Real-time monitoring and data analysis method and system for multi-source information fusion

Abstract

The invention belongs to the technical field of equipment early warning, and particularly relates to a multi-source information fusion real-time monitoring and data analysis method and system, and the method comprises the steps: achieving the automatic recognition and data collection, remote assistance and support, and intelligent path planning and navigation functions in the inspection process of a chemical device, so as to standardize the inspection process; the inspection efficiency and the result accuracy are improved; establishing a chemical device image recognition model, collecting and preprocessing chemical device image data, performing model training by using a convolutional neural network deep learning model, and deploying the model to actual equipment; an abnormal feature extraction algorithm is developed, the operation state of the equipment is monitored in real time in combination with an intelligent early warning and processing system, early warning is quickly given out, and corresponding processing measures are taken; by adopting the anomaly detection and classification technology, an anomaly detection algorithm is realized, an efficient alarm information transmission mechanism and an intelligent alarm information processing system are developed, and an alarm information feedback and closed-loop management mechanism is established, so that the accuracy and timeliness of anomaly detection are improved.

Description

Technical Field

[0001] This invention belongs to, but is not limited to, the field of equipment early warning technology, and particularly relates to a real-time monitoring and data analysis method and system for multi-source information fusion. Background Technology

[0002] The current status and trends of intelligent interactive chemical process key equipment status early warning analysis and decision-making systems at home and abroad show the following characteristics: Data-driven intelligent equipment fault warning and health management systems: These systems are mainly applied to critical equipment such as high-speed rail, aircraft, turbines, and wind turbines, involving multiple aspects such as system modeling and simulation, condition monitoring and fault warning, and reliability control. Currently, most domestic systems are still in the traditional time-domain and frequency-domain analysis stage, while large foreign enterprises have begun to implement predictive maintenance. The global predictive maintenance solutions market reached 25.2 billion yuan in 2019 and is expected to reach 92.9 billion yuan by 2026.

[0003] Applications of Artificial Intelligence in the Chemical Industry: The application of artificial intelligence technology in the chemical industry mainly focuses on materials research and synthesis, intelligent production scheduling and optimization, quality control and testing, predictive maintenance and equipment health management, and safety risk assessment and early warning. These applications help improve production efficiency, reduce costs, and enhance safety.

[0004] Deep Learning-Based Fault Detection and Diagnosis in Chemical Processes: Deep learning technology has been widely applied in fault detection and diagnosis in chemical processes, including methods based on autoencoders, deep belief networks, convolutional neural networks, and recurrent neural networks. These methods have attracted considerable attention in both academia and industry, but still face challenges such as issues related to data, models, and visualization.

[0005] Application of Data-Driven Approaches in Fault Diagnosis of Chemical Processes: As a black-box model, data-driven approaches have shown significant advantages in fault diagnosis of chemical processes. Currently, deep learning and ensemble learning are the research focus of data-driven approaches. Domestic and international research indicates that combining multiple data-driven methods can effectively solve problems in chemical processes.

[0006] Artificial intelligence-based chemical process control technologies: The application of artificial intelligence in chemical process control includes expert systems, machine learning, natural language processing, computer vision, and intelligent control. These technologies contribute to the automation, intelligentization, and optimized control of chemical production processes.

[0007] In summary, intelligent interactive systems for early warning analysis and decision-making regarding the status of key equipment in chemical processes are rapidly developing both domestically and internationally, particularly in the application of data-driven and artificial intelligence technologies. In the future, these technologies will be more deeply integrated into all aspects of the chemical industry, improving production efficiency and safety performance.

[0008] Because chemical production enterprises are often characterized by flammability, explosiveness, high temperature, and high pressure, and some production processes involve highly corrosive and toxic substances, higher demands are placed on the safe operation and emergency response of chemical production. Since the beginning of the 21st century, the overall incidence of natural disasters and public accidents in China has shown a downward trend, but the accident control situation in the chemical industry remains severe. Insufficient safety supervision and outdated production equipment seriously affect the safe production within chemical enterprises, leading to frequent major and serious chemical (hazardous materials) safety accidents.

[0009] Plant inspection is a crucial aspect of chemical production safety assurance, aiming to ensure the safe operation of chemical production facilities and the stability of production quality. During inspections, inspectors, following a pre-established plan, arrive at designated locations to collect data, identify risks, and investigate potential hazards. This effectively ensures that risks and abnormal data related to safe production and equipment operation are strictly controlled. This work involves a comprehensive inspection and monitoring of production equipment, processes, and environmental conditions. The inspection process includes visual inspection and functional testing of equipment to ensure its integrity; inspection of safety devices to ensure safety during production; monitoring process parameters to ensure a safe and stable production process; and monitoring environmental indicators to ensure environmental safety at the production site. Furthermore, chemical plant inspections also include the calibration and maintenance of instruments and meters, as well as regular equipment upkeep. Timely recording and reporting of problems and anomalies facilitates timely handling and improvement measures. The rigor and timeliness of chemical plant inspections effectively ensure that risks and abnormal data related to safe production and equipment operation are strictly controlled.

[0010] Chemical plants typically operate under harsh conditions such as high temperature and pressure, making them prone to accidents such as explosions and leaks. Therefore, regular inspections are crucial. Chemical plants often involve hazardous chemicals, such as corrosive, flammable, and explosive substances. Leaks or accidents can pollute the environment and threaten the health of workers. Inspections can promptly detect problems such as pipeline leaks and valve malfunctions, preventing accidents and ensuring the safety of workers and the environment. Furthermore, chemical plants usually consist of various complex pieces of equipment, and the operating status of this equipment directly affects product quality and production efficiency. Regular inspections can promptly identify equipment problems, allowing for repairs and replacements, ensuring normal production and improving efficiency and product quality. Therefore, chemical companies should attach great importance to plant inspections to ensure the safe and stable operation of the plant.

[0011] Currently, the commonly used RFID inspection method uses RFID technology to identify and mark equipment, facilities, or areas. Inspection personnel carry handheld devices equipped with RFID readers. As they move near the equipment or area to be inspected, the reader scans and reads information from the RFID tags, such as the equipment number and location. The status of the equipment or area is then confirmed by comparing the inspector's observation with pre-set information, and the inspection results are recorded. However, this method suffers from low efficiency, poor standardization, and difficulties in supervision and data processing, making it highly susceptible to false positives, missed inspections, and even data fabrication by inspection personnel—significant safety hazards.

[0012] Based on the above analysis, the urgent technical problems that need to be solved by the existing technology are: the currently widely used RFID inspection methods have problems such as low efficiency and poor standardization, and the existing methods have drawbacks such as difficulty in supervision and low data processing efficiency, which can easily lead to major security risks such as false detection, missed detection, and even inspection personnel fabricating data. Summary of the Invention

[0013] To address the problems existing in the prior art, this invention provides a method and system for real-time monitoring and data analysis based on multi-source information fusion.

[0014] This invention is implemented as follows: a real-time monitoring and data analysis method based on multi-source information fusion, comprising: S1. Chemical Plant Inspection Path Planning: During the inspection of chemical plants, automatic identification and data collection, remote assistance and support, and intelligent path planning and navigation functions are realized to standardize the inspection process and improve inspection efficiency and accuracy. S2. Image Recognition of Chemical Plants: Establish an image recognition model for chemical plants, collect and preprocess image data of chemical plants, train the model using a convolutional neural network (CNN) deep learning model, and deploy it to actual equipment; S3. Abnormal operating condition identification: Develop an abnormal feature extraction algorithm, combine it with an intelligent early warning and processing system to monitor the equipment operating status in real time, quickly issue early warnings and take corresponding measures; S4. Alarm Management: Employ anomaly detection and classification technologies to implement anomaly detection algorithms, develop an efficient alarm information transmission mechanism and an intelligent alarm information processing system, and establish an alarm information feedback and closed-loop management mechanism to improve the accuracy and timeliness of anomaly detection.

[0015] Furthermore, S1 specifically includes: establishing an inspection environment equipment model, using a hierarchical algorithm framework for path planning, and developing a dynamic path optimization module to cope with real-time changes in equipment status and emergencies.

[0016] Furthermore, S2 also includes: developing a real-time image preprocessing platform for image recognition under special circumstances to improve the accuracy of image recognition.

[0017] Furthermore, S3 also includes: establishing an intelligent early warning and processing mechanism through deep learning and real-time data analysis technology to ensure timely response and effective handling of abnormal operating conditions.

[0018] Another objective of this invention is to provide a real-time monitoring and data analysis system for multi-source information fusion that realizes the aforementioned real-time monitoring and data analysis method for multi-source information fusion, comprising: Inspection route dynamic planning module: Automatically plans inspection routes based on inspection tasks and optimizes the inspection routes with the goal of minimizing the distance. Deep learning-based image recognition module: Uses image recognition algorithms to accurately identify the reading status of chemical instruments; Anomaly detection and classification module: Employs anomaly feature extraction algorithms to provide accurate data support for intelligent early warning and processing systems; Alarm tracing module: An alarm tracing system is adopted, including an alarm information transmission and processing system as well as an alarm information feedback and closed-loop management mechanism; Defect Library Module: Establish a defect library to guide the judgment of inspection results during the inspection process, including a defect standard library and a defect knowledge base; Inspection Management Module: The inspection management software consists of three software components: the inspection management system, the inspection system, and the inspection APP.

[0019] Furthermore, the inspection management system is deployed on a server and adopts a B / S architecture, allowing users to access the management system via a web page. The system implements basic database maintenance, equipment inspection point management, inspection task management, inspection result management, and inspection traceability functions.

[0020] Another object of the present invention is to provide a computer device, the computer device including a memory and a processor, the memory storing a computer program, and when the computer program is executed by the processor, causing the processor to perform the steps of the multi-source information fusion real-time monitoring and data analysis method.

[0021] Another object of the present invention is to provide a computer-readable storage medium storing a computer program, which, when executed by a processor, causes the processor to perform the steps of the multi-source information fusion real-time monitoring and data analysis method.

[0022] Another objective of this invention is to provide an information data processing terminal, which includes the aforementioned real-time monitoring and data analysis system for multi-source information fusion.

[0023] Based on the above technical solutions and the technical problems solved, the advantages and positive effects of the technical solution to be protected by this invention are as follows: 1. Scientific value: The project adopts advanced technologies such as mixed reality, which promotes technological innovation in the field of chemical plant inspection and introduces innovative thinking and methods for industrial production and equipment maintenance.

[0024] By combining advanced technologies such as image recognition and deep learning, the automatic identification of the reading status of chemical instruments has been realized, improving data utilization efficiency and intelligence.

[0025] 2. Social benefits: It enables real-time monitoring of equipment status and accurate identification of abnormal operating conditions, effectively preventing and reducing the probability of accidents in chemical production processes and ensuring production safety.

[0026] The use of automated inspection and data acquisition systems has reduced the workload of inspection personnel, lowered inspection costs, and improved work efficiency.

[0027] It has promoted the innovation and transformation of chemical plant inspection methods, improved production efficiency and product quality, and facilitated the innovation and progress of the chemical industry.

[0028] 3. Economic benefits: The implementation of automated inspection systems reduces labor costs and inspection cycles, saving companies on inspection and maintenance expenses.

[0029] By optimizing inspection paths and implementing automated inspections, the utilization efficiency and production efficiency of production equipment have been improved, production downtime has been reduced, and production revenue has been increased.

[0030] Technological innovation and application have promoted the improvement of chemical production technology and industrial upgrading, creating more economic value and employment opportunities.

[0031] 4. Ecological benefits: Optimized inspection routes and reduced manpower input reduced energy and material consumption and environmental pollution.

[0032] It has reduced safety accidents and environmental pollution incidents in the chemical production process, and protected the stability and sustainable development of the ecological environment.

[0033] By improving production safety and product quality, the company has fulfilled its social responsibility and made a positive contribution to social harmony and stability. Attached Figure Description

[0034] Figure 1 This is a flowchart of the real-time monitoring and data analysis method for multi-source information fusion provided in this embodiment of the invention; Figure 2 This is a structural diagram of a real-time monitoring and data analysis system for multi-source information fusion provided in an embodiment of the present invention. Detailed Implementation

[0035] To make the objectives, technical solutions, and advantages of this invention clearer, the invention will be further described in detail below with reference to embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the invention.

[0036] like Figure 1 As shown in the embodiment of the present invention, the real-time monitoring and data analysis method for multi-source information fusion includes: S1. Chemical Plant Inspection Path Planning: During the inspection of chemical plants, automatic identification and data collection, remote assistance and support, and intelligent path planning and navigation functions are realized to standardize the inspection process and improve inspection efficiency and accuracy. S2. Image recognition of chemical plants: Establish an image recognition model for chemical plants, collect and preprocess image data of chemical plants, train the model using a convolutional neural network deep learning model, and deploy it to actual equipment; S3. Abnormal operating condition identification: Develop an abnormal feature extraction algorithm, combine it with an intelligent early warning and processing system to monitor the equipment operating status in real time, quickly issue early warnings and take corresponding measures; S4. Alarm Management: Employ anomaly detection and classification technologies to implement anomaly detection algorithms, develop an efficient alarm information transmission mechanism and an intelligent alarm information processing system, and establish an alarm information feedback and closed-loop management mechanism to improve the accuracy and timeliness of anomaly detection.

[0037] S1 specifically includes: establishing an inspection environment equipment model, using a hierarchical algorithm framework for path planning, and developing a dynamic path optimization module to cope with real-time changes in equipment status and emergencies.

[0038] S2 also includes: developing a real-time image preprocessing platform for image recognition under special circumstances to improve image recognition accuracy.

[0039] S3 also includes: establishing an intelligent early warning and processing mechanism through deep learning and real-time data analysis technology to ensure timely response and effective handling of abnormal operating conditions.

[0040] This invention provides a real-time monitoring and data analysis method based on multi-source information fusion, which is applicable to the inspection and operation status monitoring scenarios of chemical plants. By coordinating the processing of inspection paths, image information, equipment status data and abnormal operating conditions, it enables real-time perception, anomaly identification and closed-loop management of the operation status of chemical plants.

[0041] In step S1, the chemical inspection path planning specifically includes: first, establishing a digital model of the inspection environment and chemical equipment, modeling the distribution of equipment, channel structure, hazardous areas, and equipment operating levels within the inspection area; based on this, a hierarchical path planning algorithm is adopted to divide the inspection task into a global planning layer and a local optimization layer. The global planning layer is used to generate a basic inspection path that meets the inspection coverage and safety constraints, while the local optimization layer is used to dynamically adjust the predetermined inspection path by combining real-time collected equipment operating status, environmental change information, and emergencies, thereby realizing intelligent optimization and navigation control of the inspection path.

[0042] In step S2, the image recognition of the chemical plant specifically includes: constructing an image recognition model for the chemical plant; collecting image data of the plant, including instrument readings, valve status, equipment appearance, and warning signs; and performing preprocessing operations such as denoising, distortion correction, and brightness normalization on the image data; subsequently, constructing a deep learning model based on a convolutional neural network, training the model using labeled samples, and deploying the trained model to an inspection terminal or edge computing device to achieve real-time recognition and status determination of images collected during the inspection process. For special situations such as insufficient lighting or occlusion, a real-time image preprocessing module is also provided to improve the stability and accuracy of image recognition.

[0043] In step S3, abnormal operating condition identification specifically includes: extracting abnormal features based on multi-source state information, combining equipment operating parameters, image recognition results and historical operating data, constructing an abnormal judgment model through deep learning model and real-time data analysis technology, and triggering an abnormal warning when a significant difference is detected between the current operating state and the historical normal state, and generating corresponding handling suggestions or control instructions to achieve rapid response to abnormal operating conditions.

[0044] In step S4, alarm management specifically includes: classifying and grading the detected abnormal information, sending the alarm information to the management terminal through an efficient information transmission mechanism, and recording the abnormal handling process and results; on this basis, by retrospectively analyzing the alarm feedback information, dynamically updating the abnormal judgment rules and threshold parameters, thereby forming a closed-loop management mechanism for alarm information and improving the accuracy and timeliness of abnormal detection and handling.

[0045] This invention provides a real-time monitoring and data analysis method based on multi-source information fusion, which is applied to the inspection scenario of chemical plants. Through the coordinated operation of path planning algorithm, deep learning image recognition algorithm and anomaly detection algorithm, it realizes real-time monitoring and anomaly management of the operating status of chemical plants.

[0046] In step S1, the chemical plant inspection path planning specifically includes: First, establishing an inspection environment topology model based on the spatial structure of the chemical plant, constructing a weighted directed graph of plant nodes, passage nodes, and hazardous area nodes, where edge weights comprehensively consider travel distance, safety level, and inspection priority; based on this, an improved A* path planning algorithm is used to generate an initial inspection path, with equipment risk weights introduced into the heuristic function to avoid invalid detours to high-risk areas. Simultaneously, equipment status change information is collected in real time during the inspection process. When abnormal equipment status or passage obstruction is detected, a dynamic path replanning module based on the D*Lite algorithm is triggered to quickly correct the local path, ensuring the continuity and safety of the inspection task.

[0047] In step S2, the image recognition of the chemical plant specifically includes: constructing an image recognition model based on a convolutional neural network, preferably using a ResNet network structure as the backbone feature extraction network to extract multi-scale features from the plant images; during the model training phase, the acquired plant images undergo grayscale normalization, noise suppression, and data augmentation to improve the model's generalization ability. After training, the model is deployed to the edge computing module of the inspection terminal to perform forward inference on the real-time acquired images, outputting the equipment status category and confidence information. For low-light or occluded scenes, an image preprocessing algorithm based on adaptive histogram equalization is introduced before inference to improve recognition accuracy.

[0048] In step S3, abnormal operating condition identification specifically includes: constructing an abnormality detection model based on multi-source state information. First, feature vectors of equipment operating parameters and image recognition results are fused to form unified state features. Then, the Isolation Forest algorithm is used to train the normal operating data to establish a normal operating condition distribution model. During actual operation, the current state features are input into the abnormality detection model. When the abnormality score exceeds a preset threshold, it is determined to be an abnormal operating condition, and corresponding early warning information and handling strategies are triggered based on the abnormality type.

[0049] In step S4, alarm management specifically includes: classifying and grading abnormal operating conditions; using a support vector machine classification algorithm to determine the type of abnormality; and sending alarm information to the management terminal via a message queue mechanism. Simultaneously, the time of occurrence of the abnormality, the handling measures, and the processing results are recorded. The system periodically performs statistical analysis on historical alarm data and adaptively adjusts the abnormality threshold using a sliding time window, thereby continuously optimizing the abnormality judgment model and forming a closed-loop management mechanism for alarm information, improving the accuracy and real-time performance of abnormality detection.

[0050] like Figure 2 As shown, the real-time monitoring and data analysis system for multi-source information fusion provided in this embodiment of the invention includes: Inspection route dynamic planning module: Automatically plans inspection routes based on inspection tasks and optimizes the inspection routes with the goal of minimizing the distance. Deep learning-based image recognition module: Uses image recognition algorithms to accurately identify the reading status of chemical instruments; Anomaly detection and classification module: Employs anomaly feature extraction algorithms to provide accurate data support for intelligent early warning and processing systems; Alarm tracing module: An alarm tracing system is adopted, including an alarm information transmission and processing system as well as an alarm information feedback and closed-loop management mechanism; Defect Library Module: Establish a defect library to guide the judgment of inspection results during the inspection process, including a defect standard library and a defect knowledge base; Inspection Management Module: The inspection management software consists of three software components: the inspection management system, the inspection system, and the inspection APP.

[0051] The inspection management system is deployed on a server and adopts a B / S architecture. Users can access the management system through a web page. The system realizes basic database maintenance system, equipment inspection point management, inspection task management, inspection result management, and inspection traceability functions.

[0052] The working principle of the multi-source information fusion real-time monitoring and data analysis system provided in this invention lies in the high degree of collaboration between hardware and software to achieve full-process automation and intelligence, from inspection task issuance, path planning, data collection, intelligent identification to anomaly handling and closed-loop management. The system relies on computer equipment, storage media, and information data processing terminals to construct an efficient and accurate ecosystem for chemical production safety monitoring. Its specific workflow and operating mechanism are as follows: 1. Task Initialization and Dynamic Path Planning: The system's operation begins with the inspection management module. Deployed on the server, the inspection management system uses a B / S architecture. Administrators access the basic database via a web interface to set equipment inspection points and generate inspection tasks. Once a task is issued, the dynamic path planning module immediately starts. This module doesn't simply connect the various points; instead, based on the specific distribution of the current tasks, it uses optimization algorithms (such as ant colony optimization or genetic algorithms) to calculate and automatically plan an optimal inspection route with the goal of minimizing the distance traveled. This mechanism effectively avoids inspection personnel repeatedly traveling back and forth within the complex chemical plant area, significantly improving inspection efficiency.

[0053] 2. On-site Data Acquisition and Intelligent Recognition: Inspection personnel, equipped with an inspection app, use information processing terminals to reach designated locations along planned routes. At this point, the system enters its core data acquisition phase. For the crucial step of reading chemical instrument data, the system no longer relies on manual visual interpretation and data entry, but instead activates a deep learning-based image recognition module. This module utilizes pre-trained deep learning algorithms such as convolutional neural networks to process instrument images captured by cameras in real time. It automatically locates the instrument panel, eliminates environmental interference such as lighting and dirt, accurately identifies pointer positions or digital displays, and converts image information into high-precision digital readings, thereby eliminating human estimation errors.

[0054] 3. Anomaly Feature Extraction and Comparison: The collected data then enters the anomaly detection and classification module. This module uses anomaly feature extraction algorithms to perform in-depth data analysis and interacts with the defect database module in real time. The defect database includes a "defect standard library" and a "defect knowledge base," storing the normal operating parameter ranges and historical failure modes of various chemical equipment. The system compares the real-time identified data with the thresholds in the standard library. If the data exceeds the normal range, or if the image features match a certain failure mode in the knowledge base (such as dial damage, zero reading, etc.), the system immediately determines it as an anomaly and automatically classifies the anomaly type, providing accurate data support for subsequent decision-making.

[0055] 4. Alarm Tracing and Closed-Loop Management: Once an anomaly is confirmed, the alarm tracing module is immediately triggered. The system does not merely issue an alarm, but initiates a complete alarm information transmission and processing flow. Alarm information is pushed to the terminals of relevant management and maintenance personnel immediately, and the system traces the source of the anomaly and related historical information according to the alarm tracing mechanism.

[0056] The entire process follows a closed-loop management mechanism: from identifying potential hazards (intelligent identification), reporting hazards (automatic alarm), handling hazards (repair dispatch), to feedback of results (re-inspection and verification after repair), all steps are recorded in the inspection management system. The inspection traceability function ensures that the source of every piece of data is traceable and the handling process of every defect is controllable.

[0057] 5. Conclusion In summary, this system integrates dynamic path planning, computer vision recognition, big data comparison and analysis, and closed-loop management processes through processor instruction execution. It transforms traditional "human-based" safety measures into a multi-source information fusion model combining "technology-based" and "human-based" measures, achieving all-weather, high-precision real-time monitoring and data analysis of the chemical production environment, greatly improving the response speed and scientific basis of safety management decisions.

[0058] I. Diversified Implementation of Multi-Source Information Acquisition and Support Configuration In this embodiment, the multi-source state information can originate from various types of sensing units. For example, the sensing unit may include one or more combinations of an image acquisition device, an operating parameter acquisition device, and an environmental state acquisition device. The image acquisition device may be an industrial camera, an infrared imaging device, or an explosion-proof camera; the operating parameter acquisition device may be a pressure sensor, a temperature sensor, a flow sensor, or a current detection module; and the environmental state acquisition device may include a gas concentration sensor, a vibration sensor, or a noise acquisition unit.

[0059] The aforementioned sensing units with different configurations can be used independently or combined and configured according to the needs of the inspection scenario, thereby supporting the broad expression of "multi-source state information" in the claims.

[0060] II. Multi-path Implementation of Multi-source Information Fusion During the information fusion phase, various fusion strategies can be employed to generate comprehensive state data. For example, fusion methods may include feature concatenation based on time synchronization, feature weighting based on weight allocation, or state mapping based on model inference.

[0061] In some embodiments, data from different sources can be time-aligned before fusion; in other embodiments, features can be extracted from each source data separately, and then comprehensive state data can be generated through association rules or mapping relationships. These multiple implementation paths ensure that "association fusion" is not limited to a single algorithm or processing flow.

[0062] III. Multi-layered implementation of the abnormal triggering and handling closed loop When the comprehensive status data meets the abnormal conditions, the abnormality can be triggered in various ways. For example, the abnormal conditions can be determined based on threshold judgment, trend change judgment, or historical status comparison judgment. The threshold can be set as a single threshold or as an interval threshold, and its numerical range can cover low value interval, medium value interval, and high value interval to adapt to different device states.

[0063] The generation of anomaly handling instructions can take the form of alarm prompts, task adjustments, inspection route replanning, or the generation of handling work orders. After the handling is completed, the handling results can be recorded and fed back to the anomaly judgment logic to update the subsequent anomaly identification criteria, thus forming a continuously evolving closed-loop mechanism.

[0064] IV. Supplementary Explanation of Strategic Non-Essential Details (Strategic Data Tracking) In some preferred embodiments, the system deployment can adopt a server-terminal collaborative architecture. The server can be configured as an industrial server or a cloud server, and the terminal can include a mobile inspection terminal or a fixed monitoring terminal. Data transmission can be wired or wireless, and the communication protocol can be selected according to the site environment.

[0065] In specific implementations, some modules can be deployed using containerization, abnormal data can be stored using relational databases or time-series databases, and defect information can be categorized and stored according to device type and abnormality level. The above implementation methods are merely illustrative examples and are not intended to limit the scope of this invention.

[0066] It should be noted that embodiments of the present invention can be implemented in hardware, software, or a combination of both. The hardware portion can be implemented using dedicated logic; the software portion can be stored in memory and executed by a suitable instruction execution system, such as a microprocessor or dedicated-design hardware. Those skilled in the art will understand that the above-described devices and methods can be implemented using computer-executable instructions and / or included in processor control code, for example, such code provided on a carrier medium such as a disk, CD, or DVD-ROM, a programmable memory such as read-only memory (firmware), or a data carrier such as an optical or electronic signal carrier. The devices and modules of the present invention can be implemented by hardware circuitry such as very large-scale integrated circuits or gate arrays, semiconductors such as logic chips, transistors, or programmable hardware devices such as field-programmable gate arrays, programmable logic devices, etc., or by software executed by various types of processors, or by a combination of the above-described hardware circuitry and software, such as firmware.

[0067] The above description is merely a specific embodiment of the present invention, but the scope of protection of the present invention is not limited thereto. Any modifications, equivalent substitutions, and improvements made by those skilled in the art within the scope of the technology disclosed in the present invention, and within the spirit and principles of the present invention, should be covered within the scope of protection of the present invention.

Claims

1. A real-time monitoring and data analysis method based on multi-source information fusion, characterized in that, include: Obtain multi-source status information from the inspected objects; Based on the multi-source state information, a comprehensive state data is formed to characterize the operating state of the inspected object. When the comprehensive status data meets the preset abnormal conditions, an abnormality determination is triggered and a corresponding handling instruction is generated.

2. The method according to claim 1, characterized in that, The multi-source status information includes inspection environment information, and the method further includes: An inspection path is generated based on the inspection environment information, and the inspection path is dynamically adjusted according to changes in the status of the inspected object.

3. The method according to claim 1, characterized in that, The multi-source status information includes image information of the inspected objects. The image information, after preprocessing, is used to extract feature information for state determination.

4. The method according to claim 1, characterized in that, The anomaly determination is based on the difference between historical state data and current state data.

5. The method according to claim 1, characterized in that, The transmission and feedback of abnormal information includes the recording and retrospection of abnormal handling results in order to update subsequent abnormal judgment rules.

6. A real-time monitoring and data analysis system that integrates multi-source information, characterized in that, include: The data acquisition module is used to acquire multi-source status information of the inspected objects; The fusion analysis module is used to correlate and fuse the multi-source state information to generate comprehensive state data. The exception triggering module is used to generate handling instructions when the comprehensive status data meets the exception conditions; The feedback update module is used to update the anomaly judgment criteria based on the handling results. This forms a closed-loop system for anomaly identification and handling.

7. The system according to claim 6, characterized in that, The data acquisition module includes an image acquisition unit for acquiring image information of the inspected objects.

8. The system according to claim 6, characterized in that, The feedback update module is connected to the defect database module, which stores exception types and corresponding handling rules.

9. A closed-loop management module for abnormal alarms in a multi-source monitoring system, characterized in that, The module includes: An anomaly information receiving unit is used to receive anomaly determination results; The exception handling instruction generation unit is used to generate exception handling instructions; The result feedback unit is used to receive the processing results and update the exception handling rules. To achieve closed-loop management of abnormal alarms.

10. The module according to claim 9, characterized in that, The module is deployed on the server and interacts with the terminal device via the network.

Images