An intelligent governance and application closed-loop system and method for internet-of-things detection data based on a knowledge graph

Abstract

The application discloses a kind of based on knowledge graph's internet of things detection data intelligent management and application closed system and method, belong to the cross field of industrial internet, artificial intelligence and detection big data technology, the system includes: internet of things perception layer;Data processing and knowledge conversion layer, including data standardization cleaning module, knowledge graph ontology modeling module and graph construction and real-time updating module;Intelligent analysis and application service layer;Feedback and control layer, including decision instruction generation and issue module, equipment reverse control and process scheduling module.The core is to build a unified, intelligent data platform, the original data generated by the detection equipment (internet of things layer) of physical world is processed through a series of processes, and is finally enabled to diversified business applications (application layer) as structured, computable knowledge (knowledge layer), while the wisdom results of application can guide the behavior of physical equipment in reverse, forming a closed loop.

Description

Technical Field

[0001] This invention belongs to the interdisciplinary field of industrial internet, artificial intelligence and big data testing technology. In particular, it relates to a data lifecycle management system and method that deeply integrates the Internet of Things (IoT), knowledge graph (KG), artificial intelligence (AI) with specific industry business such as testing and certification, equipment monitoring and quality control. In particular, it relates to a closed-loop system and method for intelligent governance and application of IoT testing data based on knowledge graph. Background Technology

[0002] In complex industrial systems such as manufacturing, energy and power, new energy vehicles, and new energy storage, equipment inspection, testing, and quality control are core aspects of ensuring product safety and performance. Currently, the industry commonly faces the following technical challenges: The problem of data silos is severe: test data is scattered across various independent instruments and equipment, manual paper records, and different business systems, lacking a unified format and interface. For example, in the testing of an electric vehicle charging station, the insulation withstand voltage test data, communication protocol consistency test data, and functional test data are stored on different devices, or even in different departments, forming "islands".

[0003] Data and knowledge disconnect: Behind the massive amounts of testing data lies a complex system of professional knowledge (such as standard clauses in GB / T, NB / T, IEC, etc., equipment parameter thresholds, and safety judgment logic). Currently, this knowledge mostly exists in expert experience, standard documents, and offline reports, failing to form a structured connection with real-time / historical data. When an "insulation resistance test fails," engineers need to manually consult standard clauses; there is no automatic push for associated failure modes, possible causes, and handling suggestions.

[0004] Superficial Data Value Mining: A large amount of testing data is used only to generate a single pass / fail report, and its potential value is not fully explored. For example, it is impossible to optimize design parameters based on historical data (such as efficiency and harmonic data of electrochemical energy storage converters under different operating conditions) or to use models to predict potential risks of equipment.

[0005] Application feedback chain breaks: The decision results and optimization suggestions of upper-layer applications (such as intelligent diagnostics, predictive maintenance, and supply chain quality traceability) are difficult to form effective feedback instructions to automatically reach physical equipment and guide subsequent testing operations, equipment parameter adjustments, or maintenance strategy changes, forming an open loop and failing to form a true "data-driven business" closed loop.

[0006] Lagging Standard Update Tracking: Industry standards are updated frequently, with standard number versions spanning several years. Traditional methods rely on manual tracking, which can easily lead to the invalidation of testing methods and judgment criteria, resulting in compliance risks for test results. Summary of the Invention

[0007] To address the shortcomings and deficiencies of existing technologies, this invention proposes a closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs. This invention is a "IoT-Knowledge-Application" closed-loop system. Its core lies in constructing a unified, intelligent data platform. This platform processes the raw data generated by physical detection devices (IoT layer) into structured, computable knowledge (knowledge layer), ultimately empowering diverse business applications (application layer). Simultaneously, the intelligent results of these applications can guide the behavior of physical devices, forming a closed loop.

[0008] Specifically, the technical solution adopted in this invention is as follows: A closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs, comprising: The IoT sensing layer includes a testing instrument networking access module, a product and asset information acquisition module, and a metadata automatic annotation module, which are used for end-to-end digital data collection and access. The data processing and knowledge transformation layer includes a data standardization and cleaning module, a knowledge graph ontology modeling module, and a graph construction and real-time update module. The intelligent analysis and application service layer includes intelligent diagnosis and root cause analysis service modules, predictive maintenance and trend analysis service modules, quality big data analysis and traceability service modules, standard compliance automatic review service modules, intelligent report automatic generation service modules, and a visual interaction platform; The feedback and control layer includes a decision instruction generation and issuance module, and a device reverse control and process scheduling module.

[0009] Furthermore, the IoT sensing layer, data processing and knowledge transformation layer, intelligent analysis and application service layer, and feedback and control layer are sequentially set to form a closed-loop system with a cyclic configuration. Standardized API interfaces are established between each layer, and the API interfaces are based on RESTful API and MQTT protocol for data exchange and control command issuance.

[0010] Furthermore, in the data processing and knowledge transformation layer, the knowledge graph ontology modeling module is used to construct a knowledge graph ontology model, which includes: "Object" entity: Define various types of test objects, which have attributes such as rated power, manufacturer, and model. "Regulation" as its core: Structured analytical testing standards, with judgment conditions within a specific range; "Field" entity: Defines the test scenario, laboratory, field, and environmental conditions such as temperature and humidity; The "Law" itself describes the specific testing methods, testing principles, and instrument operating procedures. "Result" entity: Defines the detection result as "qualified", "unqualified", numerical or waveform file data, or associated abnormal events.

[0011] Furthermore, in the data processing and knowledge transformation layer, the graph construction and real-time update module, based on the cleaned data, uses natural language processing and a rule engine to extract knowledge from semi-structured / unstructured standard files and establish a connection with the real-time collected data; the knowledge graph is updated in real time.

[0012] Furthermore, in the intelligent analysis and application service layer, The intelligent diagnosis and root cause analysis service module: When the test data is marked as "unqualified", the system automatically queries all related standards, historical similar cases, and possible failure mode libraries in the knowledge graph and quickly pushes the possible root causes to the engineer; The predictive maintenance and trend analysis service module performs time-series modeling on massive historical testing data of specific equipment models to predict their performance degradation trends and potential failure points, providing decision support for proactive maintenance. The quality big data analysis and traceability service module: statistically analyzes the quality data of all products, cross-compares them by supplier, model, inspector, batch, etc., to discover systemic quality defects and support supply chain optimization; one-click traceability allows you to trace all historical testing data and correlation graphs of any defective product; The standard compliance automatic review service module: When a new testing task is issued, the system automatically recommends corresponding testing items and parameter configurations to the testing equipment based on the list of applicable standards associated with the tested object in the knowledge graph. This ensures the comprehensiveness and compliance of the testing plan; The intelligent report automatic generation service module: Based on graph-based data and standard templates, and combined with national standard and specification formats, it can automatically generate a well-structured, accurate, and richly illustrated test report, and can also include problem diagnosis suggestions; The visualization and interaction platform provides a unified web interface or mobile app, offering data insights and decision support to different roles such as management, engineers, and field operators through graphical visualization, 3D dynamic dashboards, and multi-dimensional reports.

[0013] A knowledge graph-based intelligent governance and application closed-loop method for IoT detection data, applied to the system described in claim 1, includes the following steps: S1 full-link digital acquisition and access collects physical world data and performs preliminary labeling by combining product information and standard metadata; S2 is a data cleaning and knowledge mapping based on a knowledge graph ontology model. It cleans and transforms the collected data into standardized data, and establishes a mapping relationship between the standardized data and the corresponding standard knowledge nodes according to the knowledge graph ontology. The S3 multimodal knowledge graph construction and fusion continuously structures the knowledge information extracted from standard documents, expert experience, and historical data into standard knowledge, integrates it into the knowledge graph ontology model, and dynamically associates it with real-time streaming data to form an interconnected knowledge network centered on devices, products, and standards. S4 intelligent application incubation and value release, based on knowledge graph, launches intelligent application services; S5 application feedback forms a closed loop, transforming the analysis conclusions obtained in S4 into executable instructions or programs, which in turn act on the detection equipment, production equipment or process flow connected to the IoT sensing layer. S6 performance evaluation and continuous optimization monitor the closed-loop execution effect and re-input the result data into the system to optimize the rules, model parameters and process strategies in the knowledge graph, so that the entire system has the ability to continuously learn and evolve.

[0014] Furthermore, in S2, The data cleaning process includes standardization of the collected data, such as deduplication, noise reduction, and format conversion, to remove invalid data. The knowledge mapping involves establishing associations between cleaned, standardized data and corresponding standard knowledge nodes based on the knowledge graph ontology.

[0015] Furthermore, in S4, based on the knowledge graph, an intelligent application service is launched, which includes: (1) Intelligent diagnosis and root cause analysis: Non-compliant data triggers a graph query, which is associated with the corresponding standard, historical similar cases, and failure mode library, and pushes root cause suggestions; (2) Predictive maintenance and trend analysis: Modeling historical time-series data to predict performance degradation trends and potential failure points; (3) Quality big data analysis and traceability: Cross-analyze the full amount of data by supplier, model, batch and other dimensions to find systemic defects and support the traceability of defective products throughout the entire chain; (4) Automatic standard compliance review: When a new task is assigned, the system automatically matches the standards to be tested and recommends the test items and instrument parameter configurations; (5) Intelligent report automatic generation: Based on the spectral data and national standard templates, a compliant test report containing data, waveforms and diagnostic suggestions can be generated with one click, and the conclusion can be traced back to the original data and standards.

[0016] Furthermore, in S5, the steps for applying feedback to form a closed loop include: Decision instruction generation: Transforms analysis conclusions into executable work orders / control instructions; Device reverse control: Commands are sent to the IoT sensing layer to automatically configure instrument test modes and parameter thresholds; Cross-system collaboration: Push optimization suggestions to external systems.

[0017] A computer-readable storage device, wherein computer program instructions are permanently stored therein, and when the computer program instructions are invoked and loaded by a processor and executed, the method described in any of the preceding claims is performed.

[0018] Compared with the prior art, the present invention has the following beneficial effects: 1. Break down data silos and achieve full-domain integration: By building a unified data model and knowledge graph, scattered testing equipment, standard knowledge, and business data are integrated, solving the problem of information fragmentation.

[0019] 2. Accumulate structured knowledge to empower business applications: Free experts' standard knowledge and experience models from documents and transform them into digital assets that can be calculated, searched, and reasoned, providing a powerful "brain" for upper-level intelligent applications.

[0020] 3. Achieve closed-loop control from detection to optimization: Seamlessly feed back the intelligent analysis results from the application layer to the execution layer, forming a virtuous cycle of "data-driven optimization and data-verified optimization," thereby promoting continuous improvement in product quality, testing efficiency, and operating costs.

[0021] 4. Dynamically adapt to environmental changes and improve compliance: Knowledge graphs can quickly integrate new standards and regulations, automatically update testing processes, ensure real-time compliance of testing activities, and reduce quality risks.

[0022] 5. Significantly improve work efficiency and quality: Automated data collection, report generation, and intelligent diagnosis greatly reduce manual operations and repetitive work, and reduce human error; advanced analysis functions enable the deep release of data value, realizing the transformation from "post-event judgment" to "pre-event prediction and in-event intervention". Attached Figure Description

[0023] Figure 1 This is a schematic diagram of the system in this embodiment; Figure 2 This is a flowchart of the method in this embodiment. Detailed Implementation

[0024] The present invention will now be described in further detail with reference to the accompanying drawings and specific embodiments.

[0025] A closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs, as shown in the attached figure. Figure 1As shown, the system includes: I. IoT Sensing Layer: Testing Instrument Network Access Module: Responsible for connecting various testing instruments, such as intelligent safety comprehensive analyzers, harmonic scintillation analyzers, intelligent on-site verification devices for off-board chargers, and power quality analyzers, through industrial bus and wired / wireless networks for protocol adaptation and data acquisition.

[0026] Product and asset information acquisition module: Automatically acquires the model, batch, serial number, and specification information of the sample to be tested through scanning (RFID, QR code, etc.); acquires environmental context information from the physical environment, such as temperature and humidity sensors.

[0027] Metadata auto-labeling module: During data acquisition, automatically or semi-automatically label the instrument model, channel, test time, and test item of the data source. For example, the standard number referenced for "1.2 / 50us and 8 / 20us surge immunity test" is "GB / T 7251.1-2023 10.12".

[0028] II. Data Processing and Knowledge Transformation Layer: Data standardization and cleaning module: Cleans, transforms, and aligns raw data from different sources and in different formats (analog signals, digital messages, images, videos), and defines a unified data model according to national standards or industry specifications (e.g., detection time, equipment ID, detection items, parameter names, measured values, units, standard limits, conclusions).

[0029] Knowledge Graph Ontology Modeling Module: Establishes a multi-dimensional knowledge graph ontology model of "object-rule-field-law-effect", referred to as the unified model.

[0030] "Object" entity: Defines various types of detection objects, such as "electric vehicle off-board conductive charger", "energy storage converter", "photovoltaic grid-connected inverter", etc. These objects have attributes, such as rated power, manufacturer, and model.

[0031] The core of the "regulation" is a structured analysis of a massive number of testing standards (such as GB / T, NB / T, etc. referenced in the standards). For example, under "Standard GB / T 34120-2023", there is a test item called "protection function", the "technical specification" is "7.4", the judgment condition is a specific range, and the recommended instrument is "energy storage converter test device".

[0032] "Field" entity: Defines the test scenario, laboratory, field, and environmental conditions such as temperature and humidity.

[0033] The "Law" itself describes the specific test methods, test principles, and instrument operating procedures (e.g., the steps of the "surge immunity test").

[0034] "Result" entity: Defines the test result as "qualified" or "unqualified", the data (numerical value, waveform file), and the associated abnormal event ("insulation breakdown").

[0035] The knowledge graph construction and real-time update module: Based on cleaned and standardized data, it uses Natural Language Processing (NLP) and a rule engine to extract knowledge from semi-structured / unstructured standard documents and establish associations with real-time collected data. For example, a specific "grid-connected power factor test" result (attribute: measured value 0.99, within the range of 0.95 or above as defined in standard NB / T 32004-2018 8.3.1.2) will be automatically associated with the specific inverter under test, the corresponding standard clause knowledge node, the instrument used, and the current test environment. The knowledge graph is updated in real time, maintaining a dynamic evolution.

[0036] III. Intelligent Analysis and Application Service Layer: This layer includes various advanced intelligent application service components, specifically: Intelligent Diagnosis and Root Cause Analysis Service Module: When test data is marked as "non-compliant," the system automatically queries the knowledge graph for all related standards, historical similar cases, and possible failure mode libraries, quickly pushing the possible root causes to engineers. For example, if a "resistance" test fails, it can be linked to knowledge nodes such as "historical records of exceeding environmental humidity limits" and "multiple other samples in the same batch failing," and recommends checking "Chapter 12 Insulation Requirements in GB / T 18487.1-2023."

[0037] Predictive Maintenance and Trend Analysis Service Module: Performs time-series modeling on massive historical testing data of specific equipment models (such as charging pile efficiency and key component temperature rise data) to predict performance degradation trends and potential failure points, providing decision support for proactive maintenance.

[0038] Quality Big Data Analysis and Traceability Service Module: Statistically analyzes quality data for all products, cross-compares data by supplier, model, inspector, batch, etc., to identify systemic quality defects and support supply chain optimization. One-click traceability allows you to track all historical testing data and correlation maps of any defective product.

[0039] The standard compliance automatic review service module: When a new testing task is assigned, the system automatically recommends corresponding testing items and parameter configurations to the testing equipment based on the list of applicable standards associated with the tested object in the knowledge graph. This ensures the comprehensiveness and compliance of the testing plan.

[0040] Intelligent Report Automatic Generation Service Module: Based on graph-based data and standard templates, combined with national standard specifications, it automatically generates well-structured, accurate, and richly illustrated test reports, and can also include problem diagnosis suggestions.

[0041] Visualization and Interaction Platform: Provides a unified web interface or mobile app, offering data insights and decision support to different roles such as management, engineers, and field operators through graph visualization, 3D dynamic dashboards, and multi-dimensional reports.

[0042] IV. Feedback and Control Layer (Key to Closed-Loop Formation): Decision command generation and issuance module: Based on the analysis results of the upper layer application (for example, intelligent diagnostics suggests performing "deep BMS communication protocol detection" on a certain type of battery management system, and automatically generating control commands or task work orders).

[0043] Equipment reverse control and process scheduling module: Commands are sent to the IoT sensing layer through a standard interface, which can automatically configure and adjust the test modes and parameter thresholds of relevant detection instruments; when the system detects a suspected harmonic problem, it can automatically send a command to require the power quality analyzer to perform high-precision data sampling for a longer period of time, and link the oscilloscope to capture transient waveforms.

[0044] This embodiment establishes standardized API interfaces between each layer, for example, data exchange and control command issuance based on RESTful API and MQTT protocol.

[0045] Example: Implementation of a closed-loop "IoT-Knowledge-Application" system for third-party testing laboratories of electric vehicle charging stations Assume this system is deployed in a third-party laboratory that provides certification and testing services for charging pile manufacturers.

[0046] Physical layer access: The laboratory's equipment, such as the "Intelligent On-site Verification Device for Off-board Chargers," "Intelligent Safety Regulation Comprehensive Analyzer," "Power Quality Analyzer," and "Harmonic Scintillation Analyzer," are connected to the system via standard protocols such as Industrial Ethernet or OPC UA. When each device comes online, its model, testing capabilities, accuracy, and other information are registered in the system as a "Testing Instrument Node."

[0047] Task Trigger: The laboratory business system receives a new order for a certain model of BYD DC charging pile (including model "BYD-DC1000W" and applicable standards "GB / T 18487.5, GB / T 34657.1-2025"). The system automatically parses the data using a knowledge graph to generate a detailed "testing workflow" for the order, which includes a series of test items (such as "vehicle interface locking function test", "communication protocol conformance test", "EMC conducted emission test", etc.).

[0048] Intelligent execution: Task assignment: Based on the real-time status of the equipment (idle / busy), geographical location, and capability matching, the system optimally assigns the "vehicle interface locking function test" to the off-board charger intelligent on-site verification device of test stand No. 1.

[0049] One-click configuration: Engineers swipe their cards to log in to Test Station No. 1. The system interface (large screen) automatically pushes the complete standard requirements for this test (referencing relevant GB / T clauses), visual operation guidance, and has completed the automatic pre-configuration of instrument parameters (such as output voltage, current curve, and detection logic).

[0050] Automatic data acquisition and correlation: As engineers begin testing, all sensor data (voltage, current, temperature, locking mechanism status signals, etc.) are synchronized to the system in real time. Each data point is automatically tagged with "Test Item ID", "Standard Number and Clause Number" (e.g., linked to the "Vehicle Plug Locking Function Test" node in the knowledge graph, which references the "NB / T 33001-2018 6.5" standard in historical appendix information), "Device ID", and "Specific Instrument ID Used".

[0051] Real-time monitoring and intelligent analysis: During the testing process, the system compares the "measured locking force curve" with the "standard tolerance range curve stored in the graph" in real time. Once abnormal fluctuations are detected (such as failure to lock completely at the specified distance), an alert is immediately issued on the interface, and the graph visualization prompts possible related "mechanical structure" knowledge nodes or similar historical failure cases.

[0052] Closed-loop feedback and dynamic optimization: A systemic risk was identified: The "electrical clearance and creepage distance test" of a certain batch of charging piles showed critical performance across multiple suppliers. Knowledge graph analysis revealed that the problems all stemmed from the PCB layout design stage, specifically related to clause "GB / T 42008-2022 6.4.3" in the "Standard Body". This knowledge was then pushed to the application layer.

[0053] Application-layer intelligent service output: The quality big data analysis and traceability service module generates a "Structural Layout Design Optimization" report and provides suggestions for subsequent design improvements through the "Predictive Maintenance and Trend Analysis Service." Simultaneously, this analysis result is input into the "Intelligent Report Automatic Generation Service" to generate in-depth analysis chapters for the final report.

[0054] Closed-loop optimized testing process: For this high-risk project, the system automatically optimizes future testing strategies based on feedback: increasing the sampling rate, increasing the test voltage level, or referencing more "similar historical precedents" for this test on similar products. It can even provide suggestions to the PLM (Product Lifecycle Management) systems of collaborating design units / front-end customers to achieve design collaboration.

[0055] Reporting and Traceability: After testing, the system automatically fills in the data, images, and anomaly logs into the nationally standardized report template based on the graph, generating an official certification report with one click. Each pass / fail conclusion in the report allows users to trace back to the original data, judgment criteria, and real-time graph.

[0056] Knowledge Graph Evolution: All the data, method effectiveness analysis, and notes added by experts when handling warnings in this test (e.g., "It was found to be a common phenomenon caused by the tolerance of a certain batch of connectors, which should be marked in the test report") were also systematically learned and abstracted, transformed into "tacit knowledge", and enriched into the knowledge graph, forming a self-learning and improvement capability of "coming from practice and going back to practice".

[0057] This invention constructs a dynamic, self-optimizing, and tightly connected "IoT-knowledge-application" full-link closed loop, effectively solving the serious problem of information disconnection in various links of the traditional system, and providing a core solution for the digital upgrade of the testing field and even the broader manufacturing industry.

[0058] A closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs, see [link to relevant documentation]. Figure 2 The diagram shows the overall process. This process is a closed-loop, iterative structure, with six key stages. Each stage uses standardized interfaces such as RESTful API and MQTT to achieve data interaction and command flow. The specific steps are as follows: S1: End-to-end digital acquisition and access Collect physical world data and perform preliminary annotation by combining product information and standard metadata. Specifically, this includes: S101 IoT device access: Testing instruments (such as intelligent safety comprehensive analyzers, power quality analyzers, etc.) access the system through industrial Ethernet / OPC UA protocol, complete the registration of information such as instrument model, testing capabilities, and accuracy, and form "testing instrument nodes".

[0059] S102 Multi-Source Data Acquisition and Labeling: Real-time acquisition of physical layer detection data (voltage, current, temperature, mechanical status signals, etc.), combined with the information of the product under test (such as the charging pile model BYD-DC1000W), the entrusted task information, and standard metadata (such as GB / T 18487.5) to complete the initial labeling.

[0060] S2: Data Cleaning and Knowledge Mapping Based on a Unified Model The aforementioned "unified model" refers to a knowledge graph ontology model centered on "object-rule-field-method-effect," which is a standardized knowledge framework supporting the closed loop of intelligent governance and application of IoT detection data. Specific details are as follows: Core constituent dimensions: “Object”: refers to physical objects such as the tested equipment entity and the testing instrument node; "Regulations": refers to standard knowledge such as various standard clauses, testing items, and threshold ranges; “Field”: refers to scenario information such as the testing environment; "Law": refers to procedural knowledge such as experimental steps and testing methods; "Result": This refers to result-type information such as test data, fault cases, and diagnostic conclusions.

[0061] The core function of the unified model is to guide the standardized cleaning of collected data, completing processes such as deduplication, noise reduction, and format conversion, and eliminating invalid data. It achieves precise mapping between cleaned data and standard knowledge nodes, establishing associations between scattered test data and corresponding test item nodes, standard clauses, tested device IDs, instrument IDs, etc., transforming the data into structured knowledge-related data. The unified model acts as a core hub connecting the end-to-end digital data acquisition and access with the construction and integration of multimodal knowledge graphs. It provides a unified knowledge foundation for the dynamic updating of knowledge graphs and the incubation of intelligent applications (such as intelligent diagnosis and automatic report generation), ensuring the collaborative flow of data and knowledge throughout the entire process.

[0062] The collected data is cleaned and transformed, and based on the knowledge graph ontology, a mapping relationship is established between it and the corresponding standard knowledge nodes (detection items, thresholds, methods, etc.). This includes: S201 Data Cleaning: Standardize the collected data by performing deduplication, noise reduction, and format conversion, and remove invalid data.

[0063] S202 Ontology Mapping: Based on the core knowledge graph ontology of "object-regulation-field-law-effect", the cleaned data is associated with the corresponding standard knowledge nodes. For example, the "locking force measured curve" is mapped to the "vehicle interface locking function test" node, and associated with the corresponding GB / T clauses, the ID of the device under test, and the instrument ID.

[0064] S3: Multimodal knowledge graph construction and fusion. This involves continuously structuring knowledge extracted from standard documents, expert experience, and historical data, integrating it into a unified knowledge graph, and dynamically associating it with real-time streaming data to form an interconnected knowledge network centered on devices, products, and standards. This includes: S301 Knowledge Extraction and Fusion: Extract structured knowledge (such as test procedures, threshold ranges, and failure modes) from standard documents, expert experience, and historical failure cases, and integrate it into a unified knowledge graph.

[0065] S302 Dynamic Association Update: Establishes a dynamic association between real-time detection data and the tested device entity, standard clauses, test environment, and instrument nodes in the spectrum; synchronously updates the spectrum content, such as associating "measured grid-connected power factor value 0.99" with clause 8.3.1.2 of NB / T 32004-2018 and the tested inverter entity.

[0066] S4: Intelligent Application Incubation and Value Release. With the support of knowledge graphs, various intelligent application services are launched, such as diagnosis, prediction, and analysis. For example, if the statistical average value of a key parameter (such as "power quality adaptability") of a batch of equipment shows a slow trend deviation, the system will automatically trigger an early warning and associate it with information on upstream and downstream processes or components that may affect the parameter.

[0067] Five types of intelligent services are triggered based on knowledge graphs: (1) Intelligent diagnosis and root cause analysis: Non-compliant data triggers spectrum query, associates with corresponding standards, historical similar cases, failure mode library, and pushes root cause suggestions (such as non-compliant insulation resistance as associated with excessive ambient humidity, cases in the same batch and Chapter 12 of GB / T18487.1-2023).

[0068] (2) Predictive maintenance and trend analysis: Model historical time series data (such as charging pile temperature rise and efficiency) to predict performance degradation trends and potential failure points.

[0069] (3) Quality big data analysis and traceability: Cross-analyze the full data by supplier, model, batch and other dimensions to find systemic defects (such as electrical clearance criticality caused by PCB layout) and support full-chain traceability of defective products.

[0070] (4) Automatic standard compliance review: When a new task is assigned, the standard to be tested is automatically matched and the test items and instrument parameter configuration are recommended.

[0071] (5) Intelligent report automatic generation: Based on the spectral data and national standard templates, a compliant test report containing data, waveforms and diagnostic suggestions can be generated with one click, and the conclusion can be traced back to the original data and standards.

[0072] S5: Applying intelligence to provide feedback to the physical world, forming a closed loop. The analytical conclusions obtained in S4 (such as early warnings, diagnoses, and optimization suggestions) are transformed into executable instructions or programs, which then act in reverse on the detection equipment, production equipment, or processes at the IoT layer. For example, if the system identifies that the contact resistance of a charging gun is too high at a specific temperature, it automatically updates the final testing criteria for that batch and recommends more stringent re-inspection items to the on-site calibrator, such as "insertion and extraction force / mechanical strength test".

[0073] S501 Decision Instruction Generation: Transforms analysis conclusions into executable work orders / control instructions (e.g., generating a "Plug-in / Pull-out Force / Mechanical Strength Test" re-inspection work order for high contact resistance).

[0074] S502 device reverse control: commands are sent to the IoT sensing layer to automatically configure instrument test modes and parameter thresholds (such as triggering power quality analyzer to perform long-term high-precision sampling for harmonic problems + oscilloscope to capture transient waveforms).

[0075] S503 cross-system collaboration: pushes optimization suggestions (such as PCB layout optimization solutions) to external systems (such as customer PLM systems).

[0076] S6: Performance Evaluation and Continuous Optimization (Closed-Loop Iteration) Monitor the closed-loop execution effect (such as re-inspection results, false alarm rate, and efficiency improvement after parameter optimization), and input the results data back into the system to optimize the rules, model parameters, and process strategies in the knowledge graph, so that the entire system has the ability to continuously learn and evolve.

[0077] S601 Performance Monitoring: Tracks data such as command execution results (e.g., re-inspection pass rate, efficiency improvement after parameter optimization), and false alarm rate.

[0078] S602 Knowledge Graph Evolution: It transforms monitoring data, expert processing notes, and practical experience into structured knowledge to enrich the graph, while optimizing detection strategies (such as increasing the sampling ratio of high-risk projects) to achieve system self-learning and improvement.

[0079] The S603 process loop logic is as follows: data acquisition → knowledge modeling → intelligent application → feedback control → knowledge evolution → a new round of data acquisition, forming a virtuous cycle of "data-driven optimization and optimization verified by data".

[0080] The serial numbers in the above embodiments are for descriptive purposes only and do not represent the chronological order of assembly or use of the components. Those skilled in the art will understand that the embodiments of this application can be provided as methods, systems, or computer program products. Therefore, this application can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application can take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0081] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0082] Unless otherwise specifically stated, the relative steps, numerical expressions, and values ​​of the components and steps described in these embodiments do not limit the scope of the invention.

[0083] This application also provides an electronic device, which may include at least one processor and at least one memory. The processor may include one or more processing cores. The processor connects to various parts of the server using various interfaces and lines, and performs various functions of the server and processes data by running or executing instructions, programs, code sets, or instruction sets stored in the memory, and by calling data stored in the memory.

[0084] This application also provides a computer-readable storage medium storing instructions. When executed by one or more processors, these instructions cause an electronic device to perform one or more of the methods described in the above embodiments.

[0085] Furthermore, the functional units in this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0086] The foregoing description is merely an exemplary embodiment of this disclosure and should not be construed as limiting the scope of this disclosure. Any equivalent changes and modifications made in accordance with the teachings of this disclosure shall still fall within the scope of this disclosure. Those skilled in the art will readily conceive of other embodiments of this disclosure upon considering the specification and the disclosure of practical truth. This application is intended to cover any variations, uses, or adaptations of this disclosure that follow the general principles of this disclosure and include common knowledge or customary techniques in the art not described in this disclosure. The specification and embodiments are considered exemplary only, and the scope and spirit of this disclosure are defined by the claims.

Claims

1. A closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs, characterized in that, include: The IoT sensing layer includes a testing instrument networking access module, a product and asset information acquisition module, and a metadata automatic annotation module, which are used for end-to-end digital data collection and access. The data processing and knowledge transformation layer includes a data standardization and cleaning module, a knowledge graph ontology modeling module, and a graph construction and real-time update module. The intelligent analysis and application service layer includes intelligent diagnosis and root cause analysis service modules, predictive maintenance and trend analysis service modules, quality big data analysis and traceability service modules, standard compliance automatic review service modules, intelligent report automatic generation service modules, and a visual interaction platform; The feedback and control layer includes a decision instruction generation and issuance module, and a device reverse control and process scheduling module.

2. The closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs as described in claim 1, characterized in that, The IoT sensing layer, data processing and knowledge transformation layer, intelligent analysis and application service layer, and feedback and control layer are sequentially set to form a closed-loop system. Standardized API interfaces are established between each layer. The API interfaces are based on RESTful API and MQTT protocol for data exchange and control command issuance.

3. The closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 1, characterized in that, In the data processing and knowledge transformation layer, the knowledge graph ontology modeling module is used to construct a knowledge graph ontology model, which includes: "Object" entity: Define various types of detection objects, which have attributes such as rated power, manufacturer, and model; "Regulation" ontology: Structured analytical testing standards, with judgment conditions within a specific range; "Field" entity: Defines the test scenario, laboratory, field, and environmental conditions such as temperature and humidity; The "Law" itself describes the specific testing methods, testing principles, and instrument operating procedures; "Result" entity: Defines the detection result as "qualified", "unqualified", numerical or waveform file data, or associated abnormal events.

4. The closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 1, characterized in that, In the data processing and knowledge transformation layer, the knowledge graph construction and real-time update module, based on the cleaned data, uses natural language processing and a rule engine to extract knowledge from semi-structured / unstructured standard files and establish a connection with the real-time collected data; the knowledge graph is updated in real time.

5. The closed-loop system for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 1, characterized in that, In the intelligent analysis and application service layer, The intelligent diagnosis and root cause analysis service module: When the test data is marked as "unqualified", the system automatically queries all related standards, historical similar cases, and possible failure mode libraries in the knowledge graph and quickly pushes the possible root causes to the engineer; The predictive maintenance and trend analysis service module performs time-series modeling on massive historical testing data of specific equipment models to predict their performance degradation trends and potential failure points, providing decision support for proactive maintenance. The quality big data analysis and traceability service module: statistically analyzes the quality data of all products, cross-compares them by supplier, model, inspector, batch, etc., to discover systemic quality defects and support supply chain optimization; one-click traceability allows you to trace all historical testing data and correlation graphs of any defective product; The standard compliance automatic review service module: When a new testing task is issued, the system automatically recommends corresponding testing items and parameter configurations to the testing equipment based on the list of applicable standards associated with the tested object in the knowledge graph. This ensures the comprehensiveness and compliance of the testing plan; The intelligent report automatic generation service module: Based on graph-based data and standard templates, and combined with national standard and specification formats, it can automatically generate a well-structured, accurate, and richly illustrated test report, and can also include problem diagnosis suggestions; The visualization and interaction platform provides a unified web interface or mobile app, offering data insights and decision support to different roles such as management, engineers, and field operators through graphical visualization, 3D dynamic dashboards, and multi-dimensional reports.

6. A closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs, characterized in that, Applied to the system of claim 1, comprising the steps of: S1 full-link digital acquisition and access collects physical world data and performs preliminary labeling by combining product information and standard metadata; S2 is a data cleaning and knowledge mapping based on a knowledge graph ontology model. It cleans and transforms the collected data into standardized data, and establishes a mapping relationship between the standardized data and the corresponding standard knowledge nodes according to the knowledge graph ontology. The S3 multimodal knowledge graph construction and fusion continuously structures the knowledge information extracted from standard documents, expert experience, and historical data into standard knowledge, integrates it into the knowledge graph ontology model, and dynamically associates it with real-time streaming data to form an interconnected knowledge network centered on devices, products, and standards. S4 intelligent application incubation and value release, based on knowledge graph, launches intelligent application services; S5 application feedback forms a closed loop, transforming the analysis conclusions obtained in S4 into executable instructions or programs, which in turn act on the detection equipment, production equipment or process flow connected to the IoT sensing layer. S6 performance evaluation and continuous optimization monitor the closed-loop execution effect and re-input the result data into the system to optimize the rules, model parameters and process strategies in the knowledge graph, so that the entire system has the ability to continuously learn and evolve.

7. The closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 6, characterized in that, In S2, The data cleaning process includes standardization of the collected data, such as deduplication, noise reduction, and format conversion, to remove invalid data. The knowledge mapping involves establishing associations between cleaned, standardized data and corresponding standard knowledge nodes based on the knowledge graph ontology.

8. The closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 6, characterized in that, In S4, based on the knowledge graph, an intelligent application service is launched, which includes: (1) Intelligent diagnosis and root cause analysis: Non-compliant data triggers a graph query, which is associated with the corresponding standard, historical similar cases, and failure mode library, and pushes root cause suggestions; (2) Predictive maintenance and trend analysis: Modeling historical time-series data to predict performance degradation trends and potential failure points; (3) Quality big data analysis and traceability: Cross-analyze the full amount of data by supplier, model, batch and other dimensions to find systemic defects and support the traceability of defective products throughout the entire chain; (4) Automatic standard compliance review: When a new task is assigned, the system automatically matches the standards to be tested and recommends the test items and instrument parameter configurations; (5) Intelligent report automatic generation: Based on the spectral data and national standard templates, a compliant test report containing data, waveforms and diagnostic suggestions can be generated with one click, and the conclusion can be traced back to the original data and standards.

9. The closed-loop method for intelligent governance and application of IoT detection data based on knowledge graphs according to claim 6, characterized in that, In S5, the steps to apply feedback and form a closed loop include: Decision instruction generation: Transforms analysis conclusions into executable work orders / control instructions; Device reverse control: Commands are sent to the IoT sensing layer to automatically configure instrument test modes and parameter thresholds; Cross-system collaboration: Push optimization suggestions to external systems.

10. A computer-readable storage device, characterized in that, The storage device contains computer program instructions, which are invoked and loaded by the processor and executed to perform the method as described in any one of claims 6 to 9.

Images