health of manufacturing equipment, characterized in that: Include: a code book database for storing code book data, wherein the code book data is a plurality of monitoring parameters of a manufacturing equipment in each process; a historical data database for storing real-time high-frequency data of all parameters of the manufacturing equipment to provide historical high-frequency data of all parameters of the manufacturing equipment; a modeling server configured to extract historical high-frequency data of the plurality of monitoring parameters from the historical data database based on the codebook data, and extract features of the historical high-frequency data of the plurality of monitoring parameters to thereby establish a health model of the manufacturing equipment; and An edge computing server is used to analyze the real-time high-frequency data of the manufacturing equipment in a current process and the health model in real time, thereby generating a health index score of the manufacturing equipment in the current process.

2. The system according to claim 1, wherein The modeling server is also used to: Calculating a statistical characteristic value and a relational characteristic value of the historical high-frequency data of the plurality of monitoring parameters; as well as The health model of the manufacturing equipment is established according to the statistical characteristic value and the relationship characteristic value.

3. The system according to claim 1, wherein: Also includes: An analysis server is used to determine whether the health index score of the manufacturing equipment is qualified, and to analyze the cause of the abnormality when it is determined that the health index score is unqualified.

4. The system according to claim 1, wherein: Also includes: A model database is used to store the health model of the manufacturing equipment.

5. The system according to claim 4, wherein: The modeling server, the analysis server and the code book database are set up in a cloud, while the edge computing server, the historical data database, the model database and the manufacturing equipment are set up in a factory.

6. A method for determining the health of manufacturing equipment, characterized in that: Include: Setting a code book data for a plurality of monitoring parameters of a manufacturing equipment in each process; Storing real-time high-frequency data of all parameters of the manufacturing equipment to provide historical high-frequency data of all parameters of the manufacturing equipment; extracting the historical high-frequency data of the plurality of monitoring parameters from the historical high-frequency data of all parameters of the manufacturing equipment according to the codebook data, and extracting features of the historical high-frequency data of the plurality of monitoring parameters to thereby establish a health model of the manufacturing equipment; as well as The real-time high-frequency data of the manufacturing equipment during a current process and the health model are analyzed in real time, thereby generating a health index score of the manufacturing equipment during the current process.

7. The method according to claim 6, wherein Also includes: Before extracting features of the historical high-frequency data of the plurality of monitoring parameters, data pre-processing is performed on the historical high-frequency data of the plurality of monitoring parameters.

8. The method according to claim 6, wherein Also includes: Calculating a statistical characteristic value and a relational characteristic value of the historical high-frequency data of the plurality of monitoring parameters; and The health model of the manufacturing equipment is established according to the statistical characteristic value and the relationship characteristic value.

9. The method according to claim 8, wherein: The statistical characteristic value includes at least one of a median, a maximum value, a minimum value, and an average value of each piece of historical high-frequency data of the plurality of monitoring parameters; and The relationship characteristic value includes at least one of an average directional outlier value, a variation directional outlier value, and a distance from a data median curve for each piece of historical high-frequency data of the plurality of monitoring parameters.

10. The method according to claim 6, wherein Also includes: Determine whether the health index score of the manufacturing equipment is qualified, and if it is determined that the health index score is unqualified, analyze the cause of the abnormality and handle it.

Images