Chip high-density packaging process management method and system based on network slicing

Abstract

The invention discloses a chip high-density packaging process management method and system based on network slicing, and belongs to the technical field of network slicing technology and chip integrated control. According to the invention, classification is carried out according to time delay, precision and transmission frequency, and according to a classification result, the state parameters, transmission parameters and collection data of the corresponding network slices, uploaded and packaged equipment are transmitted to the process control server and then transmitted to the process management server, so that the problem of unstable communication is avoided, and the working efficiency is improved. The process management efficiency of the packaging equipment is further improved; through the network slicing technology, uploading state parameters, transmission parameters, collection data and issuing of collection requests are transmitted according to the set network slices, multi-channel communication is achieved, and the process management efficiency of the packaging equipment is further improved.

Description

Technical Field

[0001] The present invention relates to the fields of network slicing technology and chip integration control technology, and in particular to a chip high-density packaging process management method and system based on network slicing. Background Art

[0002] With the continuous development of the Internet of Things, it is increasingly used in chip integration and packaging process management.

[0003] In the process management of high-density chip packaging, the existing Internet of Things technology often adopts the networking of the Internet of Things for the convenience of management, and realizes the process management of high-density chip packaging based on the Internet of Things.

[0004] However, it is well known that high-density chip packaging has high requirements for precision, that is, for the Internet of Things, the requirements for the accuracy, latency and transmission frequency of its data transmission are high. In the operation and management of packaging equipment, the management requests and uploaded data generated by different processes have different requirements for accuracy, latency and transmission frequency. Especially when errors occur in the process, some high-precision and high-latency data often need to be transmitted first. However, due to the structure of the Internet of Things, the above data can only be transmitted according to the inherent single network structure, which leads to poor communication stability and reduces the process management efficiency and operation and maintenance network performance of the packaging equipment. Summary of the invention

[0005] In order to solve the problems of the prior art, an embodiment of the present invention provides a method and system for managing a high-density chip packaging process based on network slicing. The technical solution is as follows: On the one hand, a method for managing a chip high-density packaging process based on network slicing is provided, the method comprising: The process control server configures a data classification model and a request classification model; The packaging devices respectively obtain their own state parameters and transmission parameters; The packaging device collects data collected by the monitoring device; The process control server classifies the state parameter, the transmission parameter and the collected data according to the data classification model, and uploads the state parameter, the transmission parameter and the collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Also includes: The process management server sends a collection request; The process control server classifies the acquisition request according to a request classification model, wherein the categories include latency, accuracy, and transmission frequency; The process control server allocates the corresponding network slice according to the second classification result; The process control server uploads the status parameters, the transmission parameters and the collected data according to the allocated network slice.

[0006] Optionally, the process control server configuration data classification model includes: Setting the data classification model through a probabilistic neural network; The number of the data classification models is the same as the number of the state parameters, the transmission parameters and the collected data.

[0007] Optionally, the process control server configuration request classification model includes: Based on BP neural network, set the request classification model; Setting a correction database and a correction cycle corresponding to the request classification model; After the correction cycle, the request classification model is corrected through the correction database.

[0008] Optionally, the process control server classifies the state parameter, the transmission parameter and the collected data according to a data classification model, including: identifying, according to the data classification model, abnormal events indicated by the state parameter, the transmission parameter, and the collected data; The abnormal event is classified, and the first classification result includes at least one of latency, accuracy, and transmission frequency.

[0009] Optionally, uploading the state parameter, the transmission parameter, and the collected data to a corresponding process management server through corresponding multiple network slices includes: According to the first classification result, slice identifiers corresponding to the multiple network slices are respectively set for the state parameter, the transmission parameter, and the collected data; According to the slice identifier, the state parameter, the transmission parameter and the collected data are uploaded.

[0010] Optionally, the process control server classifies the collection request according to the request classification model, including: Identify the basic request corresponding to the collection request according to the request classification model; Obtaining keyword correction parameters in the correction database; Correcting the acquisition request by using keyword correction parameters; Identify the basic request corresponding to the corrected collection request; The basic request is classified.

[0011] Optionally, the classifying the basic request includes: Preset request classification model; The basic requests are classified according to the request classification model.

[0012] Optionally, the process control server allocates a corresponding network slice according to the second classification result, including: Identifying the state parameters, transmission parameters and collected data indicated by the collection request; Sending to the packaging device a collection sub-request corresponding to the state parameter, transmission parameter and collection data indicated by the collection request; Set the slice ID corresponding to the feedback data of the acquisition subrequest.

[0013] On the other hand, a chip high-density packaging process management system based on network slicing is provided, the system comprising a packaging device, a process control server and a process management server, the packaging device, the process control server and the process management server are connected through a 5G network, and the 5G network configures a plurality of network slices corresponding to the process management server through network slicing technology; wherein: The process control server is used to configure a data classification model and request a classification model; The encapsulation device is used to obtain its own state parameters and transmission parameters respectively; The packaging device is used to collect data collected by the monitoring device; The process control server is used to classify the state parameter, the transmission parameter and the collected data according to the data classification model, and upload the state parameter, the transmission parameter and the collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Also includes: The process management server is used to issue a collection request; The process control server is used to classify the acquisition request according to the request classification model, and the categories include delay, accuracy and transmission frequency; The process control server is used to allocate the corresponding network slice according to the second classification result; The process control server is used to upload the status parameters, the transmission parameters and the collected data according to the allocated network slice.

[0014] The technical solution provided by the embodiment of the present invention has the following beneficial effects: 1. Classify according to latency, accuracy and transmission frequency, and upload the status parameters, transmission parameters and collected data of the packaging equipment to the corresponding network slice based on the classification results, transmit them to the process control server, and then transmit them to the process management server, avoiding the problem of unstable communication and further improving the process management efficiency of the packaging equipment.

[0015] 2. Through network slicing technology, upload status parameters, transmission parameters and collected data, as well as sending collection requests are transmitted according to the set network slices to achieve multi-channel communication and further improve the process management efficiency of packaging equipment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0017] Figure 1 It is a flow chart of a chip high-density packaging process management method based on network slicing provided by an embodiment of the present invention; Figure 2 It is a flow chart of a chip high-density packaging process management method based on network slicing provided by an embodiment of the present invention; Figure 3 It is a schematic diagram of a chip high-density packaging process management system based on network slicing provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0018] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution in the embodiment of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiment of the present invention. Obviously, the described embodiment is only a part of the embodiment of the present invention, not all the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without making creative work are within the scope of protection of the present invention.

[0019] The method described in the embodiment of the present invention is mainly applied to the management of the high-density chip packaging process, and is particularly applied to the server used by different process managers in the high-density chip packaging process management, that is, the process management server. In the high-density chip packaging process management, different process managers or management devices need to be configured with a high-precision, low-latency and high-transmission frequency network to achieve high-precision management of the high-density chip packaging process. The method described in the implementation of the present invention aims to configure different network slices for different process managers or management devices, so as to meet the different needs of different process managers or management devices through multiple logically independent network slices; in actual applications, process managers or management devices need to configure the quality of service (QoS), bandwidth requirements and delay requirements of the network slices they use to meet their needs. The embodiment of the present invention does not limit the specific configuration process.

[0020] In the embodiment of the present invention, it is necessary to configure network slices corresponding to the network slice delay, accuracy and transmission frequency respectively as the first-level network slices; In actual applications, due to the different parameters and standards required in chip packaging, in order to facilitate management, different network slices can be set up according to status parameters, transmission parameters and collection data such as external dimensions, pin layout, number of pins, packaging materials, thermal management, pin arrangement density and welding method data, as secondary network slices, and dedicated management equipment can be configured to manage the above network slices and their corresponding parameters.

[0021] In addition, it should be noted that there are multiple packaging devices described in the embodiment of the present invention, and the multiple packaging devices are used to execute different processes. It can also be that multiple processes are integrated in the same packaging device, and the same packaging device sets network slices according to different processes. As a three-level network slice, the embodiment of the present invention describes the situation where there are multiple packaging devices. For multiple processes integrated in the same packaging device, the method provided by the embodiment of the present invention is also applicable.

[0022] Reference Figure 1 As shown, a chip high-density packaging process management method based on network slicing is provided, including: 101. The process control server configures a data classification model and a request classification model; 102. The packaging devices respectively obtain their own state parameters and transmission parameters; 103. Data collected by the packaging equipment and monitoring equipment; It should be noted that, in actual applications, the state parameters of the packaging equipment mainly refer to the operating parameters of the packaging equipment, the processes performed by the packaging equipment, and the target parameters required by the processes; among them, the target parameters include target external dimensions, target pin layout, target number of pins, target packaging materials, target thermal management parameters, target pin arrangement density and target welding method data.

[0023] Correspondingly, the transmission parameters of the packaging equipment are the actual external dimensions, actual pin layout, actual number of pins, actual packaging materials, actual thermal management parameters, actual pin arrangement density and actual welding method data obtained through detection after the current process is completed; The data collected by the monitoring equipment mainly includes the operating parameters of each component of the packaging equipment obtained through monitoring, as well as the environmental parameters of the environment where the packaging equipment is located, including temperature and electrostatic parameters (obtained through equipment such as electrometers) 104. The process control server classifies the state parameters, transmission parameters and collected data according to the data classification model, and uploads the state parameters, transmission parameters and collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Reference Figure 2 As shown, it also includes: 201. The process management server sends a collection request; 202. The process control server classifies the collection request according to the request classification model, wherein the categories include latency, accuracy, and transmission frequency; 203. The process control server allocates a corresponding network slice according to the second classification result; 204. The process control server uploads status parameters, transmission parameters and collected data according to the allocated network slice.

[0024] The process control server configuration data classification model described in step 101 includes: For each network slice, that is, for each process, a different data classification model is set, that is, the number of data classification models is the same as the number of network slices. For a single data classification model, the configuration process can be: Set up the initial data classification model through probabilistic neural network; Acquire a data set, the data set including actual state parameters, actual transmission parameters and actual collected data; the actual state parameters, actual transmission parameters and actual collected data are actually obtained in the chip high-density packaging process management, and cover abnormal events (such as failures, unqualified inspections, etc.) involved in the chip high-density packaging process management; the embodiment of the present invention does not limit the collection time of the data set, such as the collection time can be every month or every week; Taking electrostatic parameters as an example, perform the following steps: Obtain the electrostatic parameters corresponding to the abnormal events from the data set respectively and set them as training samples; if the number of abnormal events is n, the number of electrostatic parameters in the training samples is n, and the training samples are normalized; set the input number of the initial data classification model to n dimensions; The initial data classification model is trained by the training sample to obtain a final data classification model; wherein the classification results are abnormal and normal; Continue to perform the above steps for other actual state parameters, actual transmission parameters and actual collected data until a primary data classification model corresponding to all actual state parameters, actual transmission parameters and actual collected data is obtained; It should be noted that the setting process of n above is: If the acquisition frequency per second is 4, n is set to 4, that is, one of the actual state parameters, actual transmission parameters and actual acquisition data corresponding to 4 abnormal events is selected from the data set, such as 4 electrostatic parameters; the specific number of n is not limited in the embodiment of the present invention; In the process of uploading actual status parameters, actual transmission parameters and actual collected data to the corresponding process management server through the corresponding multiple network slices, the upload frequency is also n per second.

[0025] Optionally, the process control server configuration request classification model includes: Based on BP neural network, set the request classification model; The request classification model mainly includes two sub-modules. The first one is the request recognition model based on BP neural network; the other one is the request classification model based on probabilistic neural network; The construction process of the request recognition model can be: The request described in the embodiment of the present invention can be implemented by fuzzy text input; For texts containing different keywords, multiple basic requests corresponding to the different keywords are set. The multiple basic requests are manually set, such as a size fault query request, a thermal management fault query request, and an electrostatic anomaly query request, etc. The multiple basic requests at least cover abnormal events; Setting a request recognition training data set, where the request recognition training data set includes multiple texts containing different keywords, and the expected output is multiple basic requests; The BP neural network is trained through the request recognition training data set and expected output to obtain the request recognition model.

[0026] Setting a correction database and correction cycle corresponding to the requested classification model; After the correction cycle, the request classification model is corrected through the correction database.

[0027] In practical applications, the correction database is updated in real time. If the correction database is updated based on historical data, the correction cycle is the update cycle of the correction database. Continue to perform the above steps for other actual state parameters, actual transmission parameters and actual collected data until a primary data classification model corresponding to all actual state parameters, actual transmission parameters and actual collected data is obtained; It should be noted that the setting process of n above is: If the acquisition frequency per second is 4, n is set to 4, that is, one of the actual state parameters, actual transmission parameters and actual acquisition data corresponding to 4 abnormal events is selected from the data set, such as 4 electrostatic parameters; the specific number of n is not limited in the embodiment of the present invention; In the process of uploading actual status parameters, actual transmission parameters and actual collected data to the corresponding process management server through the corresponding multiple network slices, the upload frequency is also n per second.

[0028] Optionally, the process control server configuration request classification model includes: Based on BP neural network, set the request classification model; The request classification model mainly includes two sub-modules. The first one is the request recognition model based on BP neural network; the other one is the request classification model based on probabilistic neural network; The construction process of the request recognition model can be: The request described in the embodiment of the present invention can be implemented by fuzzy text input; For texts containing different keywords, multiple basic requests corresponding to the different keywords are set. The multiple basic requests are manually set, such as a size fault query request, a thermal management fault query request, and an electrostatic anomaly query request, etc. The multiple basic requests at least cover abnormal events; Setting a request recognition training data set, where the request recognition training data set includes multiple texts containing different keywords, and the expected output is multiple basic requests; The BP neural network is trained through the request recognition training data set and expected output to obtain the request recognition model.

[0029] Setting a correction database and correction cycle corresponding to the requested classification model; After the correction cycle, the request classification model is corrected through the correction database.

[0030] In practical applications, the correction database is updated in real time. If the correction database is updated based on historical data, the correction cycle is the update cycle of the correction database. The abnormal event is classified by using an abnormal event classification model, and a first classification result includes at least one of delay, accuracy and transmission frequency.

[0031] Among them, for a certain abnormal event, through the abnormal event processing history and the probabilistic neural network, the abnormal event classification model is set up including: The processing history includes at least the actual state parameters of the abnormalities during the processing, the actual transmission parameters and the delay, accuracy and transmission frequency corresponding to the actual collected data; For a certain abnormal event, obtain the event processing time, event urgency, and event processing completion degree respectively; The urgency of the event includes whether the event will affect whether the packaging equipment will fail or the chip inspection will fail. The urgency is set manually or determined by counting the probability of failure or chip inspection failure after the abnormal event occurs. For example, if the probability of failure or chip inspection failure after the abnormal event occurs is 50%, the urgency of the event is high. The processing completion degree of an event is whether the event is processed within one time; if so, the completion degree is high, otherwise, the completion degree is low; The training set is set as the actual state parameters, actual transmission parameters and delay, accuracy and transmission frequency corresponding to the actual collected data in the abnormal processing history; the output of the probabilistic neural network is set as the processing time of the event, the urgency of the event and the processing completion of the event; The probabilistic neural network is trained through the training set and output to obtain the abnormal event classification model.

[0032] Continue to perform the above steps for the next abnormal event until the abnormal event classification models corresponding to all abnormal events are obtained.

[0033] The abnormal event is classified by the abnormal event classification model, and the first classification result includes at least one of the delay, accuracy and transmission frequency: For the identified abnormal event, obtaining an abnormal event classification model corresponding to the abnormal event; Through the abnormal event classification model, the actual state parameters, actual transmission parameters and delay, accuracy and transmission frequency corresponding to the actual collected data are uploaded to the abnormal event recognition model to classify the abnormal events. The first classification result includes at least one of the delay, accuracy and transmission frequency.

[0034] Optionally, uploading state parameters, transmission parameters, and collected data to a corresponding process management server through corresponding multiple network slices includes: According to the first classification result, slice identifiers corresponding to the plurality of network slices are respectively set for the state parameter, the transmission parameter, and the collected data; Specifically, the process may be: If the probability of anomaly indicated by the final data classification model is greater than 50%, the state parameters, transmission parameters and collected data corresponding to the final data classification model are marked, and the slice identifier of the corresponding secondary network slice is configured; for example, the slice identifier of the secondary network slice corresponding to the final data classification model of the electrostatic parameter is the electrostatic parameter slice identifier; The abnormal event identification model identifies abnormal events based on the actual state parameters, actual transmission parameters and actual collected data uploaded by the final data classification model; and obtains the abnormal event classification model corresponding to the abnormal event through the identified abnormal event; uploads the abnormal actual state parameters, actual transmission parameters and the latency, accuracy and transmission frequency corresponding to the actual collected data to the abnormal event identification model through the abnormal event classification model, and classifies the abnormal events. The first classification result includes at least one of the latency, accuracy and transmission frequency, and according to the identified first classification result, configures the slice identifier of the corresponding first-level network slice.

[0035] According to the slice identifier, upload status parameters, transmission parameters and collection data.

[0036] Through the slice identifier of the secondary network slice, the actual state parameters, actual transmission parameters and actual collected data of the abnormality are uploaded to the process control server where the abnormal event recognition model is located using the corresponding secondary network slice; Through the slice identification of the first-level network slice and using the corresponding first-level network slice, the actual status parameters, actual transmission parameters and actual collected data of the abnormality, as well as their corresponding delay, accuracy and transmission frequency are uploaded to the process management server.

[0037] Optionally, the process control server classifies the collection request according to the request classification model, including: According to the request classification model, the basic request corresponding to the collection request is identified; this process is to identify the fuzzy input collection request through the request identification model to obtain the corresponding basic request; If the recognition fails, continue to execute after completion: Obtain keyword correction parameters in the correction database; Correct the basic request by using keyword correction parameters; The corrected collection request is identified through the request identification model to obtain the corresponding basic request; The basic request is classified according to the request classification model.

[0038] Optionally, the classification of the corrected collection request also includes: Preset request classification model; the process can be: Acquire a request classification data set, the request classification data set includes an abnormal process, the abnormal process is determined by an abnormal state parameter, an abnormal transmission parameter and a process where the abnormal collection data is located; Obtain abnormal process identifiers (such as abnormal process names) corresponding to abnormal events from the request classification data set; For the same abnormal process identifier, assuming that the number of abnormal events is m, extract the m abnormal process identifiers corresponding to the m abnormal events from the request classification training sample; set them as the request classification training sample; The probabilistic neural network is trained by the training sample to obtain a request classification model, wherein the classification results are related and irrelevant; Continue to perform the above steps for other process identifiers until the request classification model corresponding to the process identifier is obtained; The corrected collection request is classified according to the request classification model.

[0039] Optionally, the process control server allocates the corresponding network slice according to the second classification result, including: Identify the state parameters, transmission parameters and collected data indicated by the collection request; the process may be: Identify the abnormal process indicated by the collection request and the corresponding abnormal process identifier, and configure the third-level network slice identifier corresponding to the abnormal process according to the abnormal process identifier; Sending to the packaging device a collection sub-request corresponding to the state parameter, transmission parameter and collection data indicated by the collection request; Decomposing the collection request into a plurality of collection sub-requests corresponding to the abnormal processes respectively according to different processes; Set the slice ID corresponding to the feedback data of the acquisition subrequest.

[0040] Through the third-level network slice identifier, use the corresponding third-level network slice to issue a collection sub-request; After receiving the collection sub-request, the encapsulation device executes: If the probability of anomaly indicated by the final data classification model is greater than 50%, the state parameters, transmission parameters and collected data corresponding to the final data classification model are marked, and the slice identifier of the corresponding secondary network slice is configured; The abnormal event identification model identifies abnormal events based on the actual state parameters, actual transmission parameters and actual collected data uploaded by the final data classification model; and obtains the abnormal event classification model corresponding to the abnormal event through the identified abnormal event; uploads the abnormal actual state parameters, actual transmission parameters and the latency, accuracy and transmission frequency corresponding to the actual collected data to the abnormal event identification model through the abnormal event classification model, and classifies the abnormal events. The first classification result includes at least one of the latency, accuracy and transmission frequency, and according to the identified first classification result, configures the slice identifier of the corresponding first-level network slice.

[0041] According to the slice identifier, upload status parameters, transmission parameters and collection data.

[0042] Through the slice identifier of the secondary network slice, the actual state parameters, actual transmission parameters and actual collected data of the abnormality are uploaded to the process control server where the abnormal event recognition model is located using the corresponding secondary network slice; Through the slice identification of the first-level network slice and using the corresponding first-level network slice, the actual status parameters, actual transmission parameters and actual collected data of the abnormality, as well as their corresponding delay, accuracy and transmission frequency are uploaded to the process management server.

[0043] Reference Figure 3 As shown, a chip high-density packaging process management system based on network slicing is provided, the system includes packaging equipment, a process control server and a process management server, the packaging equipment, the process control server and the process management server are connected through a 5G network, and the 5G network configures multiple network slices corresponding to the process management server through network slicing technology; wherein: The process control server is used to configure the data classification model and request the classification model; The encapsulation device is used to obtain its own state parameters and transmission parameters respectively; The encapsulation device is used to collect the data collected by the monitoring device; The process control server is used to classify the state parameters, transmission parameters and collected data according to the data classification model, and upload the state parameters, transmission parameters and collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Also includes: The process management server is used to issue collection requests; The process control server is used to classify the collection requests according to the request classification model, and the categories include latency, accuracy and transmission frequency; The process control server is used to allocate the corresponding network slice according to the second classification result; The process control server is used to upload status parameters, transmission parameters and collect data according to the allocated network slice.

[0044] Optionally, the process control server is specifically used for: Set up a data classification model through a probabilistic neural network; The number of neurons in the data classification model is the same as the number of state parameters, transmission parameters and collected data.

[0045] Optionally, the process control server is specifically used for: Based on BP neural network, set the request classification model; Setting a correction database and correction cycle corresponding to the requested classification model; After the correction cycle, the request classification model is corrected through the correction database.

[0046] Optionally, the process control server is specifically used for: Identify abnormal events indicated by status parameters, transmission parameters, and collected data based on the data classification model; The abnormal event is classified, and a first classification result includes at least one of latency, accuracy, and transmission frequency.

[0047] Optionally, the process control server is specifically used for: According to the first classification result, slice identifiers corresponding to the plurality of network slices are respectively set for the state parameter, the transmission parameter, and the collected data; According to the slice identifier, upload status parameters, transmission parameters and collection data.

[0048] Optionally, the process control server is specifically used for: According to the request classification model, identify the basic request corresponding to the collection request; Obtain keyword correction parameters in the correction database; Correct the collection request by using keyword correction parameters; Identify the basic request corresponding to the corrected collection request; Categorize the underlying request.

[0049] Optionally, the process control server is specifically used for: Preset request classification model; The basic request is classified according to the request classification model.

[0050] Optionally, the process control server is specifically used for: Identify the status parameters, transmission parameters and collection data indicated by the collection request; Sending to the packaging device a collection sub-request corresponding to the state parameter, transmission parameter and collection data indicated by the collection request; Set the slice ID corresponding to the feedback data of the acquisition subrequest.

[0051] All the above optional technical solutions can be arbitrarily combined to form optional embodiments of the present invention, which will not be described in detail here.

[0052] It should be noted that: the chip high-density packaging process management system based on network slicing provided in the above embodiment only uses the division of the above-mentioned functional modules as an example when executing the chip high-density packaging process management method based on network slicing. In actual applications, the above-mentioned functions can be assigned to different functional modules as needed, that is, the internal structure of the system is divided into different functional modules to complete all or part of the functions described above. In addition, the chip high-density packaging process management method based on network slicing and the system embodiment provided in the above embodiment belong to the same concept. The specific implementation process is detailed in the method embodiment, which will not be repeated here.

[0053] A person skilled in the art will understand that all or part of the steps to implement the above embodiments may be accomplished by hardware or by instructing related hardware through a program, and the program may be stored in a computer-readable storage medium, and the above-mentioned storage medium may be a read-only memory, a disk or an optical disk, etc.

[0054] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions, improvements, etc. made within the spirit and principle of the present invention should be included in the protection scope of the present invention.

Claims

1. A chip high-density packaging process management method based on network slicing, characterized in that: The method comprises: The process control server configures a data classification model and a request classification model; The packaging devices respectively obtain their own state parameters and transmission parameters; The packaging device collects data collected by the monitoring device; The process control server classifies the state parameter, the transmission parameter and the collected data according to the data classification model, and uploads the state parameter, the transmission parameter and the collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Also includes: The process management server sends a collection request; The process control server classifies the acquisition request according to a request classification model, wherein the categories include latency, accuracy, and transmission frequency; The process control server allocates the corresponding network slice according to the second classification result; The process control server uploads the status parameters, the transmission parameters and the collected data according to the allocated network slice.

2. The method according to claim 1, characterized in that The process control server configuration data classification model includes: Setting the data classification model through a probabilistic neural network; The number of the data classification models is the same as the number of the state parameters, the transmission parameters and the collected data.

3. The method according to claim 1, characterized in that The process control server configuration request classification model includes: Based on BP neural network, set the request classification model; Setting a correction database and a correction cycle corresponding to the request classification model; After the correction cycle, the request classification model is corrected through the correction database.

4. The method according to claim 2, characterized in that: The process control server classifies the state parameter, the transmission parameter and the collected data according to the data classification model, including: identifying, according to the data classification model, abnormal events indicated by the state parameter, the transmission parameter, and the collected data; The abnormal event is classified, and the first classification result includes at least one of latency, accuracy, and transmission frequency.

5. The method according to claim 1, wherein uploading the state parameter, the transmission parameter and the collected data to the corresponding process management server through the corresponding plurality of network slices according to the first classification result comprises: According to the first classification result, slice identifiers corresponding to the multiple network slices are respectively set for the state parameter, the transmission parameter, and the collected data; According to the slice identifier, the state parameter, the transmission parameter and the collected data are uploaded.

6. The method according to claim 3, characterized in that The process control server classifies the collection request according to the request classification model, including: Identify the basic request corresponding to the collection request according to the request classification model; Obtaining keyword correction parameters in the correction database; Correcting the acquisition request by using keyword correction parameters; Identify the basic request corresponding to the corrected collection request; The basic request is classified.

7. The method according to claim 6, characterized in that The classifying the basic request comprises: Preset request classification model; The basic requests are classified according to the request classification model.

8. The method according to claim 1, characterized in that The process control server allocating the corresponding network slice according to the second classification result includes: Identifying the state parameters, transmission parameters and collected data indicated by the collection request; Sending to the packaging device a collection sub-request corresponding to the state parameter, transmission parameter and collection data indicated by the collection request; Set the slice ID corresponding to the feedback data of the acquisition subrequest.

9. A chip high-density packaging process management system based on network slicing, characterized in that: The system includes a packaging device, a process control server and a process management server, wherein the packaging device, the process control server and the process management server are connected via a 5G network, and the 5G network configures a plurality of network slices corresponding to the process management server via network slicing technology; wherein: The process control server is used to configure a data classification model and request a classification model; The encapsulation device is used to obtain its own state parameters and transmission parameters respectively; The packaging device is used to collect data collected by the monitoring device; The process control server is used to classify the state parameter, the transmission parameter and the collected data according to the data classification model, and upload the state parameter, the transmission parameter and the collected data to the corresponding process management server through the corresponding multiple network slices according to the first classification result; Also includes: The process management server is used to issue a collection request; The process control server is used to classify the acquisition request according to the request classification model, and the categories include delay, accuracy and transmission frequency; The process control server is used to allocate the corresponding network slice according to the second classification result; The process control server is used to upload the status parameters, the transmission parameters and the collected data according to the allocated network slice.

10. The system according to claim 9, characterized in that The process control server is specifically used for: Setting the data classification model through a probabilistic neural network; The number of neurons in the data classification model is the same as the number of the state parameters, the transmission parameters and the collected data; and, Based on BP neural network, set the request classification model; Setting a correction database and a correction cycle corresponding to the request classification model; After the correction cycle, the request classification model is corrected through the correction database.

Images