Smart park security method and device based on unstructured data processing

Abstract

The present application discloses a smart park security method and device based on unstructured data processing. The server-side method includes: collecting and preprocessing unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park. The multiple data sources include equipment sensors and supply chain systems. The structured data is generated based on the correlation relationship between the target equipment status information and the target production information mined by the pre-trained correlation analysis model; judging the structured data to determine the pain point information existing in the production process of the target park, and the pain point information is used to characterize the risk points that appear in the target park; when the number of risk points that appear in the target park is greater than a preset threshold, generating early warning information of the target park, and sending the early warning information to the client for early warning. Therefore, the use of this application can timely identify and respond to potential problems in the production process, thereby improving production efficiency and reducing manufacturing costs.

Description

Technical Field

[0001] The present application relates to the field of computer technology, and in particular to a smart park security method and device based on unstructured data processing. Background Art

[0002] In modern manufacturing, production efficiency and quality control are the key to corporate competitiveness. With the rise of Industry 4.0, intelligent manufacturing systems play an important role in improving production automation and optimizing resource allocation. Especially in production lines with high output and complex processes, real-time monitoring and early warning mechanisms are essential to prevent production failures, reduce downtime and improve product quality.

[0003] In the existing technology, most manufacturing companies rely on traditional resource planning systems to manage their production processes in industrial parks. Due to the large space, large volume, and scattered management resources, these systems are barely competent in handling tasks such as order management, inventory tracking, and supply chain coordination. However, they have obvious limitations in managing parks as smart parks and integrating and analyzing unstructured data generated during the production process. These data, including machine status, production parameters, and environmental conditions, are usually displayed on their respective devices, scattered throughout the park, and exist in multiple formats and structures, making it difficult for the park system to effectively integrate and deeply analyze this information. This limitation prevents the system from identifying and responding to potential problems in the production process in a timely manner, which in turn affects production efficiency and increases manufacturing costs. Summary of the invention

[0004] The embodiments of the present application provide a smart park security method and device based on unstructured data processing. In order to have a basic understanding of some aspects of the disclosed embodiments, a simple summary is given below. This summary is not a general review, nor is it intended to identify key / important components or describe the scope of protection of these embodiments. Its only purpose is to present some concepts in a simple form as a preface to the detailed description that follows.

[0005] In a first aspect, an embodiment of the present application provides a smart park security method based on unstructured data processing, which is applied to a server, and the method includes:

[0006] Collect and pre-process unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park, the multiple data sources include equipment sensors and supply chain systems, the structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model, the target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors, the target production information is obtained by processing the unstructured data fed back by the supply chain system, the pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between target equipment status information and target production information;

[0007] The structured data is judged to determine the pain points in the production process of the target park. The pain point information is used to characterize the risk points in the target park.

[0008] When the number of risk points appearing in the target park is greater than a preset threshold, early warning information of the target park is generated and sent to a client for early warning.

[0009] Optionally, unstructured data fed back by multiple data sources in the target park are collected and preprocessed to obtain structured data of the target park, including:

[0010] Receive unstructured data from equipment sensors and supply chain systems in real time;

[0011] Eliminate invalid or erroneous data records from the unstructured data fed back by equipment sensors and the unstructured data fed back by the supply chain system to obtain initial equipment status information and initial production information;

[0012] Optimize repeated data entries in initial equipment status information and initial production information to obtain target equipment status information and target production information;

[0013] Input the target device status information and the target production information into a pre-trained correlation analysis model, and output the association relationship between the target device status information and the target production information corresponding to the information; wherein the association relationship includes a mapping relationship between the feature identifier of the key feature of the device status and the feature identifier of the key feature of the production;

[0014] Based on the mapping relationship, the target equipment status information and the target production information are standardized to obtain the structured data of the target park.

[0015] Optionally, based on the mapping relationship, the target device status information and the target production information are standardized to obtain structured data of the target park, including:

[0016] Inputting the target device status information and the target production information into a pre-established sliding window to extract key features in the target device status information and the target production information to obtain device status key features and production key features, wherein the pre-established sliding window includes pre-defined key feature identifiers;

[0017] Perform Box-Cox transformation on the equipment status key features and production key features to obtain the final equipment status key features and final production key features;

[0018] Through the mapping relationship, the final equipment status key features are mapped with the final production key features to obtain a mapping feature sequence;

[0019] The mapping feature sequence is standardized to eliminate the influence of different dimensions and obtain the final mapping feature sequence;

[0020] The final mapping feature sequence is used as the structured data of the target park;

[0021] The transformation formula of Box-Cox transformation is:

[0022] in,

[0023] The final equipment status key characteristics or final production key characteristics of the transformation, yes The transformation parameter is the key feature of the equipment status or the key feature of production. is a positive number to avoid negative logarithms.

[0024] Optionally, the final equipment status key features are mapped to the final production key features through a mapping relationship to obtain a mapping feature sequence, including:

[0025] Traverse and obtain the first key feature in the final key features of the device status;

[0026] Determining a first feature identifier of a first key feature;

[0027] From the mapping relationship, obtain a second feature identifier of the production key feature corresponding to the first feature identifier;

[0028] From the final production key feature, obtain a second key feature corresponding to the second feature identifier of the production key feature;

[0029] Binding the second key feature to the first key feature to obtain a mapping feature;

[0030] Continue to perform the step of traversing and obtaining the first key feature in the final device state key features until all the final device state key features are traversed to obtain a mapping feature sequence.

[0031] Optionally, generate a pre-trained correlation analysis model by following these steps, including:

[0032] collecting first unstructured data and second unstructured data from equipment sensors and supply chain systems;

[0033] extracting first device status information from the first unstructured data, and extracting first production information from the second unstructured data;

[0034] Performing feature conversion on the first device state information and the first production information respectively to obtain a first device state feature sequence and a second production feature sequence;

[0035] Marking the first feature identifier of the key feature of the equipment state and the second feature identifier of the key feature of the production for the first equipment state feature sequence and the second production feature sequence;

[0036] Constructing a first mapping relationship between the first feature identifier and the second feature identifier;

[0037] Binding the first device status information, the first production information, and the first mapping relationship to obtain training data;

[0038] Based on the training data, a pre-trained correlation analysis model is generated.

[0039] Optionally, generate a pre-trained correlation analysis model based on the training data, including:

[0040] The original model was created using a neural network;

[0041] Creating a target loss function for analyzing the mapping relationship between feature identifiers of key features of equipment status and feature identifiers of key features of production;

[0042] Replace the original loss function in the original model with the target loss function to obtain a correlation analysis model;

[0043] Input the training data into the correlation analysis model and output the model loss value;

[0044] Generate a pre-trained correlation analysis model based on the model loss value.

[0045] Optionally, based on the model loss value, generate a pre-trained correlation analysis model, including:

[0046] When the model loss value reaches the minimum, a pre-trained correlation analysis model is generated;

[0047] Alternatively, when the model loss value has not reached the minimum, the model loss value is back-propagated to update the model parameters of the correlation analysis model, and the step of inputting the training data into the correlation analysis model is continued;

[0048] The objective loss function is:

[0049]

[0050] in, is the loss value, is the total number of samples of training data, is the prediction function of the model, The training data set The first device state feature vector of samples, is a weight matrix used to learn the mapping relationship between the first equipment status feature and the second production feature, is a bias term that provides the model with translational freedom along each dimension in the feature space. is with The corresponding second production eigenvector, is the regularization coefficient, is the L2 norm of the weight matrix, which is used to penalize excessive weight values. is the absolute value of the bias term, is the number of feature identifier pairs in the mapping, is a weight parameter used to balance the penalty term of the mapping probability, For the Mapping probability to feature identifier, It is The logarithm of the feature identity mapping probability is used to encourage the model to learn high-probability mapping relationships.

[0051] Optionally, the structured data includes a final mapping feature sequence between the equipment status key features and the production key features;

[0052] The structured data is judged to determine the pain points of the target park in the production process, including:

[0053] Obtain a pre-established pain point information database, which stores a sequence of key-value pairs between mapping features and pain point information, each key-value pair including a key and a value, the key being the mapping feature and the value being the pain point information;

[0054] Match each mapping feature in the final mapping feature sequence with the key of each key-value pair to obtain a matching result for each mapping feature;

[0055] Based on the matching result of each mapping feature, determine the target mapping feature that matches the key;

[0056] Get the value corresponding to the key matched by each target mapping feature, and obtain the pain point information of each target mapping feature;

[0057] The pain point information of each target mapping feature is used as the pain point information existing in the production process of the target park.

[0058] Optionally, generate warning information for the target park, including:

[0059] Extract key parameters from the pain point information of the target park during the production process;

[0060] Input key parameters into the preset warning template to generate warning information for the target park.

[0061] In a second aspect, an embodiment of the present application provides a smart park security device based on unstructured data processing, the device comprising:

[0062] An unstructured data processing module is used to collect and pre-process unstructured data fed back by multiple data sources in a target park to obtain structured data of the target park. The multiple data sources include equipment sensors and supply chain systems. The structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model. The target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors. The target production information is obtained by processing the unstructured data fed back by the supply chain system. The pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between target equipment status information and target production information.

[0063] The pain point information determination module is used to determine the structured data to determine the pain point information existing in the production process of the target park. The pain point information is used to characterize the risk points that appear in the target park.

[0064] The early warning module is used to generate early warning information for the target park when the number of risk points appearing in the target park is greater than a preset threshold, and send the early warning information to the client for early warning.

[0065] The technical solution provided by the embodiments of the present application may have the following beneficial effects:

[0066] In an embodiment of the present application, on the one hand, a pre-trained correlation analysis model can be used to analyze multiple data sources to extract the correlation between target device status information and target production information. This relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, unstructured data can be converted into insightful structured data. This process can effectively integrate and deeply analyze unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds a preset threshold, timely early warning can prevent potential risk points from evolving into larger production problems. This preventive approach not only reduces the frequency of emergency repairs and production interruptions, but also reduces waste caused by production quality problems, thereby effectively reducing manufacturing costs.

[0067] It should be understood that the foregoing general description and the following detailed description are exemplary and explanatory only and are not restrictive of the present application. BRIEF DESCRIPTION OF THE DRAWINGS

[0068] The accompanying drawings, which are incorporated in and constitute a part of this specification, illustrate embodiments consistent with the present application and, together with the description, serve to explain the principles of the present application.

[0069] Figure 1 It is a flow chart of a smart park security method based on unstructured data processing provided in an embodiment of the present application;

[0070] Figure 2 It is a data representation intent of unstructured data provided by an embodiment of the present application;

[0071] Figure 3 It is a schematic diagram of a data table of pain point information provided by this application;

[0072] Figure 4 This is a schematic diagram of early warning information displayed by a client provided by this application;

[0073] Figure 5 This is a schematic diagram of the interaction architecture between a client and a server provided by this application;

[0074] Figure 6 It is a flowchart of a model training method provided by this application;

[0075] Figure 7 It is a structural schematic diagram of a smart park security device based on unstructured data processing provided by the present application;

[0076] Figure 8 It is a structural schematic diagram of an electronic device provided in an embodiment of the present application. DETAILED DESCRIPTION

[0077] The following description and the accompanying drawings sufficiently illustrate specific embodiments of the present application to enable those skilled in the art to practice them.

[0078] It should be clear that the described embodiments are only part of the embodiments of the present application, rather than all the embodiments. Based on the embodiments in the present application, all other embodiments obtained by ordinary technicians in the field without creative work are within the scope of protection of the present application.

[0079] When the following description refers to the drawings, unless otherwise indicated, the same numbers in different drawings represent the same or similar elements. The embodiments described in the following exemplary embodiments do not represent all embodiments consistent with the present application. Instead, they are only examples of devices and methods consistent with some aspects of the present application as detailed in the attached claims.

[0080] In the description of the present application, it should be understood that the terms "first", "second", etc. are used for descriptive purposes only and should not be understood as indicating or implying relative importance. For those of ordinary skill in the art, the specific meanings of the above terms in the present application can be understood according to the specific circumstances. In addition, in the description of the present application, unless otherwise specified, "multiple" refers to two or more. "And / or" describes the association relationship of associated objects, indicating that three relationships may exist. For example, A and / or B can represent: A exists alone, A and B exist at the same time, and B exists alone. The character " / " generally indicates that the objects associated before and after are in an "or" relationship.

[0081] Currently, most manufacturing companies rely on traditional enterprise resource planning systems to manage their production processes. These systems are good at handling tasks such as order management, inventory tracking and supply chain coordination.

[0082] The inventors realized that they have obvious limitations in integrating and analyzing unstructured data generated during the production process. These data, including machine status, production parameters, and environmental conditions, usually exist in multiple formats and structures, making it difficult for the system to effectively integrate and deeply analyze this information. This limitation causes the system to be unable to identify and respond to potential problems in the production process in a timely manner, which in turn affects production efficiency and increases manufacturing costs.

[0083] In order to solve the problem of low response speed and accuracy, the present application provides a smart park security method and device based on unstructured data processing to solve the problems existing in the above-mentioned related technical problems. In the embodiment of the present application, on the one hand, the correlation analysis model trained in advance can be used to analyze the correlation relationship between the target device status information and the target production information extracted from multiple data sources. The relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, the unstructured data can be converted into insightful structured data. The process can effectively integrate and deeply analyze the unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds the preset threshold, the potential risk points can be prevented from evolving into larger production problems through timely warning. This preventive method not only reduces the frequency of emergency maintenance and production interruptions, but also reduces the waste caused by production quality problems, thereby effectively reducing manufacturing costs. The following exemplary embodiments are used for detailed description.

[0084] The following will be combined with the attached Figure 1 -Attached Figure 6 , a smart park security method based on unstructured data processing provided by an embodiment of the present application is introduced in detail. The method can be implemented by relying on a computer program and can be run on a smart park security device based on unstructured data processing based on the von Neumann system. The computer program can be integrated into an application or run as an independent tool application.

[0085] See also Figure 1 , which is a flowchart of a smart park security method based on unstructured data processing, applied to the server. Figure 1 As shown, the method of the embodiment of the present application may include the following steps:

[0086] S101, collecting and preprocessing unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park;

[0087] Among them, multiple data sources include equipment sensors and supply chain systems. Structured data is generated based on the correlation relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model. The target equipment status information is obtained by processing unstructured data feedback from equipment sensors, and the target production information is obtained by processing unstructured data feedback from the supply chain system. The pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the correlation relationship between target equipment status information and target production information.

[0088] Among them, the target park is the park that needs to be warned, and unstructured data refers to data without a fixed format or pattern. In an industrial environment, the data generated by equipment sensors include temperature, pressure, vibration, etc., and the supply chain system includes production quantity and raw material supply. The data source is the starting point or system that provides data, such as equipment sensors and supply chain systems. Equipment sensors refer to sensors installed on machines or equipment, which are used to monitor and collect physical parameters of the equipment's operating status in real time. Target equipment status information refers to information about the equipment's operating status extracted from equipment sensor data, such as whether the equipment is operating normally and whether there is a fault. Target production information refers to information about the production process extracted from the supply chain system, such as raw material supply, production volume, and production progress.

[0089] In some embodiments of the present application, the specific process of collecting and preprocessing unstructured data fed back by multiple data sources in a target park to obtain structured data of the target park includes: receiving unstructured data fed back by equipment sensors and unstructured data fed back by a supply chain system in real time; eliminating invalid or erroneous data records in the unstructured data fed back by equipment sensors and the unstructured data fed back by the supply chain system to obtain initial equipment status information and initial production information; optimizing repeated data entries in the initial equipment status information and the initial production information to obtain target equipment status information and target production information; inputting the target equipment status information and the target production information into a pre-trained correlation analysis model, and outputting the association relationship between the target equipment status information and the target production information corresponding to the information; wherein the association relationship includes a mapping relationship between feature identifiers of key features of the equipment status and feature identifiers of key features of the production; based on the mapping relationship, standardizing the target equipment status information and the target production information to obtain structured data of the target park.

[0090] Invalid or erroneous data records refer to inaccurate or non-compliant data that may be generated during the data collection process and need to be identified and eliminated. Feature identification refers to labels or marks used to identify and distinguish different features.

[0091] In the embodiment of the present application, by receiving and processing unstructured data from equipment sensors and supply chain systems in real time, invalid or erroneous data can be effectively eliminated, and accurate equipment status and production information can be optimized and refined. Using the pre-trained correlation analysis model, the system can deeply explore and identify the key relationship between equipment status and production information, and then map out the complex relationship between the two through feature identification. Based on these mapping relationships, the system further standardizes the target information and generates structured data, providing the park with a clear, accurate and easy-to-analyze data foundation.

[0092] For example Figure 2 , Figure 2 It is a data table of unstructured data, and the data type identifies the type of data. For example, the unstructured data of equipment sensors include temperature readings, pressure readings, vibration data, image data, and sound recordings. For example, the unstructured data of a supply chain system includes order information or inventory records. The data content is the actual unstructured data content, which may include text descriptions, lists of values, file names or file paths, etc. The data source is the device or system that generates the data, such as sensors, cameras, supply chain systems, etc. The receiving time is the time when the data is received by the system. Remarks are additional descriptions of the data, such as the data status (normal, error, warning).

[0093] In some embodiments of the present application, based on the mapping relationship, the target equipment status information and the target production information are standardized to obtain the structured data of the target park. The specific process includes: inputting the target equipment status information and the target production information into a pre-established sliding window to extract the key features in the target equipment status information and the target production information to obtain equipment status key features and production key features, and the pre-established sliding window contains predefined key feature identifiers; performing Box-Cox transformation on the equipment status key features and the production key features to obtain the final equipment status key features and the final production key features; mapping the final equipment status key features and the final production key features through the mapping relationship to obtain a mapping feature sequence; standardizing the mapping feature sequence to eliminate the influence of different dimensions to obtain the final mapping feature sequence; and using the final mapping feature sequence as the structured data of the target park.

[0094] Among them, the transformation formula of Box-Cox transformation is:

[0095] in,

[0096] The final equipment status key characteristics or final production key characteristics of the transformation, yes The transformation parameter is the key feature of the equipment status or the key feature of production. is a positive number to avoid negative logarithms.

[0097] Among them, the target equipment status information and target production information refer to the data of equipment operation status and production process that have been optimized and cleaned for analysis. Sliding window is a data processing technique used to continuously select subsets in a data set, which are particularly useful in time series analysis. Key features are the most important attributes or variables in the data, which can provide an in-depth understanding of the system state or behavior. Key feature identification is a label or indicator used to identify and mark key features in the data. The Box-Cox transformation is a mathematical transformation used to stabilize data variance and make data more normally distributed, which helps to improve the performance of certain statistical analysis methods. The mapping relationship is used to characterize the correlation between equipment status and production information.

[0098] In an embodiment of the present application, by inputting the target equipment status information and the target production information into a pre-established sliding window, the system can efficiently extract key features, which are the core of understanding and predicting equipment performance and production efficiency. By using predefined key feature identifiers, it can be ensured that only those data points that are most important for analysis are focused on. Furthermore, by performing Box-Cox transformation on these key features, the variance of the data can be stabilized and its distribution characteristics can be improved, thereby providing a more robust basis for subsequent analysis. Through the mapping relationship, these transformed features are mapped to obtain mapping feature sequences, which reveal the intrinsic connection between equipment status and production efficiency. Finally, the mapping feature sequence is standardized to eliminate the influence of different dimensions and ensure the consistency and comparability of the data. The final mapping feature sequence is used as the structured data of the target park, which provides accurate and reliable information for production monitoring, fault prediction and decision support, thereby improving the park's operational efficiency and product quality while reducing potential risks.

[0099] In some embodiments of the present application, a pre-trained correlation analysis model is generated according to the following steps, including: collecting first unstructured data and second unstructured data from equipment sensors and supply chain systems; extracting first equipment status information from the first unstructured data, and extracting first production information from the second unstructured data; performing feature conversion on the first equipment status information and the first production information, respectively, to obtain a first equipment status feature sequence and a second production feature sequence; marking a first feature identifier of a key feature of the equipment status and a second feature identifier of a key feature of the production for the first equipment status feature sequence and the second production feature sequence; constructing a first mapping relationship between the first feature identifier and the second feature identifier; data binding the first equipment status information, the first production information and the first mapping relationship to obtain training data; and generating a pre-trained correlation analysis model based on the training data.

[0100] In some embodiments of the present application, the specific process of generating a pre-trained correlation analysis model based on training data includes: using a neural network to create an original model; creating a target loss function for analyzing the mapping relationship between feature identifiers of key features of equipment status and feature identifiers of key production features; replacing the original loss function in the original model with the target loss function to obtain a correlation analysis model; inputting the training data into the correlation analysis model and outputting the model loss value; and generating a pre-trained correlation analysis model based on the model loss value.

[0101] In some embodiments of the present application, the specific process of generating a pre-trained correlation analysis model based on the model loss value includes: generating a pre-trained correlation analysis model when the model loss value reaches the minimum; or, when the model loss value does not reach the minimum, back-propagating the model loss value to update the model parameters of the correlation analysis model, and continuing to execute the step of inputting the training data into the correlation analysis model.

[0102] Among them, the objective loss function is:

[0103]

[0104] in, is the loss value, is the total number of samples of training data, is the prediction function of the model, The training data set The first device state feature vector of samples, is a weight matrix used to learn the mapping relationship between the first equipment status feature and the second production feature, is a bias term that provides the model with translational freedom along each dimension in the feature space. is with The corresponding second production eigenvector, is the regularization coefficient, is the L2 norm of the weight matrix, which is used to penalize excessive weight values. is the absolute value of the bias term, is the number of feature identifier pairs in the mapping, is a weight parameter used to balance the penalty term of the mapping probability, For the Mapping probability to feature identifier, It is The logarithm of the feature identity mapping probability is used to encourage the model to learn high-probability mapping relationships.

[0105] S102, judging the structured data to determine the pain point information existing in the production process of the target park, where the pain point information is used to characterize the risk points occurring in the target park;

[0106] The structured data includes the final mapping feature sequence between the key features of the equipment status and the key features of the production.

[0107] In some embodiments of the present application, the specific process of judging structured data to determine the pain point information existing in the production process of the target park includes: obtaining a pre-established pain point information library, the pain point information library stores a key-value pair sequence between mapping features and pain point information, each key-value pair includes a key and a value, the key is the mapping feature, and the value is the pain point information; matching each mapping feature in the final mapping feature sequence with the key of each key-value pair to obtain the matching result of each mapping feature; based on the matching result of each mapping feature, determining the target mapping feature that matches the key; obtaining the value corresponding to the key matched by each target mapping feature to obtain the pain point information of each target mapping feature; using the pain point information of each target mapping feature as the pain point information existing in the production process of the target park.

[0108] Among them, the pain point information library is a pre-established database or knowledge base that stores the association between mapping features and potential problems or production bottlenecks (pain point information). Key-value pair sequence: a data structure in which each key-value pair contains two elements, one is the key (mapping feature) and the other is the value (pain point information). The matching result is the output obtained during the matching process, indicating whether the mapping feature has found the corresponding key. The target mapping feature is the mapping feature that matches the key in the pain point information library after the match is successful. The pain point information is the specific problem description or bottleneck information associated with the target mapping feature, which is used to identify the problem points in the production process.

[0109] In an embodiment of the present application, by obtaining a pre-established pain point information library, which stores a sequence of key-value pairs between mapping features and pain point information, potential problems in the production process can be quickly identified. Using this library, the system matches the final mapping feature sequence with each key in the library to find the pain point information corresponding to each mapping feature. This matching process not only improves the efficiency of problem identification, but also ensures the accuracy of problem identification. Based on the matching results, the system can determine the target mapping features and obtain the corresponding pain point information, thereby clearly pointing out the specific pain points in the production process of the target park. This method helps to quickly locate problems, accelerate the decision-making process, and take targeted measures to solve production bottlenecks or optimize production processes, ultimately improving production efficiency and product quality, and reducing downtime and costs.

[0110] For example Figure 3 As shown, Figure 3It is a data table of risk points that appear in the target park. Mapping features are features associated with risk points, such as temperature, pressure, vibration, etc. Pain point information is text information that specifically describes the risk point, pointing out specific problems in the production process. Risk level is a level divided according to the severity and urgency of the risk, such as high, medium, and low. The affected production line is the specific production line or production area affected by the risk point. The discovery time is the time when the risk point is identified and recorded by the system.

[0111] S103: When the number of risk points appearing in the target park is greater than a preset threshold, early warning information of the target park is generated and sent to a client for early warning.

[0112] In some embodiments of the present application, the specific process of generating warning information of the target park includes: extracting key parameters from the pain point information of the target park in the production process; inputting the key parameters into a preset warning template to generate warning information of the target park. Figure 4 shown.

[0113] In some implementations, assume that during a day's production process in the target park, the system detects the following risk points. Risk point 1: The motor temperature of production line A exceeds the normal operating range; Risk point 2: The pressure sensor reading of production line B is abnormal; Risk point 3: The vibration sensor of production line C detects an abnormal vibration pattern. Since the number of risk points reaches 3, which is equal to the preset threshold, the system generates the following warning information:

[0114] Warning title: Production line risk warning. Warning content: Currently, 3 risk points have been detected in the park. Please check the following production lines immediately and take corresponding measures: Production line A: Motor overheating, please check the cooling system. Production line B: Pressure sensor abnormality, please check the sensor and related equipment. Production line C: Equipment vibration is too large, please check the stability and balance of the equipment. The system sends this warning information to the park managers via email, SMS or a dedicated park management system client. After receiving the information, the managers can respond quickly according to the warning content to ensure production safety and efficiency.

[0115] For example Figure 5 As shown, the server analyzes the number of risk points in the target park, and the client displays warning information.

[0116] In an embodiment of the present application, on the one hand, a pre-trained correlation analysis model can be used to analyze multiple data sources to extract the correlation between target device status information and target production information. This relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, unstructured data can be converted into insightful structured data. This process can effectively integrate and deeply analyze unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds a preset threshold, timely early warning can prevent potential risk points from evolving into larger production problems. This preventive approach not only reduces the frequency of emergency repairs and production interruptions, but also reduces waste caused by production quality problems, thereby effectively reducing manufacturing costs.

[0117] See also Figure 6 , is a flow chart of a model training method for an elevator operation status recognition model provided in an embodiment of the present application. Figure 6 As shown, the method of the embodiment of the present application may include the following steps:

[0118] S201, collecting first unstructured data and second unstructured data from a device sensor and a supply chain system;

[0119] S202, extracting first equipment status information from the first unstructured data, and extracting first production information from the second unstructured data;

[0120] S203, performing feature conversion on the first device state information and the first production information respectively to obtain a first device state feature sequence and a second production feature sequence;

[0121] S204, marking a first feature identifier of a key feature of an equipment state and a second feature identifier of a key feature of a production for the first equipment state feature sequence and the second production feature sequence;

[0122] S205, constructing a first mapping relationship between the first feature identifier and the second feature identifier;

[0123] S206, data binding the first device status information, the first production information, and the first mapping relationship to obtain training data;

[0124] S207, generating a pre-trained correlation analysis model based on the training data.

[0125] In an embodiment of the present application, on the one hand, a pre-trained correlation analysis model can be used to analyze multiple data sources to extract the correlation between target device status information and target production information. This relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, unstructured data can be converted into insightful structured data. This process can effectively integrate and deeply analyze unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds a preset threshold, timely early warning can prevent potential risk points from evolving into larger production problems. This preventive approach not only reduces the frequency of emergency repairs and production interruptions, but also reduces waste caused by production quality problems, thereby effectively reducing manufacturing costs.

[0126] The following is an embodiment of the device of the present application, which can be used to execute the embodiment of the method of the present application. For details not disclosed in the embodiment of the device of the present application, please refer to the embodiment of the method of the present application.

[0127] See also Figure 7 , which shows a schematic diagram of the structure of a smart park security device based on unstructured data processing provided by an exemplary embodiment of the present application. The smart park security device based on unstructured data processing can be implemented as all or part of an electronic device through software, hardware, or a combination of both. The device 1 includes an unstructured data processing module 10, a pain point information determination module 20, and an early warning module 30.

[0128] The unstructured data processing module 10 is used to collect and pre-process the unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park, wherein the multiple data sources include equipment sensors and supply chain systems, and the structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model, the target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors, and the target production information is obtained by processing the unstructured data fed back by the supply chain system, and the pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between target equipment status information and target production information;

[0129] The pain point information determination module 20 is used to determine the structured data to determine the pain point information existing in the target park during the production process. The pain point information is used to characterize the risk points that appear in the target park.

[0130] The early warning module 30 is used to generate early warning information of the target park when the number of risk points appearing in the target park is greater than a preset threshold, and send the early warning information to a client for early warning.

[0131] It should be noted that the smart campus security device based on unstructured data processing provided in the above embodiment only uses the division of the above functional modules as an example when executing the smart campus security method based on unstructured data processing. In actual applications, the above functions can be assigned to different functional modules as needed, that is, the internal structure of the device is divided into different functional modules to complete all or part of the functions described above. In addition, the smart campus security device based on unstructured data processing provided in the above embodiment and the smart campus security method based on unstructured data processing belong to the same concept. The implementation process is detailed in the method embodiment, which will not be repeated here.

[0132] The serial numbers of the above-mentioned embodiments of the present application are for description only and do not represent the advantages or disadvantages of the embodiments.

[0133] In an embodiment of the present application, on the one hand, a pre-trained correlation analysis model can be used to analyze multiple data sources to extract the correlation between target device status information and target production information. This relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, unstructured data can be converted into insightful structured data. This process can effectively integrate and deeply analyze unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds a preset threshold, timely early warning can prevent potential risk points from evolving into larger production problems. This preventive approach not only reduces the frequency of emergency repairs and production interruptions, but also reduces waste caused by production quality problems, thereby effectively reducing manufacturing costs.

[0134] The present application also provides a computer-readable medium having program instructions stored thereon, which, when executed by a processor, implement the smart park security method based on unstructured data processing provided by the above-mentioned various method embodiments.

[0135] The present application also provides a computer program product comprising instructions, which, when executed on a computer, enables the computer to execute the smart campus security method based on unstructured data processing of the above-mentioned various method embodiments.

[0136] See also Figure 8 , is a schematic diagram of the structure of an electronic device provided in an embodiment of the present application. Figure 8 As shown, the electronic device 1000 may include: at least one processor 1001 , at least one network interface 1004 , a user interface 1003 , a memory 1005 , and at least one communication bus 1002 .

[0137] The communication bus 1002 is used to realize the connection and communication between these components.

[0138] The user interface 1003 may include a display screen (Display) and a camera (Camera). Optionally, the user interface 1003 may also include a standard wired interface and a wireless interface.

[0139] The network interface 1004 may optionally include a standard wired interface or a wireless interface (such as a WI-FI interface).

[0140] Among them, the processor 1001 may include one or more processing cores. The processor 1001 uses various interfaces and lines to connect various parts of the entire electronic device 1000, and executes various functions and processes data of the electronic device 1000 by running or executing instructions, programs, code sets or instruction sets stored in the memory 1005, and calling data stored in the memory 1005. Optionally, the processor 1001 can be implemented in at least one hardware form of digital signal processing (Digital Signal Processing, DSP), field programmable gate array (Field-Programmable Gate Array, FPGA), and programmable logic array (Programmable Logic Array, PLA). The processor 1001 can integrate one or a combination of a central processing unit (Central Processing Unit, CPU), a graphics processing unit (Graphics Processing Unit, GPU) and a modem. Among them, the CPU mainly processes the operating system, user interface and application programs; the GPU is responsible for rendering and drawing the content to be displayed on the display screen; the modem is used to process wireless communications. It can be understood that the above-mentioned modem may not be integrated into the processor 1001, and it can be implemented separately through a chip.

[0141] Among them, the memory 1005 may include a random access memory (Random Access Memory, RAM) and may also include a read-only memory (Read-Only Memory). Optionally, the memory 1005 includes a non-transitory computer-readable storage medium. The memory 1005 can be used to store instructions, programs, codes, code sets or instruction sets. The memory 1005 may include a program storage area and a data storage area, wherein the program storage area may store instructions for implementing an operating system, instructions for at least one function (such as a touch function, a sound playback function, an image playback function, etc.), instructions for implementing the above-mentioned various method embodiments, etc.; the data storage area may store data involved in the above-mentioned various method embodiments, etc. The memory 1005 may also be optionally at least one storage system located away from the aforementioned processor 1001. As Figure 8 As shown, the memory 1005 as a computer storage medium may include an operating system, a network communication module, a user interface module, and a smart campus security application based on unstructured data processing.

[0142] exist Figure 8 In the electronic device 1000 shown, the user interface 1003 is mainly used to provide an input interface for the user and obtain the data input by the user; and the processor 1001 can be used to call the smart campus security application based on unstructured data processing stored in the memory 1005, and specifically perform the following operations:

[0143] Collect and pre-process unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park, the multiple data sources include equipment sensors and supply chain systems, the structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model, the target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors, the target production information is obtained by processing the unstructured data fed back by the supply chain system, the pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between target equipment status information and target production information;

[0144] The structured data is judged to determine the pain points in the production process of the target park. The pain point information is used to characterize the risk points in the target park.

[0145] When the number of risk points appearing in the target park is greater than a preset threshold, early warning information of the target park is generated and sent to a client for early warning.

[0146] In one embodiment, when the processor 1001 collects and pre-processes unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park, the processor 1001 specifically performs the following operations:

[0147] Receive unstructured data from equipment sensors and supply chain systems in real time;

[0148] Eliminate invalid or erroneous data records from the unstructured data fed back by equipment sensors and the unstructured data fed back by the supply chain system to obtain initial equipment status information and initial production information;

[0149] Optimize repeated data entries in initial equipment status information and initial production information to obtain target equipment status information and target production information;

[0150] Input the target device status information and the target production information into a pre-trained correlation analysis model, and output the association relationship between the target device status information and the target production information corresponding to the information; wherein the association relationship includes a mapping relationship between the feature identifier of the key feature of the device status and the feature identifier of the key feature of the production;

[0151] Based on the mapping relationship, the target equipment status information and the target production information are standardized to obtain the structured data of the target park.

[0152] In one embodiment, when the processor 1001 performs standardized processing on the target device status information and the target production information based on the mapping relationship to obtain structured data of the target park, the processor 1001 specifically performs the following operations:

[0153] Inputting the target device status information and the target production information into a pre-established sliding window to extract key features in the target device status information and the target production information to obtain device status key features and production key features, wherein the pre-established sliding window includes pre-defined key feature identifiers;

[0154] Perform Box-Cox transformation on the equipment status key features and production key features to obtain the final equipment status key features and final production key features;

[0155] Through the mapping relationship, the final equipment status key features are mapped with the final production key features to obtain a mapping feature sequence;

[0156] The mapping feature sequence is standardized to eliminate the influence of different dimensions and obtain the final mapping feature sequence;

[0157] The final mapping feature sequence is used as the structured data of the target park;

[0158] The transformation formula of Box-Cox transformation is:

[0159] in,

[0160] The final equipment status key characteristics or final production key characteristics of the transformation, yes The transformation parameter is the key feature of the equipment status or the key feature of production. is a positive number to avoid negative logarithms.

[0161] In one embodiment, when the processor 1001 maps the final equipment status key feature with the final production key feature through the mapping relationship to obtain the mapping feature sequence, the processor 1001 specifically performs the following operations:

[0162] Traverse and obtain the first key feature in the final key features of the device status;

[0163] Determining a first feature identifier of a first key feature;

[0164] From the mapping relationship, obtain a second feature identifier of the production key feature corresponding to the first feature identifier;

[0165] From the final production key feature, obtain a second key feature corresponding to the second feature identifier of the production key feature;

[0166] Binding the second key feature to the first key feature to obtain a mapping feature;

[0167] Continue to perform the step of traversing and obtaining the first key feature in the final device state key features until all the final device state key features are traversed to obtain a mapping feature sequence.

[0168] In one embodiment, when the processor 1001 generates a pre-trained correlation analysis model, the processor 1001 specifically performs the following operations:

[0169] collecting first unstructured data and second unstructured data from equipment sensors and supply chain systems;

[0170] extracting first device status information from the first unstructured data, and extracting first production information from the second unstructured data;

[0171] Performing feature conversion on the first device state information and the first production information respectively to obtain a first device state feature sequence and a second production feature sequence;

[0172] Marking the first feature identifier of the key feature of the equipment state and the second feature identifier of the key feature of the production for the first equipment state feature sequence and the second production feature sequence;

[0173] Constructing a first mapping relationship between the first feature identifier and the second feature identifier;

[0174] Binding the first device status information, the first production information, and the first mapping relationship to obtain training data;

[0175] Based on the training data, a pre-trained correlation analysis model is generated.

[0176] In one embodiment, when the processor 1001 generates a pre-trained correlation analysis model based on training data, the processor 1001 specifically performs the following operations:

[0177] The original model was created using a neural network;

[0178] Creating a target loss function for analyzing the mapping relationship between feature identifiers of key features of equipment status and feature identifiers of key features of production;

[0179] Replace the original loss function in the original model with the target loss function to obtain a correlation analysis model;

[0180] Input the training data into the correlation analysis model and output the model loss value;

[0181] Generate a pre-trained correlation analysis model based on the model loss value.

[0182] In one embodiment, when the processor 1001 generates a pre-trained correlation analysis model based on the model loss value, the processor 1001 specifically performs the following operations:

[0183] When the model loss value reaches the minimum, a pre-trained correlation analysis model is generated;

[0184] Alternatively, when the model loss value has not reached the minimum, the model loss value is back-propagated to update the model parameters of the correlation analysis model, and the step of inputting the training data into the correlation analysis model is continued;

[0185] The objective loss function is:

[0186]

[0187] in, is the loss value, is the total number of samples of training data, is the prediction function of the model, The training data set The first device state feature vector of samples, is a weight matrix used to learn the mapping relationship between the first equipment status feature and the second production feature, is a bias term that provides the model with translational freedom along each dimension in the feature space. is with The corresponding second production eigenvector, is the regularization coefficient, is the L2 norm of the weight matrix, which is used to penalize excessive weight values. is the absolute value of the bias term, is the number of feature identifier pairs in the mapping, is a weight parameter used to balance the penalty term of the mapping probability, For the Mapping probability to feature identifier, It is The logarithm of the feature identity mapping probability is used to encourage the model to learn high-probability mapping relationships.

[0188] In one embodiment, when the processor 1001 performs judgment on the structured data to determine the pain point information existing in the production process of the target park, the processor 1001 specifically performs the following operations:

[0189] Obtain a pre-established pain point information database, which stores a sequence of key-value pairs between mapping features and pain point information, each key-value pair including a key and a value, the key being the mapping feature and the value being the pain point information;

[0190] Match each mapping feature in the final mapping feature sequence with the key of each key-value pair to obtain a matching result for each mapping feature;

[0191] Based on the matching result of each mapping feature, determine the target mapping feature that matches the key;

[0192] Get the value corresponding to the key matched by each target mapping feature, and obtain the pain point information of each target mapping feature;

[0193] The pain point information of each target mapping feature is used as the pain point information existing in the production process of the target park.

[0194] In one embodiment, when the processor 1001 generates warning information of the target park, the processor 1001 specifically performs the following operations:

[0195] Extract key parameters from the pain point information of the target park during the production process;

[0196] Input key parameters into the preset warning template to generate warning information for the target park.

[0197] In an embodiment of the present application, on the one hand, a pre-trained correlation analysis model can be used to analyze multiple data sources to extract the correlation between target device status information and target production information. This relationship reveals the specific impact of the device status on the production process. Therefore, based on the correlation relationship, unstructured data can be converted into insightful structured data. This process can effectively integrate and deeply analyze unstructured data, so that the system can timely identify and respond to potential problems in the production process based on structured data, thereby improving production efficiency; on the other hand, when the number of identified risk points exceeds a preset threshold, timely early warning can prevent potential risk points from evolving into larger production problems. This preventive approach not only reduces the frequency of emergency repairs and production interruptions, but also reduces waste caused by production quality problems, thereby effectively reducing manufacturing costs.

[0198] Those skilled in the art can understand that all or part of the processes in the above-mentioned embodiments can be implemented by instructing the relevant hardware through a computer program, and the program for smart campus security based on unstructured data processing can be stored in a computer-readable storage medium, and when the program is executed, it can include the processes of the embodiments of the above-mentioned methods. Among them, the storage medium of the program for smart campus security based on unstructured data processing can be a disk, an optical disk, a read-only storage memory, or a random access memory, etc.

[0199] The above disclosure is only the preferred embodiment of the present application, which certainly cannot be used to limit the scope of rights of the present application. Therefore, equivalent changes made according to the claims of the present application are still within the scope covered by the present application.

Claims

1. A smart park security method based on unstructured data processing, characterized in that: Applied to the server, the method includes: Collect and pre-process unstructured data fed back by multiple data sources in a target park to obtain structured data of the target park, wherein the multiple data sources include equipment sensors and supply chain systems, and the structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model, the target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors, and the target production information is obtained by processing the unstructured data fed back by the supply chain system, and the pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between the target equipment status information and the target production information; wherein, the collection and pre-processing of unstructured data fed back by multiple data sources in the target park to obtain structured data of the target park includes: Receiving in real time the unstructured data fed back by the equipment sensor and the unstructured data fed back by the supply chain system; Eliminate invalid or erroneous data records in the unstructured data fed back by the equipment sensor and the unstructured data fed back by the supply chain system to obtain initial equipment status information and initial production information; Optimizing repeated data entries in the initial device status information and the initial production information to obtain target device status information and target production information; Input the target device status information and the target production information into a pre-trained correlation analysis model, and output the association relationship between the target device status information and the target production information corresponding to the information; wherein the association relationship includes a mapping relationship between the feature identifier of the key feature of the device status and the feature identifier of the key feature of the production; Based on the mapping relationship, the target device status information and the target production information are standardized to obtain the structured data of the target park; wherein, Based on the mapping relationship, the target device status information and the target production information are standardized to obtain structured data of the target park, including: Inputting the target device state information and the target production information into a pre-established sliding window to extract key features in the target device state information and the target production information to obtain device state key features and production key features, wherein the pre-established sliding window includes pre-defined key feature identifiers; Performing Box-Cox transformation on the equipment status key features and the production key features to obtain final equipment status key features and final production key features; Mapping the final equipment status key feature with the final production key feature through the mapping relationship to obtain a mapping feature sequence; Standardizing the mapping feature sequence to eliminate the influence of different dimensions and obtain a final mapping feature sequence; Using the final mapping feature sequence as structured data of the target park; The structured data is judged to determine the pain point information existing in the production process of the target park, and the pain point information is used to characterize the risk points occurring in the target park; When the number of risk points appearing in the target park is greater than a preset threshold, early warning information of the target park is generated and sent to a client for early warning.

2. The method according to claim 1, characterized in that The transformation formula of the Box-Cox transformation is: in, The final equipment status key characteristics or final production key characteristics of the transformation, yes The transformation parameter is the key feature of the equipment status or the key feature of production. is a positive number to avoid negative logarithms.

3. The method according to claim 1, characterized in that The final equipment status key feature is mapped with the final production key feature through the mapping relationship to obtain a mapping feature sequence, including: Traversing and obtaining the first key feature in the final device state key features; Determining a first feature identifier of the first key feature; From the mapping relationship, obtain a second feature identifier of the production key feature corresponding to the first feature identifier; From the final production key feature, obtaining a second key feature corresponding to the second feature identifier of the production key feature; Binding the second key feature with the first key feature to obtain a mapping feature; The step of traversing and acquiring the first key feature in the final device state key features is continued until all the final device state key features are traversed to obtain a mapping feature sequence.

4. The method according to claim 1, generating a pre-trained correlation analysis model according to the following steps, comprising: Collecting first unstructured data and second unstructured data from the device sensor and the supply chain system; extracting first equipment status information from the first unstructured data, and extracting first production information from the second unstructured data; Performing feature conversion on the first device state information and the first production information respectively to obtain a first device state feature sequence and a second production feature sequence; Marking the first feature identifier of the key feature of the equipment state and the second feature identifier of the key feature of the production for the first equipment state feature sequence and the second production feature sequence; Constructing a first mapping relationship between the first feature identifier and the second feature identifier; Binding the first device status information, the first production information, and the first mapping relationship to obtain training data; Based on the training data, a pre-trained correlation analysis model is generated.

5. The method according to claim 4, characterized in that The generating a pre-trained correlation analysis model based on the training data comprises: The original model was created using a neural network; Creating a target loss function for analyzing the mapping relationship between feature identifiers of key features of equipment status and feature identifiers of key features of production; Replacing the original loss function in the original model with the target loss function to obtain a correlation analysis model; Inputting the training data into the correlation analysis model and outputting a model loss value; Based on the model loss value, a pre-trained correlation analysis model is generated.

6. The method according to claim 5, characterized in that The generating a pre-trained correlation analysis model based on the model loss value includes: When the model loss value reaches a minimum, generating a pre-trained correlation analysis model; Alternatively, when the model loss value has not reached the minimum, the model loss value is back-propagated to update the model parameters of the correlation analysis model, and the step of inputting the training data into the correlation analysis model is continued; The objective loss function is: in, is the loss value, is the total number of samples of training data, is the prediction function of the model, The training data set The first device state feature vector of samples, is a weight matrix used to learn the mapping relationship between the first equipment status feature and the second production feature, is a bias term that provides the model with translational freedom along each dimension in the feature space. is with The corresponding second production eigenvector, is the regularization coefficient, is the L2 norm of the weight matrix, which is used to penalize excessive weight values. is the absolute value of the bias term, is the number of feature identifier pairs in the mapping, is a weight parameter used to balance the penalty term of the mapping probability, For the Mapping probability to feature identifier, It is The logarithm of the feature identity mapping probability is used to encourage the model to learn high-probability mapping relationships.

7. The method according to claim 1, characterized in that The structured data includes a final mapping feature sequence between the equipment status key features and the production key features; The determining the structured data to determine the pain point information existing in the production process of the target park includes: Obtain a pre-established pain point information database, wherein the pain point information database stores a sequence of key-value pairs between mapping features and pain point information, each key-value pair including a key and a value, the key being the mapping feature and the value being the pain point information; Matching each mapping feature in the final mapping feature sequence with the key of each key-value pair to obtain a matching result for each mapping feature; Based on the matching result of each mapping feature, determine a target mapping feature that matches the key; Get the value corresponding to the key matched by each target mapping feature, and obtain the pain point information of each target mapping feature; The pain point information of each target mapping feature is used as the pain point information existing in the production process of the target park.

8. The method according to claim 1, characterized in that The generating of the warning information of the target park includes: Extract key parameters from the pain point information existing in the production process of the target park; The key parameters are input into a preset warning template to generate warning information for the target park.

9. A smart park security device based on unstructured data processing implemented by the method described in any one of claims 1 to 8, characterized in that: The device comprises: An unstructured data processing module is used to collect and pre-process unstructured data fed back by multiple data sources in a target park to obtain structured data of the target park, wherein the multiple data sources include equipment sensors and a supply chain system, and the structured data is generated based on the association relationship between target equipment status information and target production information mined by a pre-trained correlation analysis model, wherein the target equipment status information is obtained by processing the unstructured data fed back by the equipment sensors, and the target production information is obtained by processing the unstructured data fed back by the supply chain system, and the pre-trained correlation analysis model is a mathematical model trained based on a neural network for mining the association relationship between the target equipment status information and the target production information; A pain point information determination module, used to determine the structured data to determine the pain point information existing in the production process of the target park, wherein the pain point information is used to characterize the risk points occurring in the target park; The early warning module is used to generate early warning information of the target park when the number of risk points appearing in the target park is greater than a preset threshold, and send the early warning information to a client for early warning.

Images