Lithium battery diaphragm production quality tracing method and system based on Internet of Things

Abstract

The invention discloses a lithium battery diaphragm production quality tracing method and system based on the Internet of Things, and relates to the field of lithium battery diaphragm production quality tracing, and the method comprises the steps: collecting historical diaphragm production information and real-time diaphragm production information; key nodes of lithium battery diaphragm production are determined, and a unique quality safety traceability code is set for each key node; constructing a quality anomaly knowledge graph in combination with historical diaphragm production information and quality safety traceability codes; inputting the real-time diaphragm production information into the quality anomaly knowledge graph to obtain quality anomaly parameters; carrying out credibility verification on the quality abnormal parameters; determining a quality abnormality cause by using the root cause analysis model, and generating a compensation parameter according to the quality abnormality cause; and performing parameter compensation on the quality anomaly parameter based on the compensation parameter, and performing quality tracing in the quality anomaly knowledge graph according to the compensated quality anomaly parameter. The management efficiency of lithium battery diaphragm production can be effectively improved.

Description

TECHNICAL FIELD

[0001] The embodiment of the present application relates to the field of lithium battery diaphragm production quality tracing, in particular to a lithium battery diaphragm production quality tracing method and system based on Internet of Things. BACKGROUND

[0002] As a core energy carrier in the field of new energy vehicles, energy storage equipment, consumer electronics and the like, the safety, stability and consistency of the lithium battery directly affect the performance of the terminal product. As a key component inside the battery, the lithium battery diaphragm bears the important functions of isolating the positive and negative electrodes and allowing ions to pass through, and the thickness uniformity, pore size distribution, mechanical strength and other parameters of the lithium battery diaphragm play a decisive role in the cycle life, charging and discharging efficiency and safety of the battery.

[0003] At present, the lithium battery diaphragm production quality tracing work mainly relies on manual recording and local information system. From the production process, the production of the lithium battery diaphragm covers polymerization, extrusion, stretching, coating, slitting, quality inspection and other key processes. However, the data of each process is in an independent storage state. This independent storage mode will cause the problem of "information island", so that the data between the lithium battery diaphragm production processes lack real-time interconnection in the whole chain. When the production quality of the lithium battery diaphragm is comprehensively analyzed and traced, it is difficult to obtain the complete data chain, which leads to the inability to accurately locate the position of the abnormal lithium battery diaphragm production process, and further affects the management efficiency of the lithium battery diaphragm production. For example, in actual production, the temperature change of the extruder may affect the stretching rate of the subsequent stretching process, and the stretching rate may further affect the coating effect of the coating liquid on the diaphragm. At the same time, the difference of the material batch may also lead to the difference of the coating liquid composition, and further affect the quality of the diaphragm. However, due to the lack of effective cross-process correlation analysis method, when the diaphragm quality is abnormal, it is difficult to quickly and accurately locate which process or which parameter is the problem, thereby leading to low efficiency of quality tracing and affecting the management efficiency of the lithium battery diaphragm production.

[0004] At present, there is no good solution to the above problem. SUMMARY

[0005] The embodiment of the present application provides a lithium battery diaphragm production quality tracing method and system based on Internet of Things, which is used for improving the management efficiency of the lithium battery diaphragm production.

[0006] In order to achieve the above purpose, the embodiment of the present application adopts the following technical scheme: In a first aspect, a lithium battery diaphragm production quality tracing method based on Internet of Things is provided, and the method comprises: collecting historical diaphragm production information and real-time diaphragm production information; Based on the historical separator production information, key nodes in lithium battery separator production are identified, and a unique quality and safety traceability code is assigned to each key node. By combining the historical diaphragm production information and the quality and safety traceability code, and using knife-cut Bayesian estimation and prior algorithms, a quality anomaly knowledge graph is constructed. The real-time diaphragm production information is input into the quality anomaly knowledge graph to obtain quality anomaly parameters; The reliability of the aforementioned quality anomaly parameters is verified; If the quality anomaly parameters pass the credibility verification, the root cause analysis model is used to determine the cause of the quality anomaly, and compensation parameters are generated based on the cause of the quality anomaly. The quality anomaly parameters are compensated based on the compensation parameters, and the quality anomaly is traced in the quality anomaly knowledge graph based on the compensated quality anomaly parameters.

[0007] In one possible implementation of the first aspect, the key nodes include production quality nodes and production process nodes. The step of determining the key nodes in lithium battery separator production based on the historical separator production information and assigning a unique quality and safety traceability code to each key node includes the following steps: Select production parameter data and quality indicators from the historical diaphragm production information; The correlation coefficient between the production parameter data and the quality index was calculated using the Pearson correlation coefficient. The production node corresponding to the production parameter data whose correlation coefficient exceeds the preset correlation coefficient is taken as the production quality node. The historical diaphragm production information is used to determine the diaphragm production process flow and the corresponding process steps and process nodes for each process flow. A process flow diagram is drawn for each process flow, combining the process steps and process nodes corresponding to each process flow. Determine the completion time for each of the aforementioned process nodes; The production process nodes are determined based on the process flow chart and the completion time. Assign a unique quality and safety traceability code to each of the key nodes.

[0008] In one possible implementation of the first aspect, the step of combining the historical diaphragm production information and the quality and safety traceability code and constructing a quality anomaly knowledge graph using cut-and-grab Bayes estimation and prior algorithms includes the following steps: Extract a quality anomaly knowledge ontology from the historical diaphragm production information, and construct a quality anomaly knowledge base based on the quality anomaly knowledge ontology; extracting a quality abnormality subject and a quality abnormality object in the quality abnormality knowledge base by using a preset knowledge extraction model; storing the quality abnormality subject, the quality abnormality object and the key node in a form of a triple; extracting an association relationship between all the triples by using a knife-cut Bayesian estimation and a priori algorithm, and generating an association rule; constructing a quality abnormality knowledge graph in combination with the association rule and the triples.

[0009] In a possible implementation manner of the first aspect, the step of extracting an association relationship between all the triples by using a knife-cut Bayesian estimation and a priori algorithm, and generating an association rule includes the following steps: transforming the triples into transaction data sets, wherein each transaction data set represents each quality abnormality case; constituting item set elements from the transaction data sets, wherein each item variable in the item set elements corresponds to each transaction data set; for any one item variable, calculating a number of times that the item variable appears in all the item set elements to obtain an empirical probability; calculating a bias and a variance of the empirical probability by using a knife-cut Bayesian estimation, and adjusting the empirical probability by using the bias and the variance to obtain a probability estimation value; screening out multi-item frequent sets by using a priori algorithm in combination with the probability estimation value; screening out non-empty true subsets in all the frequent sets, and generating candidate association rules between the frequent sets and the non-empty true subsets corresponding to the frequent sets; calculating a confidence degree of each candidate association rule by using a preset confidence degree formula, and taking the candidate association rules with the confidence degree greater than or equal to a preset confidence degree threshold as association rules.

[0010] In a possible implementation manner of the first aspect, the quality abnormality parameters include process abnormality parameters and equipment abnormality parameters, and the step of performing credibility verification on the quality abnormality parameters includes the following steps: when the quality abnormality parameters are the process abnormality parameters, taking process parameters corresponding to all the key nodes within a preset time stamp and the quality safety traceability code as a process combination parameter set; calculating a sliding window hash value of the process combination parameter set by using a hash function within a preset time period; comparing the sliding window hash value with a benchmark process hash value, and determining that the quality abnormality parameters pass the credibility verification if the sliding window hash value is not equal to the benchmark process hash value.

[0011] In a possible implementation manner of the first aspect, the method further includes: when the quality abnormality parameter is the equipment abnormality parameter, determining an initial state parameter set of all lithium battery separator production equipment, and generating an initial hash value based on the initial state parameter set; calculating, by a hash function, an equipment state hash value of an equipment combined parameter set of the lithium battery separator production equipment after each time of completing a production task, wherein the equipment combined parameter set is composed of a batch number of the production task and an equipment state parameter set of the current lithium battery separator production equipment; comparing the equipment state hash value with a reference equipment hash value, and determining that the quality abnormality parameter passes the credibility verification if the equipment state hash value is not equal to the reference equipment hash value.

[0012] In a possible implementation manner of the first aspect, the determining of the cause of the quality abnormality by using the root cause analysis model and the generating of the compensation parameter according to the cause of the quality abnormality include the following steps: inputting the quality abnormality parameter into the root cause analysis model to determine the cause of the quality abnormality of the lithium battery separator production; inputting the cause of the quality abnormality into a preset compensation model to generate the compensation parameter.

[0013] In a possible implementation manner of the first aspect, the determining of the cause of the quality abnormality by inputting the quality abnormality parameter into the root cause analysis model includes the following steps: taking a quality abnormality phenomenon corresponding to the quality abnormality parameter as a top event in the process of the lithium battery separator production; inputting the top event into a fault tree logic chain of the root cause analysis model to identify a minimal cut set of the top event by using Boolean algebra, wherein the minimal cut set is composed of basic events; calculating a basic event probability of each basic event and a minimal cut set probability of each minimal cut set based on a preset probability formula; calculating a probability of occurrence of the top event by using the principle of inclusion and exclusion in combination with the minimal cut set and the basic event probability; calculating a contribution degree of each basic event probability to the probability of occurrence of the top event by using a preset importance degree analysis algorithm; taking a basic event corresponding to the largest contribution degree as the cause of the quality abnormality.

[0014] In the second aspect, the present application provides a machine readable storage medium, and the machine readable storage medium stores instructions for causing a machine to execute the above-mentioned lithium battery separator production quality tracing method based on Internet of Things.

[0015] In a third aspect, the application provides a lithium battery separator production quality traceability system based on the Internet of Things, comprising: a memory configured to store instructions; and a processor configured to call the instructions from the memory and enable the above-mentioned lithium battery separator production quality traceability method based on the Internet of Things when executing the instructions.

[0016] Through the above technical solution, first, by setting a unique quality and safety traceability code for the key nodes of lithium battery separator production, combined with historical separator production information and real-time separator production information, accurate traceability of the whole production process can be achieved. This helps to quickly locate the source of the problem when quality problems occur, improving problem solving efficiency. Second, using the knife-cut Bayesian estimation and prior algorithm to construct a quality anomaly knowledge graph can systematically integrate and analyze various data in the production process, thereby more accurately identifying potential quality anomalies, helping to prevent quality problems from occurring in advance and reduce quality risks. Subsequently, inputting real-time separator production information into the quality anomaly knowledge graph can obtain real-time quality anomaly parameters, which helps to timely discover quality problems in the production process, avoid problem expansion, and reduce losses. The credibility of the quality anomaly parameters is verified to ensure the accuracy of the monitoring results. If the quality anomaly parameters pass the credibility verification, the root cause analysis model is used to determine the causes of the quality anomalies and generate compensation parameters, which helps to analyze the root cause of the problem in depth and provide a basis for formulating effective improvement measures. Based on the compensation parameters, the quality anomaly parameters are compensated, and the quality traceability is performed in the quality anomaly knowledge graph according to the compensated quality anomaly parameters, which helps to timely adjust the production process, optimize the production process, and improve product quality. At the same time, through quality traceability, the responsibility can be clearly attributed, promoting the standardization of production management. Through real-time monitoring and tracing of quality information in the production process, problems can be discovered and solved in a timely manner, avoiding production delays and waste, and improving the management efficiency of lithium battery separator production.

[0017] Other features and advantages of the embodiments of the application will be described in detail in the subsequent specific embodiments section. BRIEF DESCRIPTION OF DRAWINGS

[0018] Figure 1 A flowchart of a lithium battery separator production quality traceability method based on the Internet of Things provided by the embodiments of the application; Figure 2 A structure diagram of a quality anomaly knowledge graph provided by the embodiments of the application. DETAILED DESCRIPTION

[0019] In order to make the purposes, technical solutions and advantages of the embodiments of the present application clearer, the technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. It should be understood that the specific implementation described herein is only used to explain and illustrate the embodiments of the present application, and is not used to limit the embodiments of the present application. Based on the embodiments in the present application, all other embodiments obtained by those of ordinary skill in the art without creative labor fall within the scope of protection of the present application.

[0020] It should be noted that if the embodiments of the present application involve directional indications (such as up, down, left, right, front, back, etc.), the directional indications are only used to explain the relative positional relationship, motion condition, etc. between components in a certain specific posture (as shown in the drawings), and if the specific posture changes, the directional indications also change accordingly.

[0021] In addition, if the embodiments of the present application involve descriptions such as "first", "second", etc., the descriptions of "first", "second", etc. are only for description purposes, and cannot be understood as indicating or implying the relative importance of the indicated technical features or implicitly indicating the number of the indicated technical features. Therefore, the features limited by "first", "second" can explicitly or implicitly include at least one of the features. In addition, the technical solutions of each embodiment can be combined with each other, but it must be based on the fact that a person of ordinary skill in the art can realize it, and when the combination of technical solutions contradicts each other or cannot be realized, it should be considered that the combination of technical solutions does not exist and is not within the scope of protection claimed by the present application.

[0022] Figure 1 A flowchart of a lithium battery separator production quality traceability method based on the Internet of Things according to an embodiment of the present application is schematically shown. As shown in Figure 1 The present application provides a lithium battery separator production quality traceability method based on the Internet of Things, which can include the following steps.

[0023] S110, collecting historical separator production information and real-time separator production information; S120, determining key nodes of lithium battery separator production according to the historical separator production information, and setting a unique quality safety traceability code for each key node; S130, combining the historical separator production information and the quality safety traceability code and constructing a quality anomaly knowledge graph by using a cut Bayes estimation and a priori algorithm; S140, inputting the real-time separator production information into the quality anomaly knowledge graph to obtain a quality anomaly parameter; S150, verifying the credibility of the quality anomaly parameter; S160, if the quality abnormality parameter passes the credibility verification, determining the cause of the quality abnormality by using the root cause analysis model, and generating a compensation parameter according to the cause of the quality abnormality; S170, performing parameter compensation on the quality abnormality parameter based on the compensation parameter, and performing quality tracing in the quality abnormality knowledge graph according to the compensated quality abnormality parameter.

[0024] Firstly, collecting historical diaphragm production information can collect and organize these data through enterprise resource planning system or special data collection software. Historical production information records various parameters and events in the past production process, which helps to trace the source of the problem when quality problems occur. Collecting real-time diaphragm production information can use Internet of Things technology to collect various parameters in the production process in real time, such as temperature, humidity, pressure, speed, etc. The collected data is transmitted to the data center or cloud platform for storage and processing through wired or wireless methods. Real-time production information can timely detect abnormalities in the current production process and prevent problems from being magnified.

[0025] Secondly, according to the historical diaphragm production information, determine the key nodes of lithium battery diaphragm production, and set a unique quality safety traceability code for each key node. In the production process of lithium battery diaphragm, the key nodes refer to those links that have a significant impact on product quality, production efficiency and cost control. By in-depth analysis of historical diaphragm production information, these key nodes can be identified. For example, the quality of raw materials directly affects the performance of the diaphragm. Therefore, the procurement, inspection and storage of raw materials are one of the key nodes. In the diaphragm production process, a variety of raw materials need to be mixed and stirred to ensure their uniformity and consistency. The process parameters of this step have a significant impact on product quality, and the production link of mixing and stirring can also be regarded as a key node. After determining each key node, in order to ensure the traceability of lithium battery diaphragm production quality, a unique quality safety traceability code needs to be set for each key node. These traceability codes can be used to record the production information, quality detection results and operators of each node, so as to quickly locate the source of the problem and take corresponding measures when quality problems occur. In this embodiment, the quality safety traceability code can be determined according to the actual situation, which can be a two-dimensional code or a unique identifier of the production node in the production process. According to the coding rules set by the enterprise, a unique quality safety traceability code is generated for each key node. The traceability code can contain node name, production batch, production date, etc. Information to better identify and manage. In the production process, the traceability code is recorded on the production record and quality detection report of each node. At the same time, a traceability code database is established to quickly query and trace relevant information when needed.

[0026] Subsequently, historical diaphragm production information and quality safety traceability codes are combined and a quality anomaly knowledge graph is constructed using the jackknife Bayesian estimation and prior algorithm. In this embodiment, the jackknife method is a resampling technique used to estimate the stability and bias of statistical quantities. The jackknife method can be used to estimate the degree of influence of different features on diaphragm quality anomalies. By repeatedly deleting one sample or a group of samples and recalculating the statistical quantity, the jackknife distribution of the statistical quantity can be obtained, and the bias and variance of the statistical quantity can be estimated. Combined with Bayesian estimation, the jackknife distribution can be used as prior information to update the posterior probability distribution. The prior algorithm is an algorithm based on Bayes' theorem. The prior algorithm is used to screen multiple frequent sets, and the association between each frequent set is calculated for confidence. The association with a confidence greater than a pre-set confidence threshold is used as the association between triples in the quality anomaly knowledge graph. A pre-set knowledge extraction model is used to extract quality anomaly subjects and quality anomaly objects from the quality anomaly knowledge base. The quality anomaly subject can be the cause of the quality anomaly, such as human, equipment, materials, process or environment, etc. The quality anomaly object refers to the specific object, product or quality anomaly phenomenon affected by the quality anomaly, for example, in diaphragm production, the object can be the diaphragm product itself or its specific quality attribute. The pre-set knowledge extraction model can be an attention mechanism knowledge extraction model. The quality anomaly knowledge graph is constructed in combination with the association, the quality anomaly subject and the quality anomaly object.

[0027] After obtaining the quality abnormality knowledge graph, real-time diaphragm production information is input into the quality abnormality knowledge graph to obtain quality abnormality parameters. Real-time diaphragm production information refers to various data about the production process obtained in real time through various sensors, monitoring equipment, and manual recording during the diaphragm production process, such as real-time parameters of raw materials, operating status of production equipment, production process parameters, and related operation information of operators, etc. The quality abnormality knowledge graph is a knowledge network constructed based on historical data and domain knowledge, which stores the relationships between various factors and quality abnormalities in the diaphragm production process. The nodes in the graph can represent different production parameters, quality indicators, equipment states, raw material characteristics, etc. entities, while the edges represent the association relationships between these entities, such as causal relationships, correlations, etc. When real-time production information is input into the quality abnormality knowledge graph, the knowledge graph will analyze and reason these real-time data based on its existing knowledge and relationship network. It will compare real-time data with historical data to find out whether there are similar patterns or characteristics with known quality abnormality situations. If it finds that some production parameters are outside the normal range or the relationship between some parameters is abnormal, these parameters will be identified as quality abnormality parameters. For example, if the knowledge graph records that under a specific combination of raw material purity and equipment temperature, it is easy to have uneven diaphragm thickness quality problems, and the real-time production information happens to have similar raw material purity and equipment temperature, then these two parameters may be marked as quality abnormality parameters, prompting the operator to pay attention to the relevant links and take timely measures to avoid or correct quality problems.

[0028] Subsequently, the credibility of the quality abnormality parameters is verified. Quality abnormality parameters are various data indicators found in the production process that may be related to product quality problems. However, these parameters may have errors or misleading due to various reasons, and credibility verification first determines whether these parameters are real and effective. In this embodiment, a hash value is introduced to verify the credibility of the quality abnormality parameters. In the credibility verification of the quality abnormality parameters, a hash value is introduced to verify the credibility. The essence is to ensure the integrity and non-tamperability of data in the process of collection, transmission, and storage through cryptographic hash technology, so as to judge whether the parameters are real and effective. Specifically, a hash value can be generated for the real-time collected quality abnormality parameters, and compared with the hash value of historical credible data or block chain storage. If they are consistent, the data has not been tampered with and the credibility is high; if they are not consistent, the data may be contaminated or tampered with and needs to be investigated.

[0029] After the quality abnormality parameter passes the credibility verification, if the quality abnormality parameter passes the credibility verification, the root cause of the quality abnormality is determined by using a root cause analysis model, and a compensation parameter is generated according to the root cause of the quality abnormality. In the present embodiment, the root cause analysis model refers to tracing the root cause of the quality problem according to the quality abnormality parameter by using a systematic method. First, by comparing the multi-dimensional parameters of the normal and abnormal production batches, the root cause analysis model is used to identify the key factors that are strongly related to the quality problem, and to determine the root cause of the quality abnormality parameter. For example, it is found by the root cause analysis model that the root cause of the "abnormal increase of mold temperature" is "insufficient flow caused by wear of the cooling water pump impeller", rather than simply a temperature setting error. After determining the root cause of the quality abnormality, a compensation parameter is generated according to the root cause of the quality abnormality. The root cause of the quality abnormality can be input into a pre-set parameter compensation model, which is a neural network model in the present embodiment. The pre-set parameter compensation model determines which causes are the main factors causing the quality abnormality, and according to the nature and influence degree of the causes, a corresponding compensation strategy is developed, and the compensation strategy is converted into specific compensation parameters. For example, if the root cause is that the equipment aging causes unstable temperature control, then the compensation parameter can be to adjust the set value of the temperature controller or to increase the equipment maintenance frequency.

[0030] Subsequently, the quality abnormality parameters are compensated based on the compensation parameters, and quality traceability is performed in the quality abnormality knowledge graph according to the compensated quality abnormality parameters. The compensation parameters generated according to the previously analyzed causes of quality abnormality are used to adjust or correct the originally detected quality abnormality parameters. For example, if it is found that the size deviation of the product is caused by the excessively high temperature of the mold, the compensation parameters generated according to the root cause analysis may be to reduce the mold temperature setting value and adjust the cooling water flow rate, etc. The compensation parameters are used to adjust the current abnormal mold temperature parameters and related cooling water flow rate parameters. The quality abnormality knowledge graph is a technology for organizing and storing various knowledge and information related to quality abnormality in the form of a graph, which includes various parameters in the production process, equipment information, process information, raw material information, and the mutual relationship among them, as well as knowledge such as past quality abnormality cases and their causes. The compensated quality abnormality parameters are used to search for related information and historical records in the quality abnormality knowledge graph to trace the source and development path of the quality problem. Through the query and matching in the knowledge graph, it can be known whether similar quality abnormality has occurred in the past, what the situation was at that time, what factors caused it, and what measures were taken to solve it, etc. For example, after parameter compensation, the mold temperature and other quality abnormality parameters change. These compensated parameters are input into the quality abnormality knowledge graph for query, which may find that there was a similar case in the past, which was also in the case of similar mold temperature and product size deviation, and finally traced to the quality abnormality caused by the change of a certain property of the raw material. In this way, further analysis and control can be performed on this property of the raw material, so as to realize quality traceability and optimization of the production process.

[0031] By setting a unique quality safety trace code for the key nodes of lithium battery separator production, combining historical separator production information and real-time separator production information, precise tracing of the entire production process can be achieved. This helps to quickly locate the source of the problem when quality problems occur, improving problem-solving efficiency. Second, using the knife-cut Bayesian estimation and prior algorithm to construct a quality anomaly knowledge graph can systematically integrate and analyze various data in the production process, thereby more accurately identifying potential quality anomalies, helping to prevent quality problems from occurring and reduce quality risks. Subsequently, inputting real-time separator production information into the quality anomaly knowledge graph can obtain real-time quality anomaly parameters, helping to timely detect quality problems in the production process and avoid problem expansion and loss reduction. Verifying the credibility of the quality anomaly parameters can ensure the accuracy of the monitoring results. If the quality anomaly parameters pass the credibility verification, the root cause analysis model is used to determine the causes of the quality anomaly and generate compensation parameters, which helps to analyze the root cause of the problem and provide a basis for developing effective improvement measures. Based on the compensation parameters, the quality anomaly parameters are compensated, and the quality trace is performed in the quality anomaly knowledge graph based on the compensated quality anomaly parameters, which helps to timely adjust the production process, optimize the production process, and improve product quality. At the same time, through quality trace, responsibility can be clearly attributed, promoting the standardization of production management. By real-time monitoring and tracing of quality information in the production process, problems can be detected and solved in a timely manner, avoiding production delays and waste, and improving the management efficiency of lithium battery separator production.

[0032] In one embodiment of the present embodiment, the key nodes include production quality nodes and production process nodes. The key nodes of lithium battery separator production are determined according to historical separator production information, and a unique quality safety trace code is set for each key node, including the following steps: S201, selecting production parameter data and quality indicators from historical separator production information; S202, calculating the correlation coefficient between the production parameter data and the quality indicators using the Pearson correlation coefficient; S203, taking the production node corresponding to the production parameter data with a correlation coefficient exceeding a predetermined correlation coefficient as a production quality node; S204, determining the process flow of separator production and the process procedure and process node corresponding to each process flow through historical separator production information; S205, drawing a process flow diagram for each process flow in combination with the process procedure and process node corresponding to each process flow; S206, determining the completion time of each process node; S207, determining the production process node according to the process flow diagram and the completion time; S208, assigning a unique quality safety trace code to each key node.

[0033] First, select the production parameter data and quality indicators from historical diaphragm production information. In this embodiment, historical diaphragm production information refers to a large number of records and materials accumulated during the diaphragm production process, covering various aspects and stages of production, including production time, production equipment, raw material usage, process operation, etc. Production parameter data refers to various specific numerical values or variables used to describe and control the production process during diaphragm production. For example, temperature, pressure, speed, raw material ratio, additive dosage, etc. These parameters will directly affect the production process and the performance of the final product. Quality indicators are a series of standards and parameters used to measure the quality of diaphragm products, such as thickness uniformity, porosity, air permeability, tensile strength, puncture strength, etc. By selecting these quality indicators, we can establish a connection between production parameters and product quality, and thus study how to optimize production parameters to improve product quality, or find the production parameter factors that may cause problems when quality problems occur.

[0034] Next, use the Pearson correlation coefficient to calculate the correlation coefficient between production parameter data and quality indicators. The Pearson correlation coefficient is a statistical indicator used to measure the strength and direction of linear relationship between two variables. By calculating the Pearson correlation coefficient, we can quantify the tightness and direction of the linear relationship between production parameters and quality indicators. Specifically, after calculating the correlation coefficient between production parameter data and quality indicators using the Pearson correlation coefficient, if the correlation coefficient is close to 1, it indicates that there is a strong positive linear correlation between production parameters and quality indicators, i.e. when production parameters increase, quality indicators also tend to increase; if it is close to -1, it indicates that there is a strong negative linear correlation, i.e. when production parameters increase, quality indicators tend to decrease; if it is close to 0, it indicates that the linear relationship between the two is weak, and there may be no obvious linear correlation. By calculating this correlation coefficient, analysts can determine which production parameters have a significant impact on product quality indicators and whether the impact is positive or negative, thereby providing data support and decision-making basis for optimizing the production process and improving product quality. For example, if it is found that the Pearson correlation coefficient between a certain production parameter and a key quality indicator is very high, then this parameter can be monitored and adjusted during the production process in order to achieve better quality control results.

[0035] The production parameter data corresponding to the preset correlation coefficient is used as the production quality node. The correlation coefficient refers to the correlation strength between the production parameters and the quality indicators. In this embodiment, the preset correlation coefficient can be set according to the specific business needs of the enterprise; the production node refers to the key link or equipment in the production process, and each node corresponds to a set of production parameters. After calculating the correlation coefficient between the production parameter data and the quality indicators by using the Pearson correlation coefficient, a specific numerical value is set as the preset correlation coefficient. Then, the production link corresponding to the production parameter data whose calculated correlation coefficient is greater than or equal to the preset correlation coefficient is determined as the production quality node. In this embodiment, the production quality node refers to those production links or process nodes that have a significant impact on the quality of the diaphragm in the lithium battery diaphragm production process. For example, the preset correlation coefficient is set to 0.6, and the calculated correlation coefficient between the production parameters and the quality indicators is 0.7 for the production parameter A and 0.8 for the production parameter B. They all exceed the preset 0.6, so the production nodes corresponding to the production parameters A and B, such as specific processing procedures on the production line, a certain raw material adding link, etc., will be identified as key nodes.

[0036] The process flow of the diaphragm production and the process procedure and process node corresponding to each process flow are determined through historical diaphragm production information. In this embodiment, the process flow refers to the complete production path from raw materials to finished products, which is composed of a series of process procedures with logical order, specifically including dry process, wet process, or coating process, etc. The process procedure refers to the specific processing step in the process flow, which is the basic unit of the flow (for example, "stretching" is a process in the wet process); the process node refers to the key operation point or time point within the process, which usually corresponds to the material state change, equipment action, or quality control point (for example, the time point when the temperature in the stretching process is raised to 150°C). According to the historical diaphragm production information, the past production flow is determined, and the main stages experienced from raw material preparation to diaphragm finished product output are clarified. Based on the time sequence and logical relationship, the order of each process flow is arranged. For example, the dry process generally first prepares the raw materials, then extrudes and casts the base film, and then sequentially performs longitudinal stretching, transverse stretching, and finally rolls up to obtain the finished diaphragm. Under each main process flow, the specific process procedure is further subdivided. Taking the wet process as an example, the "raw material preparation" process includes preparing polyolefin resin, organic solvent, etc.; the "solvent preparation" process is to mix organic solvent and additives; the "casting" process is to cast the prepared solution into a sheet; the "longitudinal stretching" and "transverse stretching" processes are to stretch the cast sheet; the "extraction" process is to remove the solvent; and finally, the "rolling" process. In each process procedure, the key points that have a significant impact on the quality of the diaphragm, production efficiency, etc. are identified, i.e., the process nodes. In the "longitudinal stretching" process of the dry process, the settings of "stretching temperature", "stretching speed", and "stretching multiple" are the key process nodes, and the parameter control of these nodes is directly related to the quality indicators such as the pore structure and mechanical properties of the diaphragm.

[0037] The process procedure and process node corresponding to each process flow are combined to draw the process flow chart of each process flow. After obtaining the process procedure and process node corresponding to each process flow according to the historical diaphragm production information, professional flowchart drawing software can be used to draw the process flow chart of each process flow, taking each main process flow as a unit. Adjacent process rectangular boxes are connected by arrows, and the direction of the arrows represents the execution order of the processes. At the same time, the key node relationship between processes can also be represented by arrows or other connection methods to clarify the input-output relationship between processes and the mutual influence of key nodes.

[0038] determining the completion time of each process node, the completion time refers to the time length experienced from the beginning to the end of each process node according to historical diaphragm production information, for example, in the diaphragm production process, the coating process of the diaphragm starts from the moment when the coating is evenly applied on the base film, to the moment when the coating operation is completed, and the coated diaphragm reaches the specified drying degree or performance requirement, this period of time is the completion time of the coating process, which can be realized through historical diaphragm production information.

[0039] Subsequently, the production flow nodes are determined according to the process flowchart and the completion time, by analyzing the logical relationship and time constraints of each process node in the process flowchart, the nodes that have a key influence on the overall production flow are identified, these nodes determine the efficiency and stability of the production flow. Specifically, the following steps are included: determining the node relationship between the process nodes in each process flowchart, the node relationship includes parallel running nodes, sequential running nodes and merging nodes; drawing a diaphragm production flowchart according to each process flowchart and the node relationship; calculating the earliest start time, the earliest completion time, the latest end time and the latest start time of each process node based on the diaphragm production flowchart, the node relationship and the completion time; taking the process node with the earliest start time equal to the latest start time and the earliest completion time equal to the latest end time as the production flow node.

[0040] First, the node relationship between the process nodes in each process flowchart is determined, the node relationship includes parallel running nodes, sequential running nodes and merging nodes, in this embodiment, the parallel running nodes are process nodes that can be performed simultaneously in the production flow, on different equipment or different production lines at the same time; the sequential running nodes refer to process nodes that must be performed in a specific order, the output of the previous node is the input of the next node, and the next node must start after the previous node is completed; the merging nodes refer to the nodes that converge multiple parallel running nodes or multiple sequential running node paths in the production flow, which need to wait for all the previous nodes to complete before starting. That is, it is clear how these process nodes are related and interact with each other in the production process. The node relationship mainly includes parallel running nodes, sequential running nodes and merging nodes.

[0041] Secondly, a diaphragm production flow chart is drawn according to each process flow chart and node relationship, and in the embodiment, the diaphragm production flow chart refers to integrating each individual process flow chart into a complete diaphragm production flow chart. Specifically, firstly, the starting point and the ending point of the diaphragm production flow chart are determined according to each process flow chart; secondly, the order of each process flow chart is sorted according to the starting point and the ending point of the diaphragm production flow chart; and each process flow chart is arranged in sequence to obtain the diaphragm production flow chart. In the diaphragm production flow chart, the parallel node relationship, the sequential node relationship and the merging node relationship are marked.

[0042] Based on the diaphragm production flow chart, the node relationship and the completion time, the earliest start time, the earliest completion time, the latest end time and the latest start time of each process node are calculated. Firstly, the starting node and the ending node are determined in the diaphragm production flow chart, and the earliest start time of the starting node is initialized. Each process node is traversed from the starting node, and the earliest start time and the earliest completion time are calculated according to the node relationship of each process node. When the node relationship is the parallel running node and the sequential running node, the earliest start time of the process node is equal to the earliest completion time of the preceding node. When the node relationship is the merging node, the maximum value of the earliest completion time of all preceding nodes is taken as the earliest start time of the process node. The earliest start time of each process node is added to the completion time corresponding to each process node to obtain the earliest completion time of each process node. Similarly, each process node is traversed from the ending node, and the latest end time and the latest start time are calculated according to the node relationship of each process node. When the node relationship is the parallel running node and the sequential running node, the latest end time of the process node is equal to the latest start time of the subsequent node. When the node relationship is the merging node, the minimum value of the latest start time of all subsequent nodes is taken as the latest end time of the process node. The latest end time of each process node is subtracted by the completion time to obtain the latest start time.

[0043] From the starting node of the diaphragm production flowchart, the earliest start time (ES) of the starting node is set to 0, and then the earliest finish time (EF) is calculated, which is equal to the earliest start time plus the duration of the node, i.e. EF = ES + duration. The earliest start time (ES) of the subsequent node is calculated, specifically, for sequentially running nodes, the earliest start time of the subsequent node is equal to the earliest finish time of the previous node. For parallel running nodes, the earliest start time of all parallel nodes is the same, usually equal to the earliest finish time of the converging node. The calculation is repeated until the end point, and the earliest start time and the earliest finish time of each node are calculated in turn until the end point of the flowchart. Then, the latest finish time and the latest start time are calculated (reverse traversal), and the latest finish time (LF) of the end point node is determined, which is usually equal to the earliest finish time of the end point node of the flowchart. The latest start time (LS) is calculated, which is equal to the latest finish time minus the duration of the node, i.e. LS = LF - duration. The latest finish time (LF) of the preceding node is calculated, for sequentially running nodes, the latest finish time of the preceding node is equal to the latest start time of the subsequent node. For parallel running nodes, the latest finish time of all parallel nodes is the same, usually equal to the latest start time of the converging node. The calculation is repeated until the starting point, and the latest finish time and the latest start time of each node are calculated in turn in the direction opposite to the arrow of the flowchart, until the starting point of the flowchart.

[0044] The process node with the earliest start time equal to the latest start time and the earliest finish time equal to the latest finish time is taken as the production flow node, the earliest start time refers to the earliest time when the process can start after all the preceding processes are completed; the latest start time refers to the latest time when the process must start in order to ensure the timely completion of the entire project; the earliest finish time refers to the earliest time when the process is successfully completed after starting at the earliest start time, i.e. EF = ES + duration of the process; the latest finish time refers to the latest time when the process must be completed in order not to affect the progress of the entire project. The earliest start time equal to the latest start time, i.e. when ES = LS, indicates that the start time of the process is very critical, there is no buffer time, and it must be strictly started according to this time, otherwise it will affect the progress of the entire production process. The earliest finish time equal to the latest finish time, i.e. when EF = LF, indicates that the completion time of the process is very critical and must be completed at this time without delay. The process node satisfying the above conditions is taken as the production flow node in the production process, which is a specific link or process in the production process that has a critical impact on the progress and efficiency of the entire production process.

[0045] After determining each key node, assigning a unique quality safety traceability code to each key node refers to giving a unique code to those key links that have an important influence on product quality in the product production process, which is used for tracking and tracing product quality safety related information. The quality safety traceability code contains specific information of the node, such as production time, production location, operator, batch of raw materials used, etc. By scanning or querying this traceability code, detailed production information of the product at the key node can be obtained. Using the production management information system of the enterprise or a special traceability code generation software, a unique traceability code is generated for each key node according to the preset coding rule, and the relevant information of the key node is associated and stored, so as to obtain a unique quality safety traceability code.

[0046] By setting a unique quality safety traceability code for each key node, the detailed production information of the product at the key node can be quickly obtained through the quality safety traceability code, and it is determined whether the product meets the quality standard, which effectively improves the accuracy and efficiency of quality traceability, and helps to focus on monitoring and managing these key links.

[0047] In one of the embodiments of the present embodiment, combining historical diaphragm production information and quality safety traceability code and constructing a quality anomaly knowledge graph using knife-cut Bayesian estimation and prior algorithm includes the following steps: S310, extracting quality anomaly knowledge ontology from historical diaphragm production information, and constructing a quality anomaly knowledge base based on the quality anomaly knowledge ontology; S320, extracting quality anomaly subject and quality anomaly object from the quality anomaly knowledge base using a preset knowledge extraction model; S330, storing the quality anomaly subject, the quality anomaly object and the key node in the form of triples; S340, using knife-cut Bayesian estimation and prior algorithm to extract the association relationship between all triples, and generating association rules; S350, constructing a quality anomaly knowledge graph combining the association rules and the triples.

[0048] Figure 2 A structural diagram of a quality anomaly knowledge graph provided by the embodiments of the present application is shown in Figure 2 A plurality of quality anomaly cases are constructed into a quality anomaly knowledge graph, wherein the quality anomaly cases include an extruder temperature anomaly case, a heat setting furnace pressure fluctuation case, a batch porosity exceeding standard case, a lithium battery diaphragm thickness unevenness case, a lithium battery diaphragm wrinkle case, etc. The present diagram is a structural diagram only for illustrative reference.

[0049] In the historical diaphragm production information, quality abnormality knowledge ontology is extracted, and a quality abnormality knowledge base is constructed based on the quality abnormality knowledge ontology. In this embodiment, the quality abnormality knowledge ontology includes entity categories and relationship categories. The entity categories are all quality defect types, quality defect causes, quality defect phenomena, and the like in the lithium battery diaphragm production process. The relationship categories are relationship category mappings between the entity categories, such as the relationships between quality defects and quality defect phenomena, quality defects and quality defect parameters, and the like. From the historical diaphragm production information, the above-mentioned entity categories and relationship categories are identified through data mining, natural language processing, or manual sorting, and the quality abnormality knowledge ontology is obtained. After obtaining the quality abnormality knowledge ontology, the quality abnormality knowledge ontology is stored according to a preset framework to form a searchable and analyzable database. For example, each abnormality record corresponds to an entity instance in the ontology, such as “On May 10, 2025, the temperature of No. 3 coating machine exceeded the standard (180°C→standard 170°C), causing uneven diaphragm thickness”. The relationship between entities is connected through data association, such as the connection between “temperature exceeding the standard” and “uneven thickness” through the “causing” relationship.

[0050] Subsequently, a preset knowledge extraction model is used to extract quality abnormality subjects and quality abnormality objects from the quality abnormality knowledge base. In this embodiment, the quality abnormality subject refers to a cause entity that causes the quality abnormality to occur. For example, the quality abnormality subject of the process type can be coating speed overrun, slurry viscosity anomaly, drying time deficiency, and the like. The quality abnormality subject of the equipment type can be roll press bearing wear, sensor drift, scraper pressure instability, and the like. The quality abnormality object refers to an object entity affected in the quality abnormality event. For example, the quality abnormality object can be lithium battery diaphragm thickness unevenness, porosity overrun, puncture strength deficiency, surface scratch, and the like. The preset knowledge extraction model can be a natural language processing technology or a rule engine, which extracts the quality abnormality subject and the quality abnormality object from the knowledge base in combination with the entity categories and the relationship categories of the quality abnormality knowledge ontology. The subject and object entities in the quality abnormality knowledge base can be identified by a pre-trained named entity recognition model. For example, the abnormality record in the quality abnormality knowledge base is input into the pre-trained named entity recognition model, such as “On May 14, 2025, the scraper of the coating machine was worn (equipment type subject), causing the diaphragm of batch L20250514-001 to have coating peeling (abnormality type object)”. The identified subject is “scraper wear” (equipment type cause), and the identified object is “coating peeling” (abnormality type).

[0051] After obtaining the quality abnormality subject and the quality abnormality object, the quality abnormality subject, the quality abnormality object and the key node are stored in the form of a triple, which is a basic storage unit in a knowledge graph and is composed of a subject, a relationship and an object. The subject refers to a core object that causes a quality problem. The relationship refers to an action relationship between the subject and the object. The object refers to a specific object affected by the quality abnormality. In the embodiment, the quality abnormality subject refers to an object or a factor that causes a quality abnormality, for example, equipment failure, raw material quality problem, improper process parameter setting, operator error, etc., which is a proactive factor in a quality abnormality event. The quality abnormality object refers to an object affected or acted on by the quality abnormality, which can be a quality characteristic of a product, a certain link in a production process or the final product itself, such as product size deviation, performance failure, appearance defect, or a problem in a certain process in the production process. The relationship represents the association relationship or connecting link between the quality abnormality subject and the quality abnormality object, and describes how the subject causes the object to have a quality abnormality. For example, words such as “cause”, “cause”, “cause”, “because”, “make” can be used to represent the association relationship, which embodies the causal relationship or other logical connection between the subject and the object. Through the key node, the root cause of the problem can be traced along the causal chain of the quality abnormality. For example, “insufficient purity of raw materials (quality abnormality subject) causes product performance to decline (quality abnormality object)”, “causes” as an association relationship, according to this relationship, when the product performance is found to decline, the direction indicated by “causes” can be followed to find the source factor of insufficient purity of raw materials, and the process of the production flow corresponding to the source factor is taken as the key node, so as to determine which production process has a problem. The triple form storage is to combine the quality abnormality subject, the key node and the quality abnormality object together to form a fixed structure to store knowledge. Taking “equipment aging causes product cracking” as an example, “equipment aging” is the quality abnormality subject, “product cracking” is the quality abnormality object, and “causes” is the association relationship. Through the quality abnormality object, the quality abnormality subject and the association relationship, the equipment number or the equipment position where the equipment abnormality occurs is found, and the key node corresponding to the production link in the production process where the equipment with the abnormal phenomenon is located is composed of a triple. In this way, a large amount of complex quality abnormality information can be standardized, which is convenient for computer storage, query and analysis, and provides strong support for quality abnormality diagnosis, prevention and improvement.

[0052] The association between all triplets is extracted by using the cut-knife Bayesian estimation and the prior algorithm. The cut-knife Bayesian estimation is a method combining the cut-knife method and Bayesian inference. The cut-knife method is a resampling technique that reduces estimation bias by sampling subsets of the original data multiple times. Bayesian inference is based on Bayes' theorem, which uses prior knowledge and observed data to update the estimate of unknown parameters. In quality anomaly analysis, it is used to estimate the probability of association between triplets. The prior algorithm is an algorithm based on the "support-confidence" framework that generates frequent itemsets through hierarchical search. Specifically, the original quality anomaly data is sampled multiple times using the cut-knife Bayesian estimation, and the probability of association between triplets is calculated using Bayesian inference after each sampling. Then, based on the prior algorithm, the support and confidence are calculated to extract association rules that meet certain conditions from these triplets. For example, through the cut-knife Bayesian estimation, it is found that the probability of "raw material quality problem leading to product performance decline" is high. Then, combined with the prior algorithm, the support and confidence are calculated. If both meet the preset standard, the association rule "if raw material quality problem, then product performance decline" can be generated, thus obtaining the association rule. These association rules can help enterprises quickly locate the causes of quality anomalies and take appropriate measures to prevent and solve quality problems, improving product quality and production efficiency. Association rules are mined from data using cut-knife Bayesian estimation and prior algorithms, which can discover the association between triplets. Through the mining of association rules, more complex association patterns between quality anomaly subjects, key nodes, and quality anomaly objects can be revealed, further enriching the content of the knowledge graph.

[0053] The quality anomaly knowledge graph is constructed by combining association rules and triplets. Triplets provide the most basic factual information for the knowledge graph. Through a large number of triplets, various phenomena and relationships in the quality anomaly field can be described comprehensively, laying the foundation for the construction of the knowledge graph. Association rules provide the ability to predict and reason for the knowledge graph. Based on these rules, new quality anomalies can be predicted and analyzed to help enterprises take preventive measures against quality problems or quickly locate the causes and take appropriate solutions when quality anomalies occur. For example, if there is an association rule in the knowledge graph that "when the equipment operating temperature is too high, there is a high probability of causing product size deviation", when the equipment operating temperature is too high, the product size can be checked in advance or measures can be taken to reduce the equipment temperature to avoid the occurrence of quality anomalies. The extracted triplets and association rules are used to construct the quality anomaly knowledge graph. Special knowledge graph construction tools or databases can be used to visualize triplets and association rules, such as through a graphical interface, users can quickly find all quality anomaly subjects and key nodes related to a certain quality anomaly object, as well as their association paths and strengths.

[0054] By constructing the quality abnormality knowledge graph, various entities and relationships related to the quality abnormality can be intuitively displayed, the quality abnormality problem can be determined in advance, and effective quality traceability and problem solving can be performed on the quality abnormality problem, so that the risk early warning of the quality abnormality is front-loaded, and the product quality of the lithium battery separator production is effectively guaranteed.

[0055] In one of the embodiments of the present embodiment, the association relationship between all the triplets is extracted by using the knife-cut Bayesian estimation and the prior algorithm, and the generation of the association rule includes the following steps: S410, converting the triplets into transaction data sets, wherein each transaction data set represents each quality abnormality case; S420, constructing the transaction data set into item set elements, wherein each item variable in the item set element corresponds to each transaction data set; S430, for any item variable, the number of occurrences of the item variable in all item set elements is calculated to obtain an empirical probability; S440, the deviation and variance of the empirical probability are calculated by using the knife-cut Bayesian estimation, and the empirical probability is adjusted by using the deviation and variance to obtain a probability estimation value; S450, the multi-item frequent set is screened out by using the prior algorithm in combination with the probability estimation value; S460, non-empty true subsets in all frequent sets are screened out, and candidate association rules between the frequent sets and the non-empty true subsets corresponding to the frequent sets are generated; S470, the confidence of each candidate association rule is calculated by using a preset confidence formula, and the candidate association rule with a confidence greater than or equal to a preset confidence threshold is taken as an association rule.

[0056] Firstly, the triplets are converted into transaction data sets, wherein each transaction data set represents each quality abnormality case. The essence of converting the triplets into the transaction data sets is to convert the structured relationship data in the knowledge graph into a standard format suitable for data mining algorithm processing, and the core purpose is to adapt to the data analysis demand and improve the data utilization efficiency. The transaction data set is a data organization form, and in the present embodiment, one transaction data set represents a set of quality abnormality cases, such as extruder temperature abnormality, heat setting furnace pressure fluctuation, batch porosity exceeding standard, etc. When the triplets are converted into the transaction data sets, each triplet represented quality abnormality case is expanded into a transaction containing more related information. For example, for the triplet “excessive impurities in raw materials lead to holes in the separator”, in the converted transaction data set, in addition to the core elements, the time of the quality abnormality case, the production batch involved, the related production equipment, the detection results, etc. Information may also be included to form a complete transaction.

[0057] Secondly, the transaction dataset is constituted into item set elements, wherein each item variable in the item set elements corresponds to each transaction dataset, and the transaction dataset is a set composed of multiple transactions, and each transaction can be regarded as a specific event or case. For example, in the quality abnormality scenario of lithium battery separator production, one transaction can represent a quality abnormality in a production process, and multiple such transactions constitute a transaction dataset. The purpose of constituting the transaction dataset into the item set elements and making the item variable correspond to each transaction dataset is to convert the specific quality abnormality case into calculable structured data, so as to mine potential association rules. The item set elements are a set of all possible feature items extracted from the transaction dataset, and the items can be various information such as quality abnormality subjects, key nodes, quality abnormality objects and the like. For example, from multiple quality abnormality transaction datasets, the item set elements that can be extracted include "equipment aging", "slurry viscosity abnormality", "trigger", "product unqualified" and the like. For each item in the item set elements, an item variable is generated. The item variable is a Boolean variable, which is used to represent whether the item appears in each transaction dataset. For example, if the item set element is "equipment aging", in a certain transaction dataset, if the transaction contains the information of "equipment aging", the item variable corresponding to "equipment aging" takes the value 1; if not, the value is 0. In this way, the transaction dataset is converted into the representation form of item set elements and item variables, so that the data is more convenient for association rule mining.

[0058] Next, for any item variable, the number of times of appearance of the item variable in all item set elements is calculated to obtain an empirical probability. In the embodiment, the item variable refers to an independent element in the transaction dataset, such as an extruder temperature abnormality and a heat setting furnace pressure fluctuation; the item set element is a set composed of item variables, such as {extruder temperature abnormality, heat setting furnace pressure fluctuation}; and the empirical probability is obtained by calculating the number of times of appearance of the item variable in all transactions, and then dividing by the total number of transactions to obtain the frequency of appearance of the item, which is the empirical probability, reflecting the possibility of the appearance of the item in the dataset. The frequency of appearance of the item variable in all item set elements is calculated according to the following formula:

[0059] For each item variable, count the number of times it takes value 1 in all transactions. For example, assume there are 10 itemset elements, and the item variable corresponding to "device failure" takes value 1 in 3 itemset elements. Then the occurrence of "device failure" is 3. Divide the occurrence of the item variable by the total number of transactions to get the empirical probability. Assume the total number of transactions is 10, and the occurrence of "device failure" is 3. Then the empirical probability of "device failure" is 3 ÷ 10 = 0.3, which is 30%. The empirical probability provides a quantitative importance indicator for the item variable by calculating the frequency of the item variable appearing in all itemset elements. According to the empirical probability, the item variables can be sorted, so that those with high frequency and possibly more important item variables can be screened, which helps to focus on these key items in subsequent analysis. In quality anomaly analysis, fault prediction and other scenarios, empirical probability can help identify those item variables that are frequently associated with adverse events or faults, i.e. risk factors. Through empirical probability, support can be provided for decision-making. For example, in the case of limited resources, those item variables with high empirical probability and high risk can be prioritized for processing.

[0060] After obtaining the empirical probability, the bias and variance of the empirical probability are calculated using the jackknife method, and the empirical probability is adjusted by the bias and variance to obtain the probability estimate. The jackknife method is a resampling technique used to estimate the properties of statistical quantities such as bias and variance. It repeatedly removes one observation from the original data and then calculates the statistical quantity based on the remaining data. This process is repeated to obtain a series of statistical quantity estimates. When calculating the bias and variance of the empirical probability, the jackknife method can estimate the stability and accuracy of the empirical probability. For example, for the case where the empirical probability is calculated to be 0.3, the jackknife method is used to remove each item set element in turn and recalculate the probability of the item variable appearing. A series of probability values are obtained. The bias and variance can be calculated from these probability values. The bias reflects the average difference between the empirical probability and the true probability, and the variance measures the dispersion of the probability estimates. Bayesian estimation is an estimation method that takes into account prior information. Based on past experience or knowledge, a prior estimate of the probability of the item variable appearing is made, which is represented by a prior distribution. For example, based on the analysis of similar data in the past, it is believed that the probability of the item variable appearing is more likely to be between 0.2 and 0.4, and a suitable prior distribution can be used to describe it. Then, combining the observed item set data, the prior distribution is updated to the posterior distribution using Bayes' theorem, thereby obtaining a more accurate probability estimate that combines prior information and observed data information. Combining the jackknife method and Bayesian estimation, the bias and variance of the empirical probability are first estimated using the jackknife method, and then some parameters in the Bayesian estimation are adjusted using historical information, or the empirical probability is directly adjusted. For example, if the jackknife method estimates that the bias of the empirical probability is large, it means that the empirical probability may have some systematic errors. In the process of Bayesian estimation, the prior distribution or posterior distribution can be modified according to the direction and size of the bias, or the empirical probability can be directly adjusted to make it closer to the true probability value. The probability estimate obtained in this way not only considers the observed information of the data, but also uses prior knowledge, and at the same time, the bias and variance of the empirical probability are corrected, making the empirical probability more accurate and reliable.

[0061] Subsequently, the multi-item frequent sets are screened out by using the priori algorithm combined with the probability estimation value, which is obtained by methods such as the Bayesian estimation, and reflects the possibility of the occurrence of an item or item set in the data set. For example, in the quality abnormal data of the lithium battery diaphragm production, the probability estimation value of the item "equipment temperature is too high" is 0.3, indicating that the probability of the equipment temperature being too high in all quality abnormal events is 0.3. The probability estimation value provides basic information for the subsequent screening of frequent sets, helping to judge the possibility of the occurrence of an item set. The priori algorithm is based on the "support-confidence" framework, and generates frequent item sets through a layer-by-layer search. The support represents the frequency of the occurrence of an item set in the data set, and the confidence represents the probability of the occurrence of another item set under the condition of the occurrence of an item set. For example, for the item set {equipment temperature is too high, diaphragm thickness is uneven}, the support is the proportion of the occurrence of the item set in all transaction data sets, and the confidence can be the probability of the occurrence of the diaphragm thickness unevenness under the condition of the equipment temperature being too high. The priori algorithm starts from a single item set, and gradually generates multi-item sets, and screens out frequent item sets according to the set support and confidence thresholds. According to the set support and confidence thresholds, the two-item frequent sets are screened out, and the two-item set is a set composed of two items. If the support and confidence of the two-item set both exceed the threshold, it is retained as a frequent item set; otherwise, it is discarded. For example, if the support threshold is set to 0.2 and the confidence threshold is set to 0.6, if the support of the two-item set {equipment temperature is too high, diaphragm thickness is uneven} is 0.3 and the confidence is 0.7, it is regarded as a frequent item set. That is, the probability estimation value is determined by using the priori algorithm whether it is greater than or equal to the minimum support threshold given by the user. If there are item sets greater than or equal to the minimum support threshold given by the user, these item sets are regarded as frequent sets.

[0062] After the frequent sets are obtained, all non-empty proper subsets of the frequent sets are screened out, and candidate association rules between the frequent sets and the non-empty proper subsets corresponding to the frequent sets are generated. In the embodiment, the frequent set refers to an item set whose occurrence frequency in a given transaction data set is greater than or equal to a minimum support threshold; and the non-empty proper subset refers to a subset that does not contain an empty set and has at least one item and is not equal to itself. First, all non-empty proper subsets are found by traversing all the frequent sets. The non-empty proper subset refers to all subsets that do not contain an empty set, that is, at least one item and are not equal to themselves. Next, for each non-empty proper subset, candidate association rules between the frequent sets and the non-empty proper subsets corresponding to the frequent sets are generated. In the embodiment, the candidate association rule is an expression used to describe a potential association relationship between data items and is constructed based on the frequent set and the non-empty proper subset thereof. For example, the frequent set F = {A, B, C}, and the non-empty proper subset S = {A, B}. Then the candidate association rule is {A, B}→{C}, which means that if A and B appear, C is likely to appear. These candidate association rules are summarized based on the frequently occurring item sets in the data and are potential association relationships. Further evaluation and screening are needed through support, confidence and other indicators to determine the reliability and effectiveness of the candidate association rules in actual data.

[0063] The confidence of each candidate association rule is calculated through a preset confidence formula, and the candidate association rules with a confidence greater than or equal to a preset confidence threshold are taken as association rules. In the embodiment, the candidate association rule refers to a possible association rule generated in the association rule mining process. A large number of possible rules are generated through an association rule mining algorithm (such as Apriori and FP-Growth). The confidence is an index for measuring the reliability of the association rule and represents the proportion of transactions containing the antecedent (A) that also contain the consequent (B). The formula for calculating the confidence is as follows:

[0064] wherein Support(A∪B) is the support of the simultaneous occurrence of A and B, Support(A) is the support of the occurrence of A, and Support(B) is the support of the occurrence of B. Confidence(A B) represents the confidence.

[0065] The confidence of each candidate rule is calculated using the preset confidence formula, and the calculated confidence is compared with the preset confidence threshold. If the calculated confidence is greater than or equal to the preset confidence threshold, the candidate association rules are taken as association rules. In the embodiment, the preset confidence threshold can be determined according to actual conditions.

[0066] By extracting the correlation between all triplets by using the knife-cutting Bayesian estimation and prior algorithm, the correlation mode of abnormal events in production or service can be identified and root cause positioning can be assisted, early warning of abnormal events in production can be performed, and related early warning processes can be optimized, thereby improving the management efficiency of abnormal early warning.

[0067] In one of the embodiments of the present embodiment, the quality abnormality parameter includes a process abnormality parameter and a device abnormality parameter, and the credibility verification of the quality abnormality parameter includes the following steps: S510, when the quality abnormality parameter is a process abnormality parameter, taking the process parameters and the quality safety traceability code corresponding to all key nodes within a preset time stamp as a process combination parameter set; S520, calculating a sliding window hash value of the process combination parameter set by using a hash function within a preset time period; S530, comparing the sliding window hash value with a reference process hash value, and if the sliding window hash value is not equal to the reference process hash value, determining that the quality abnormality parameter passes the credibility verification.

[0068] In the present embodiment, the quality abnormality parameter includes a process abnormality parameter and a device abnormality parameter, the process abnormality parameter refers to the abnormality of parameters related to the production process, operation method and process control, which directly or indirectly leads to product quality problems, and the device abnormality parameter refers to the abnormality of the running state parameters of production equipment (such as coating machines, rolling machines and cutting machines), which leads to the failure of equipment functions or the decline of precision, and further causes quality problems. When the quality abnormality parameter is a process abnormality parameter, the process parameters and the quality safety traceability code corresponding to all key nodes within a preset time stamp are taken as a process combination parameter set, in the present embodiment, the preset time stamp can be determined according to actual conditions, and the process abnormality parameter refers to the parameter deviating from the normal range in the production process; that is, when the quality abnormality parameter is a process abnormality parameter, the process parameters and the quality safety traceability code corresponding to all key nodes within the preset time stamp need to be integrated into the process combination parameter set. For example, the combination of "process code + parameter value + time stamp" is taken as the process combination parameter set.

[0069] Subsequently, a sliding window hash value of the process combination parameter set is calculated using a hash function within a preset time period. The essence is to perform real-time fingerprint identification on the process parameters in the production process through the hash technology, for monitoring parameter fluctuation, tracing data integrity, or detecting abnormal changes. The preset time period is divided into multiple fixed-length subintervals, which are the time intervals for each window movement. For the process combination parameter set in each sliding window, a unique value is calculated using a hash function as the digital fingerprint of the window data, for determining the hash values of the process combination parameter sets in different time periods, so as to determine whether there is an abnormal process parameter. In this embodiment, the sliding window hash value can be calculated once every 5 seconds, covering the parameter data in the last 10 minutes. If the window hash value suddenly changes, it indicates that the process parameters in this period have a significant fluctuation, triggering a warning.

[0070] The sliding window hash value is compared with the reference process hash value. If the sliding window hash value is not equal to the reference process hash value, it is determined that the quality abnormal parameter passes the credibility verification. In this embodiment, the reference process hash value can be determined according to the actual situation. The obtained sliding window hash value is compared with the reference process hash value. When the sliding window hash value is not equal to the reference process hash value, it indicates that the credibility of the current window process parameter with the quality abnormal parameter is high, and it is determined that there is a process quality problem. Conversely, if the sliding window hash value is equal to the reference process hash value, the credibility is low.

[0071] Through the sliding window hash comparison, the accurate time node of the parameter mutation can be quickly locked, avoiding time-consuming manual investigation, and effectively preventing data tampering and missing detection.

[0072] In one of the embodiments of the present embodiment, the method further comprises: S610, when the quality abnormal parameter is a device abnormal parameter, determining an initial state parameter set of all lithium battery separator production devices, and generating an initial hash value based on the initial state parameter set; S620, calculating a device state hash value of a device combination parameter set of the lithium battery separator production device after each production task is completed by a hash function, wherein the device combination parameter set is composed of a batch number of the production task and a device state parameter set of the current lithium battery separator production device; S630, comparing the device state hash value with a reference device hash value. If the device state hash value is not equal to the reference device hash value, it is determined that the quality abnormal parameter passes the credibility verification.

[0073] When the quality abnormality parameter is the equipment abnormality parameter, initial state parameter sets of all lithium battery separator production equipment are determined, and initial hash values are generated based on the initial state parameter sets, when the equipment abnormality parameter is detected, the credibility of the abnormality is verified by comparing the initial state hash value of the equipment with the current state hash value, and it is ensured that the equipment state is not tampered or abnormally degraded. In this embodiment, the initial state parameter set refers to the baseline parameter set of the equipment in normal operation, including hardware state, configuration parameter, environment variable, etc. The initial hash value is generated by serializing the initial state parameter set and calculating the fixed length digest by the hash function, as the "digital fingerprint" of the equipment state, for subsequent comparison and verification. The parameters of the equipment in normal operation can be obtained from the equipment management system, the parameter set is converted into a JSON string, and the initial hash value is generated by applying the hash function.

[0074] The equipment state hash value of the equipment combined parameter set of the lithium battery separator production equipment after completing each production task is calculated by the hash function, wherein the equipment combined parameter set is composed of the batch number of the production task and the equipment state parameter set of the current lithium battery separator production equipment. The equipment abnormality parameter refers to the quality abnormality signal caused by the abnormal running state of the equipment in the lithium battery separator production process; the initial state parameter set refers to the key parameter set of all lithium battery separator production equipment in the initial state; the initial hash value is obtained by calculating the initial state parameter set by the hash function. The value is equivalent to the "digital fingerprint" of the equipment state, which is used for subsequent comparison of whether the equipment running state has changed abnormally. After each production task is completed, the equipment combined parameter set is collected, and the equipment combined parameter set (batch number + equipment state parameter set) is calculated again by the hash function to generate the equipment state hash value, which reflects the actual running state of the equipment in a specific batch task.

[0075] Then, the equipment state hash value is compared with the reference equipment hash value, if the equipment state hash value is not equal to the reference equipment hash value, it is determined that the quality abnormality parameter passes the credibility verification, in this embodiment, the reference equipment hash value can be determined according to the actual situation, the reference equipment hash value is generated by the hash function based on the actual state parameters of the equipment after completing a batch of production; the reference equipment hash value refers to the standard fingerprint of the equipment in normal operation, the reference hash value is set in advance, which represents the parameter combination characteristics of the equipment in the normal state. If the equipment state hash value is not equal to the reference equipment hash value, it means that the equipment state deviates from the normal range, and it is determined that the equipment abnormality has high credibility to cause quality problems.

[0076] By verifying the credibility of the equipment abnormality parameter, the hash value is sensitive to the small changes of the equipment parameters, and it is ensured that the production process strictly follows the process standard. The equipment abnormality can be accurately positioned and real-time warned, and the consistency of the lithium battery separator production quality is ensured.

[0077] In one of the embodiments of the present embodiment, the root cause analysis model is used to determine the cause of the quality abnormality, and the compensation parameters are generated according to the cause of the quality abnormality, including the following steps: S710, input the quality abnormality parameter into the root cause analysis model to determine the cause of the quality abnormality in the production of the lithium battery separator; S720, input the cause of the quality abnormality into the preset compensation model to generate the compensation parameter.

[0078] The quality abnormality parameter is input into the root cause analysis model to determine the cause of the quality abnormality in the production of the lithium battery separator. The quality abnormality parameter reflects the specific data index of the quality abnormality in the production of the lithium battery. The root cause analysis model is a model constructed based on historical data, process knowledge or algorithm, which is used to analyze the causal relationship between the "abnormal parameter" and the "quality problem", and find out the most fundamental reason for the abnormality. The real-time or historical quality abnormality parameter is input into the model. For example, if it is detected that the thickness of the lithium battery separator exceeds the standard, and it is found that the pressure fluctuation of the coating machine is large, these two groups of data are input. The root cause analysis model outputs the root cause of the quality abnormality.

[0079] Then, the cause of the quality abnormality is input into the preset compensation model to generate the compensation parameter. In the present embodiment, the preset compensation model can be determined according to the actual situation. The preset compensation model is an algorithm model constructed based on process knowledge or historical data, which can be a neural network model, used to generate adjustment parameters according to the cause, and used to generate correction values for offsetting the abnormal influence. The preset compensation model calculates the compensation parameter that can offset the influence of the abnormal cause according to the preset rules or algorithm, for example, the equipment parameter adjustment reduces the coating speed of the coating machine from 50 mm / s to 45 mm / s; the process parameter correction increases the drying temperature from 80℃ to 85℃, etc.

[0080] Through the synergistic effect of the root cause analysis model and the compensation parameter generation mechanism, the accurate management of the quality problem in the production of the lithium battery separator can be realized, and an accurate quality guarantee solution for the production of the lithium battery separator is also provided.

[0081] In one of the embodiments of the present embodiment, the quality abnormality parameter is input into the root cause analysis model to determine the cause of the quality abnormality in the production of the lithium battery separator, including the following steps: S810, in the process of the production of the lithium battery separator, the quality abnormality phenomenon corresponding to the quality abnormality parameter is taken as a top event; S820, input the top event into the fault tree logic chain of the root cause analysis model to identify the minimum cut set of the top event by using Boolean algebra, wherein the minimum cut set is composed of basic events; S830, calculate the basic event probability of each basic event and the minimal cut set probability of each minimal cut set based on a preset probability formula; S840, calculate the probability of the occurrence of the top event by using the principle of inclusion and exclusion in combination with the minimal cut sets and the basic event probabilities; S850, calculate the contribution degree of each basic event probability to the probability of the occurrence of the top event by using a preset importance analysis algorithm; S860, take the corresponding basic event with the largest contribution degree as the cause of the quality abnormality.

[0082] In the process of lithium battery separator production, the quality abnormality parameter corresponding to the quality abnormality phenomenon is taken as the top event, the quality abnormality parameter is the abnormal data detected in the lithium battery separator production, the quality abnormality phenomenon refers to an observable result such as separator wrinkle, insufficient air permeability, etc., and the top event refers to a core event located at the top layer and needing to be solved first in the fault tree logical chain, for example, insufficient air permeability of the separator leads to short circuit of the battery. The quality abnormality parameter is a quantitative representation of the top event. For example, the top event is unqualified separator thickness, the corresponding parameter is that the measured thickness value exceeds the standard range, and the parameter abnormality is direct evidence of the top event, and the fault tree analysis process can be quickly triggered by monitoring the parameter. The essence of taking the quality abnormality phenomenon as the top event is to use a systematic fault analysis methodology to decompose complex quality problems into traceable, quantifiable and solvable bottom factors.

[0083] Subsequently, the top event is input into the fault tree logical chain of the root cause analysis model to identify the minimal cut sets of the top event by using Boolean algebra, wherein the minimal cut sets are composed of basic events; the fault tree logical chain is a causal relationship network of intermediate events and bottom events obtained by decomposing the system top fault through logical gates layer by layer, reflects the complete path of how the fault is gradually transmitted from the basic event to the system level fault, and is a tool for analyzing the fault propagation mechanism and identifying the key reasons; the minimal cut set is the least combination of basic events that can cause the occurrence of the top event, and the top event will not occur when any event in the combination does not occur; first, the possible direct causes are decomposed layer by layer with the top event as the starting point, and then the basic events are further decomposed to form a logical tree structure. Subsequently, the logical gates in the fault tree are converted into Boolean expressions, for example, the top event T=A∧B (AND gate, A and B occur at the same time to cause T); the top event T=C∨D (OR gate, C or D occurs to cause T), the expression is simplified by Boolean algebra operation, and finally a combination of several minimal cut sets is obtained. The minimal cut set is obtained by Boolean algebra simplification, for example, T=(E∧F)∨G is obtained after simplification, and the minimal cut sets are {E, F} and {G}, which respectively represent that "E and F occur at the same time" will cause the top event; "G occurs alone" will also cause the top event. Each minimal cut set corresponds to a group of basic events that must exist at the same time, and is a necessary and sufficient condition for the occurrence of the top event.

[0084] Next, the basic event probability of each basic event and the minimum cut set probability of each minimum cut set are calculated based on preset probability formulas. The preset probability formulas are the first preset probability formula and the second preset probability formula. In this embodiment, the basic event probability refers to the probability of each bottom-level event (i.e., an event that cannot be further decomposed) occurring in the fault tree; the minimum cut set probability refers to the probability of each cut set occurring within all the minimum cut sets that lead to the top event; a minimum cut set is a set composed of a group of basic events. The basic event probability of each basic event can be calculated based on the first preset probability formula and historical data statistics. The first preset probability formula is:

[0085] Where P(A) represents the probability of the basic event A occurring; n represents the number of times event A occurs during the observation period; and N represents the total number of trials or the total time period during the observation period.

[0086] The probability of a basic event is determined by calculating the probability of each basic event occurring within the historical data using a first preset probability formula.

[0087] The probability of each minimal cut set can be calculated using the second preset probability formula, which is:

[0088] in It is the probability of the j-th minimum cut set. is the probability of the i-th basic event in the cut set, and r is the number of basic events in the cut set.

[0089] The minimum cut set probability of each minimum cut set is calculated using the second preset probability formula.

[0090] Utilizing the inclusion-exclusion principle combined with minimal cut sets and the probability of basic events to calculate the probability of the top event is a precise quantitative method for handling overlapping relationships between multiple minimal cut sets in fault tree analysis of lithium battery separator production. By progressively adding and subtracting the probabilities of overlapping events, duplicate calculations or omissions are avoided, thus accurately obtaining the probability of the union of all minimal cut sets, which is the total probability of the top event. The calculation process begins by calculating the probability of a single minimal cut set, followed by the probability of the intersection between minimal cut sets (two minimal cut sets occurring simultaneously, meaning their basic events occur together). Then, the inclusion-exclusion principle is applied layer by layer to calculate the union probability, and so on, until all possible intersection combinations are included, i.e., calculating the probability of the union of all minimal cut sets, yielding the total probability of the top event.

[0091] The contribution of each basic event probability to the probability of the top event is calculated by a preset importance analysis algorithm. Through importance analysis, the contribution of each basic event to the probability of the top event can be quantified, so as to identify the key components or weak links and provide a basis for optimizing system reliability. The preset importance analysis algorithm can be probability importance (first-order importance), that is, the first-order partial derivative of the basic event probability to the top event probability is calculated. The first-order importance can directly reflect the marginal influence of the change of the basic event probability on the top event probability. Subsequently, on the basis of the probability importance, the occurrence probability of the basic event itself is considered, which is essentially the product of "probability importance" and "basic event probability proportion", and can more truly reflect the actual risk contribution of the basic event. For example, in the production of lithium battery separators, assuming that the top event is "separator thickness abnormality", and the fault tree contains the following basic events: A: insufficient raw material purity (probability p A =0.1); B: coating head temperature out of control (p B =0.05); C: traction speed fluctuation (p C =0.08); Through importance analysis, the probability importance is IA=0.4, IB=0.5, and IC=0.3; the critical importance is Icr,A=0.3, Icr,B=0.4, and Icr,C=0.2. Therefore, the critical importance of the coating head temperature out of control (B) is the highest, indicating that it has the largest actual risk contribution to the separator thickness abnormality, and the temperature control system should be optimized first.

[0092] The corresponding basic event with the largest contribution is taken as the cause of the quality abnormality. The contribution refers to the influence degree of a basic event on the occurrence probability of the top event, which is calculated by the importance analysis algorithm. The larger the value is, the greater the "contribution" of the basic event to the quality abnormality is, that is, the more significant the possibility of the occurrence of the top event is improved when the basic event occurs or changes. The bottom factor that has the greatest influence on the occurrence probability of the quality abnormality is determined through quantitative analysis, and is identified as the main cause. The basic event refers to the bottommost reason leading to the quality abnormality (such as equipment failure, unqualified raw materials, operation error, etc.), which is an event that cannot be further decomposed in the fault tree. The basic event with the greatest influence is screened out as the cause of the quality abnormality through contribution sorting.

[0093] By determining the cause of the quality abnormality in the production of lithium battery separators, the basic event with the greatest contribution can be quickly locked, avoiding misjudgment caused by subjective experience, and potential quality risks can be warned in advance to achieve preventive maintenance. Moreover, the influence degree of each factor on the quality abnormality can be quantified to provide data basis for resource allocation.

[0094] The embodiment of the present application further provides a machine readable storage medium, which stores instructions for causing a machine to execute the above-mentioned lithium battery diaphragm production quality tracing method based on Internet of Things.

[0095] The embodiment of the present application further provides a lithium battery diaphragm production quality tracing system based on Internet of Things, comprising: a memory configured to store instructions; and a processor configured to call the instructions from the memory and capable of realizing the above-mentioned lithium battery diaphragm production quality tracing method based on Internet of Things when executing the instructions.

[0096] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system or a computer program product. Therefore, the present application can adopt a completely hardware embodiment, a completely software embodiment or an embodiment combining software and hardware aspects. Moreover, the present application can adopt a form of a computer program product implemented on one or more computer usable storage media including computer usable program code (including but not limited to disk storage, CD-ROM, optical storage, etc.).

[0097] The present application is described with reference to flowcharts and / or block diagrams of the methods, devices (systems), computer program products according to the embodiments of the present application. It should be understood that each flow and / or block in the flowcharts and / or block diagrams can be realized by computer program instructions, and the combination of the flows and / or blocks in the flowcharts and / or block diagrams. These computer program instructions can be provided to a general-purpose computer, a special-purpose computer, an embedded processor or other programmable data processing devices to produce a machine, so that the instructions executed by the computer or other programmable data processing devices produce a device for realizing the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus for realizing the functions specified in one or more flows and / or blocks.

[0098] These computer program instructions can also be stored in a computer readable memory capable of guiding the computer or other programmable data processing devices to work in a specific way, so that the instructions stored in the computer readable memory produce a product including instruction apparatus, which realizes the functions specified in the flowcharts and / or block diagrams. Figure 1 one or more flows and / or blocks Figure 1 an apparatus for realizing the functions specified in one or more flows and / or blocks.

[0099] These computer program instructions can also be loaded into a computer or other programmable data processing device, so that a series of operation steps are executed on the computer or other programmable device to produce a computer implemented process, so that the instructions executed on the computer or other programmable device provide a product for realizing the functions specified in the flowcharts and / or block diagrams. Figure 1one or more processes and / or blocks Figure 1 steps of a function specified in one or more blocks.

[0100] In one typical arrangement, the computing device includes one or more processors (CPUs), input / output interfaces, network interfaces, and memory.

[0101] The memory can include non-persistent memory and / or volatile memory, such as random access memory (RAM) about which the processor can execute instructions. The memory can also include non-volatile memory, such as read only memory (ROM), electrically programmable read only memory (EPROM), electrically erasable programmable read only memory (EEPROM), flash memory, or other memory technologies, about which the processor can execute instructions. The memory can be a memory storage device of any type.

[0102] Computer readable media includes permanent and non-permanent, removable and non-removable media implemented in any method or technology for storage of information such as computer readable instructions, data structures, program modules or other data. Examples of computer storage media include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read only memory (ROM), electrically programmable read only memory (EEPROM), flash memory or other memory technologies, compact disc read only memory (CD-ROM), digital versatile disc (DVD) or other optical storage, magnetic cassettes, magnetic tapes, magnetic disk storage or other magnetic storage devices, or any other non-transmission medium that can be used to store information accessible to a computing device. According to the definition herein, computer readable media does not include transitory media such as modulated data signals and carrier waves.

[0103] It should also be noted that the terms "comprising", "containing", or any other variant thereof are intended to cover a non-exclusive inclusion, such that a process, method, article or apparatus that comprises a list of elements does not include only those elements recited, but can also include other elements not expressly listed or inherent to such process, method, article or apparatus. Without further limitation, an element defined by the statement "comprising a" does not exclude the presence of additional identical elements in the process, method, article or apparatus that includes the element.

[0104] The above merely provides an embodiment of the present application and is not intended to limit the present application. The present application can have various modifications and changes for those skilled in the art. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present application shall be included in the scope of claims of the present application.

Claims

1. A lithium battery separator production quality traceability method based on Internet of Things, characterized in that, The method comprises the following steps: Collecting historical diaphragm production information and real-time diaphragm production information; Determining key nodes of lithium battery diaphragm production according to the historical diaphragm production information, and setting a unique quality safety traceability code for each key node; Combining the historical diaphragm production information and the quality safety traceability code and constructing a quality anomaly knowledge graph by using a cut Bayesian estimation and a priori algorithm; Inputting the real-time diaphragm production information into the quality anomaly knowledge graph to obtain a quality anomaly parameter; Verifying the credibility of the quality anomaly parameter; If the quality anomaly parameter passes the credibility verification, determining the cause of the quality anomaly by using a root cause analysis model, and generating a compensation parameter according to the cause of the quality anomaly; Based on the compensation parameter, compensating the quality anomaly parameter, and performing quality tracing in the quality anomaly knowledge graph according to the compensated quality anomaly parameter.

2. The method of claim 1, wherein, The key nodes include production quality nodes and production process nodes, and the determination of the key nodes of lithium battery diaphragm production according to the historical diaphragm production information and the setting of a unique quality safety traceability code for each key node comprise the following steps: Selecting production parameter data and quality indicators in the historical diaphragm production information; Calculating the correlation coefficient between the production parameter data and the quality indicators by using a Pearson correlation coefficient; Taking a production node corresponding to the production parameter data with a correlation coefficient exceeding a preset correlation coefficient as a production quality node; Determining a process flow of diaphragm production and a process procedure and a process node corresponding to each process flow through the historical diaphragm production information; Drawing a process flow chart of each process flow in combination with the process procedure and the process node corresponding to each process flow; Determining the completion time of each process node; Determining a production process node according to the process flow chart and the completion time; Allocating a unique quality safety traceability code for each key node.

3. The method of claim 1, wherein, The combination of the historical diaphragm production information and the quality safety traceability code and the construction of a quality anomaly knowledge graph by using a cut Bayesian estimation and a priori algorithm comprise the following steps: Extracting a quality anomaly knowledge ontology from the historical diaphragm production information, and constructing a quality anomaly knowledge base based on the quality anomaly knowledge ontology; Extracting a quality anomaly subject and a quality anomaly object from the quality anomaly knowledge base by using a preset knowledge extraction model; Storing the quality anomaly subject, the quality anomaly object and the key node in a triple form; Extracting the association relationship between all the triples by using a cut Bayesian estimation and a priori algorithm, and generating an association rule; Constructing a quality anomaly knowledge graph in combination with the association rule and the triples.

4. The method of claim 3, wherein, The extraction of the association relationship between all the triples by using a cut Bayesian estimation and a priori algorithm and the generation of an association rule comprise the following steps: Converting the triples into transaction data sets, wherein each transaction data set represents each quality anomaly case; Forming item set elements from the transaction data sets, wherein each item variable in the item set elements corresponds to each transaction data set; For any one of the term variables, the number of occurrences of the term variable in all the term set elements is calculated to obtain an empirical probability; The bias and variance of the empirical probability are calculated using the knife cut Bayesian estimation, and the empirical probability is adjusted by the bias and the variance to obtain a probability estimation value; A priori algorithm is used to screen out multiple frequent sets in combination with the probability estimation value; Non-empty true subsets in all the frequent sets are screened out, and candidate association rules between the frequent sets and the non-empty true subsets corresponding to the frequent sets are generated; The confidence of each candidate association rule is calculated through a preset confidence formula, and the candidate association rule with a confidence greater than or equal to a preset confidence threshold is taken as an association rule.

5. The method of claim 1, wherein, The quality abnormality parameter includes a process abnormality parameter and a device abnormality parameter, and the credibility verification of the quality abnormality parameter includes the following steps: When the quality abnormality parameter is the process abnormality parameter, all process parameters corresponding to the key nodes within a preset time stamp and the quality safety traceability code are taken as a process combination parameter set; A sliding window hash value of the process combination parameter set is calculated within a preset time period using a hash function; The sliding window hash value is compared with a reference process hash value, and if the sliding window hash value is not equal to the reference process hash value, it is determined that the quality abnormality parameter passes the credibility verification.

6. The method of claim 5, wherein, The method further includes: When the quality abnormality parameter is the device abnormality parameter, an initial state parameter set of all lithium battery separator production devices is determined, and an initial hash value is generated based on the initial state parameter set; A device state hash value of a device combination parameter set of the lithium battery separator production device after each production task is completed is calculated by a hash function, wherein the device combination parameter set is composed of a batch number of the production task and a device state parameter set of the current lithium battery separator production device; The device state hash value is compared with a reference device hash value, and if the device state hash value is not equal to the reference device hash value, it is determined that the quality abnormality parameter passes the credibility verification.

7. The method of claim 1, wherein, The determination of the quality abnormality cause by the root cause analysis model and the generation of the compensation parameter according to the quality abnormality cause include the following steps: The quality abnormality parameter is input into the root cause analysis model to determine the quality abnormality cause of the lithium battery separator production; The quality abnormality cause is input into a preset compensation model to generate a compensation parameter.

8. The method of claim 7, wherein, The determination of the quality abnormality cause of the lithium battery separator production by inputting the quality abnormality parameter into the root cause analysis model includes the following steps: During the production of the lithium battery separator, a quality abnormality phenomenon corresponding to the quality abnormality parameter is taken as a top event; The top event is input into a fault tree logic chain of the root cause analysis model to identify the minimal cut set of the top event using Boolean algebra, wherein the minimal cut set is composed of basic events; The basic event probability of each basic event and the minimal cut set probability of each minimal cut set are calculated based on a preset probability formula; The probability of occurrence of the top event is calculated using the inclusion-exclusion principle in combination with the minimal cut set and the basic event probability; calculate a contribution degree of each basic event to the probability of the top event occurrence through a preset importance analysis algorithm; take the corresponding basic event with the largest contribution degree as the quality abnormality cause.

9. A machine-readable storage medium, characterized in that, The machine readable storage medium has stored instructions for causing a machine to perform the lithium battery separator production quality tracing method based on Internet of Things according to any one of claims 1 to 8.

10. An Internet of Things-based lithium battery separator production quality traceability system, characterized in that, comprising: a memory configured to store instructions; and a processor configured to call the instructions from the memory and enable the lithium battery separator production quality tracing method based on Internet of Things according to any one of claims 1 to 8 when the instructions are executed.

Images