Intelligent highway electromechanical equipment fault early warning and maintenance system

Abstract

The invention relates to an intelligent highway electromechanical equipment fault early warning and maintenance system, belongs to the technical field of highway electromechanical equipment management, and mainly solves the problems of response lag, low operation and maintenance efficiency and lack of predictive maintenance capability of a traditional maintenance mode. According to the system, an electromechanical equipment full-life-cycle standing book is established through an equipment management module, operation data is collected in real time, equipment state data and historical maintenance data are analyzed through a statistical analysis module, predictive early warning and health degree evaluation results are generated, and a maintenance management module generates and dispatches a maintenance work order according to early warning information. The handheld terminal APP is used for field execution and feedback, the GIS service module realizes visual display of the geographic position and state of the equipment, and the inspection management module is responsible for issuing and receiving an inspection task. According to the system, intelligent early warning, efficient maintenance and whole-process closed-loop management of the electromechanical equipment can be realized, and the reliability and the operation and maintenance efficiency of the equipment are remarkably improved.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of expressway electromechanical equipment management. More particularly, the present application relates to an intelligent expressway electromechanical equipment fault early warning and maintenance system. BACKGROUND

[0002] At present, the operation and maintenance management of expressway electromechanical systems mainly relies on the traditional manual inspection and after-maintenance mode. This mode has the problems of response lag and low operation and maintenance efficiency, and it is difficult to realize early warning and effective prevention of equipment faults. Since the types of electromechanical equipment are various, widely distributed and the running environment is complex, the state monitoring data is often in a scattered state, which cannot be systematically integrated and analyzed, resulting in the inability to accurately assess the equipment health status and development trend. The operation and maintenance personnel usually intervene only after the equipment has a dominant fault, which not only increases the traffic operation risk caused by sudden failure, but also makes the maintenance work often in a passive state.

[0003] In addition, the existing maintenance management mainly relies on paper records or isolated information systems, and there is a lack of effective association between historical maintenance data and real-time operation data, which makes it difficult to support data-based decision analysis. The equipment account information is not updated in time, and the location information is not intuitive, which also increases the difficulty of operation and maintenance management. In the inspection process, the task assignment and feedback mainly rely on manual transmission, which is easy to cause information delay, omission or incomplete management loop, further reducing the overall efficiency of operation and maintenance work.

[0004] The main reasons for the above problems include: the heterogeneous and non-uniform format of equipment data makes it difficult to integrate and comprehensively analyze the data; the lack of effective data modeling and prediction algorithms makes it difficult to extract the equipment degradation rules from massive data; at the same time, due to the large number of electromechanical equipment and wide geographical distribution, it is also difficult to realize real-time visualization and accurate positioning of equipment status. In the past, when trying to improve the operation and maintenance efficiency, the system integration is low, the cooperation between business modules is insufficient, and there is a lack of mobile and visual technology support, which makes it impossible to realize truly intelligent operation and maintenance management. SUMMARY

[0005] An object of the present application is to solve at least the above problems and to provide at least the advantages to be described later.

[0006] In order to achieve these objects and other advantages according to the present application, an intelligent expressway electromechanical equipment fault early warning and maintenance system is provided, which comprises a device management module, a maintenance management module, a statistical analysis module, a GIS service module, an inspection management module and a handheld terminal APP. The device management module is configured to establish and maintain full life cycle account information of all mechanical and electrical devices in the jurisdictional section, and to receive real-time operation state data, fault data and location data from the mechanical and electrical devices; The statistical analysis module is connected to the device management module and the maintenance management module, and is configured to receive and store the device operation state data and historical maintenance data from the maintenance management module; the statistical analysis module analyzes the received data to generate predictive early warning information and health degree evaluation results, and sends the early warning information and evaluation results to the maintenance management module and the device management module; The maintenance management module is connected to the device management module and the statistical analysis module, and is configured to receive device fault reporting information and predictive early warning information from the statistical analysis module; the maintenance management module generates repair work orders based on the received information, and dispatches, tracks and closes the loop of the work orders; The handheld terminal APP interacts with the maintenance management module and the inspection management module to receive work order tasks from the maintenance management module, and allows field personnel to perform operations and feed back on-site repair information to the maintenance management module; The GIS service module is connected to the device management module to obtain geographic location information and real-time state information of the mechanical and electrical devices from the device management module, and visually displays and spatially queries the information on a map; The inspection management module is connected to the handheld terminal APP to issue inspection tasks to the handheld terminal APP and receive inspection records uploaded by the handheld terminal APP.

[0007] Preferably, the inspection management module is further configured to formulate and issue inspection plans; The handheld terminal APP receives inspection tasks from the inspection management module, and allows field personnel to check in and record inspection results by scanning RFID tags on the devices, and transmits inspection records containing check-in information and hidden danger problems to the inspection management module.

[0008] Preferably, the system further comprises a personal space module; The personal space module is connected to the maintenance management module and the device management module to obtain and centrally display announcement information, pending work orders, processed work orders, knowledge bases and system policies from the connected modules according to the roles and permissions of users.

[0009] Preferably, the system further comprises a system management module; the system management module is configured to configure and manage department organizations, basic operation parameters and type classifications of mechanical and electrical devices in the system.

[0010] Preferably, the maintenance management module is further configured to: receive a failure repair request sent by a repairer and generate a failure repair order for the request; automatically or manually assign the repair order to a maintenance unit; receive repair feedback information filled in by a repairer through the handheld terminal APP; close the failure repair order after receiving confirmation of the repair from the repairer.

[0011] Preferably, it further comprises a plurality of edge computing nodes deployed along the roadside or station end of the expressway; The edge computing node is communicatively connected to a group of mechanical and electrical equipment in its geographical area, and is configured to receive and preprocess the running state data of the equipment in real time; the preprocessing includes preliminary judgment of abnormal state and data filtering based on preset rules; The edge computing node is also connected to the statistical analysis module for uploading the preprocessed data and preliminary judgment results; The predictive warning information generated by the statistical analysis module for the equipment in the area is preferentially issued to the corresponding edge computing node, which performs the first-time warning response.

[0012] Preferably, the maintenance management module further comprises an intelligent early warning routing unit; the intelligent early warning routing unit is configured to receive the predictive warning information issued by the statistical analysis module, and perform dynamic risk assessment and grading on the warning information to generate a risk level label; The intelligent early warning routing unit is preconfigured with processing strategies corresponding to different risk levels, and the processing strategies at least include notification objects, notification urgency and expected response time; The maintenance management module automatically routes the warning information to different user terminals, superior management units or external emergency service systems based on the risk level label and the corresponding processing strategy.

[0013] Preferably, the system further comprises a model self-optimization module; The model self-optimization module is connected to the maintenance management module and the statistical analysis module; The model self-optimization module is configured to continuously collect on-site repair feedback information returned by the handheld terminal APP and obtain closed repair order records from the maintenance management module; The model self-optimization module is also configured to associate and compare the repair feedback information with the warning information that initially triggered the order to generate a repair effect evaluation index; The model self-optimization module performs parameter optimization or algorithm iteration on the mechanical and electrical maintenance analysis and prediction model in the statistical analysis module based on the maintenance effect evaluation index and maintenance result data.

[0014] The present application at least includes the following beneficial effects: First, predictive maintenance is achieved: through the deep analysis of equipment operating status and historical data by the statistical analysis module, predictive warning information and health degree evaluation can be generated, turning "after-the-fact maintenance" into "before-the-fact maintenance", greatly reducing the occurrence of sudden failures and improving the availability and reliability of equipment.

[0015] Second, the scientific nature of decision-making is improved: the device full life cycle account, real-time running data and historical maintenance data are integrated, providing comprehensive and accurate data basis for equipment state evaluation and maintenance decision-making, and changing the operation and maintenance management from experience-driven to data-driven.

[0016] Third, the operation and maintenance management process is optimized: through the cooperation of the maintenance management module and the handheld terminal APP, the automatic generation, rapid distribution, on-site execution and feedback transmission of the maintenance work order are realized, the whole process is online and closed-loop management, significantly reducing the information delay and loss of intermediate links, and improving the operation and maintenance response speed and processing efficiency.

[0017] Fourth, the intuitive and spatial ability of management is enhanced: through the GIS service module, the equipment geographic location and real-time state information are integrated and visualized, so that the management personnel can macroscopically and intuitively master the whole line equipment situation, facilitate efficient resource scheduling and spatial query, and improve the management refinement level.

[0018] Fifth, the standardization and traceability of inspection work are ensured: through the cooperation of the inspection management module and the handheld terminal APP, the standardization of inspection task and the digitization of execution process are realized, ensuring the on-site rate and quality of inspection work, and forming a traceable electronic file.

[0019] Sixth, the multi-mode communication and cache design of the edge computing node ensures the continuous operation in the unstable network area, solving the technical problem of highway full-section coverage.

[0020] Other advantages, objects and features of the present application will be partially embodied in the following description, and partially understood by those skilled in the art through research and practice of the present application. DETAILED DESCRIPTION

[0021] The present application will be further described in detail below so that those skilled in the art can implement it with reference to the description.

[0022] It should be understood that the terms such as "have", "contain" and "include" used herein do not exclude the presence or addition of one or more other elements or combinations thereof.

[0023] The application provides a kind of intelligent highway electromechanical equipment fault early warning and maintenance system, including equipment management module, maintenance management module, statistical analysis module, GIS service module, inspection management module, handheld terminal APP; The equipment management module is used to establish and maintain the full life cycle account information of all electromechanical equipment in the jurisdiction section, and real-time receive the running state data, fault data and location data from the electromechanical equipment; The statistical analysis module is connected to the equipment management module and maintenance management module, for receiving and storing the equipment running state data and historical maintenance data from the maintenance management module;The statistical analysis module analyzes and processes the received data, generates predictive warning information and health degree evaluation results, and sends the warning information and evaluation results to the maintenance management module and the equipment management module; The maintenance management module is connected to the equipment management module and the statistical analysis module, for receiving equipment fault reporting information and predictive warning information from the statistical analysis module;The maintenance management module generates repair work order based on the received information, and dispatches, tracks and closes loop management of work order; The handheld terminal APP interacts with the maintenance management module and the inspection management module, for receiving work order tasks from the maintenance management module;For field personnel to perform operation, and feedback information of on-site repair to the maintenance management module; The GIS service module is connected to the equipment management module, for obtaining the geographic location information of electromechanical equipment and its real-time state information from the equipment management module, and visualizing and spatially querying on the map; The inspection management module is connected to the handheld terminal APP, for issuing inspection tasks to the handheld terminal APP, and receiving inspection records uploaded by the handheld terminal APP.

[0024] In the above technical solution, the equipment management module can be a cloud-based equipment management system that connects various electromechanical devices via IoT gateways. The real-time equipment operating status data received by this module can include parameters such as current, voltage, and temperature, with a sampling frequency set to once per minute. When a data value exceeds ±10% of the rated value, the system determines it as abnormal data and triggers a recording. Fault data can include information such as equipment communication interruptions and hardware fault codes. Location data can be obtained using a GPS positioning module, with positioning accuracy controlled within 10 meters. Equipment ledger information can be stored in a relational database, with each record containing fields such as equipment number, installation location, equipment model, and commissioning date.

[0025] The statistical analysis module can be deployed on a server with big data processing capabilities, employing a distributed storage architecture to store equipment operating status data. Historical maintenance data can be retrieved from the maintenance management module's MySQL database via an API interface, with an update frequency set to hourly synchronization. Data analysis and processing can utilize machine learning algorithms, such as random forests or LSTM neural networks, with training data derived from the past three years of equipment operating records. The threshold for generating predictive warning information can be set when the equipment health score falls below 80 points, and health assessment results can be presented in a quantitative format from 0 to 100 points. Warning information can be pushed to the maintenance management module via a message queue, and assessment results can generate weekly reports and update the equipment management module.

[0026] The maintenance management module utilizes a workflow engine for work order management, with fault reporting information coming from automatic equipment alarms or manual entry. Repair work orders are automatically assigned to the nearest maintenance team to the fault location, with a dispatch response time controlled within 5 minutes. Work order tracking is achieved through the location function of a handheld terminal app; closed-loop management requires confirmation from both on-site personnel and the person who reported the repair upon completion. The handheld terminal app can be installed on an industrial-grade tablet, supporting 4G / 5G network communication, and data interaction with the maintenance management module can be encrypted using the HTTPS protocol. The GIS service module can be developed based on the ArcGIS platform, using location services provided by the BeiDou satellite navigation system for map data, and supporting both heatmap and hierarchical rendering modes for visualization. The inspection management module can be deployed on an application server, automatically generating inspection tasks daily or weekly, and inspection records can include on-site photos and videos.

[0027] Regarding equipment selection, multi-protocol industrial smart gateways supporting communication protocols such as Modbus and OPC UA can be chosen for the IoT gateway. Vehicle-grade navigation and positioning modules can be selected for the GPS positioning module, with a positioning update frequency set to once per second. Rack-mount servers can be selected, with a configuration of at least 32-core CPUs and 128GB of memory. Industrial-grade tablet PCs should meet IP67 protection standards and have a battery life of at least 8 hours. These devices can be procured from mainstream domestic equipment suppliers.

[0028] During system deployment, the equipment management and statistical analysis modules can be installed in the data center, while the maintenance management and GIS service modules can be deployed on a cloud service platform. The handheld terminal app can be installed on the mobile devices of maintenance personnel along each road section, and the IoT gateway can be installed in equipment boxes along the highway. The GPS positioning module can be integrated into various electromechanical equipment controllers. System wiring can use national standard flame-retardant cables, and network connections can use single-mode fiber optic leased lines.

[0029] System testing can select all electromechanical equipment on a section of a highway as the test objects, with a testing period of six months. A comparative experiment can be used, employing traditional maintenance methods for the first three months and the current system for the following three months. Data collection can include indicators such as the number of equipment failures, average repair time, and maintenance costs. Statistical analysis can use the T-test to compare the significance of various indicators under the two maintenance methods. Test results show that using this system extends the average interval between equipment failures and reduces the number of emergency repairs.

[0030] Compared to existing technologies, current highway electromechanical equipment maintenance primarily relies on periodic inspections and post-failure repairs, lacking real-time monitoring and predictive analysis of equipment status. This system, by integrating equipment status monitoring, data analysis, and predictive early warning functions, achieves a shift from reactive to proactive maintenance. The system provides more timely maintenance reminders and more accurate fault predictions, contributing to improved equipment reliability and reduced maintenance costs.

[0031] In another technical solution, the inspection management module is also used to formulate and publish inspection plans; The handheld terminal APP receives inspection tasks from the inspection management module and allows on-site personnel to sign in and record inspection results by scanning the RFID tags on the device. The inspection records, which include sign-in information and potential problems, are then sent back to the inspection management module. In the above technical solution, the inspection management module can be deployed on a Windows Server operating system, developed using Java and the Spring Boot framework. Inspection plans can be generated daily, weekly, or monthly, with daily inspection tasks scheduled to be automatically issued at 8:00 AM. The plan content can include inspection routes, equipment lists, and inspection items. Key equipment can be scheduled for daily inspection, while general equipment can be inspected three times a week. The system can set an inspection timeout reminder threshold, automatically sending a reminder notification if the task has not started within two hours of being issued.

[0032] The handheld terminal app can be installed on an industrial tablet PC with NFC functionality, and the RFID tags can be high-frequency RFID tags conforming to the ISO 15693 standard. Each piece of electromechanical equipment can be equipped with one RFID tag, and the card memory can store the equipment number and location code information. The identification distance for on-site personnel check-in can be set to 5-10 cm, and the identification response time can be controlled within 0.5 seconds. Inspection result records can include equipment status photos, text descriptions, and abnormal situation options. The photo resolution can be set to 1920×1080 pixels, and the size of each photo can be controlled within 500KB. The description of potential hazards can use a combination of structured forms and free text. The form can include required fields such as hazard type and urgency level.

[0033] RFID tags can be installed in prominent locations on the control boxes of various electromechanical equipment, encased in weather-resistant ABS plastic shells, achieving an IP65 protection rating. Handheld terminals can be equipped with wrist straps and back clips for convenient on-site operation. Inspection data transmission can utilize 4G / 5G mobile networks; in areas with poor network signal, a data caching function can be enabled, automatically retransmitting once the signal is restored. The system can be configured with data integrity checks, automatically prompting for re-collection when missing records are detected. The inspection management module's backend allows for setting data review processes, and shift leaders can review and evaluate inspection records using their accounts.

[0034] Compared to existing inspection methods, traditional methods often rely on manual recording in paper notebooks, leading to problems such as illegible handwriting, data loss, and information lag. This implementation method, through an electronic inspection process, achieves automatic allocation of inspection tasks and digital recording of the execution process, ensuring the real-time nature and accuracy of data. The use of RFID tags for check-in ensures the authenticity of inspection visits, preventing missed inspections and proxy inspections. Structured data collection facilitates subsequent statistical analysis, helping to identify patterns and common problems in equipment malfunctions.

[0035] In another technical solution, a personal space module is also included; the personal space module is connected to the maintenance management module and the device management module, and is used to obtain and centrally display announcement information, pending work orders, processed work orders, knowledge base and system policies from the connected modules according to the user's role and permissions; In the above technical solution, the personal space module can be deployed on an application server cluster, implemented using a front-end / back-end separation architecture. The front-end interface can be developed based on the Vue.js framework, using the Element UI component library to build the user interface. The back-end service can be written in Java, providing interface services through the Spring Cloud microservice architecture. This module interacts with the maintenance management module and the equipment management module via a RESTful API. The API call timeout can be set to 3 seconds, and the number of retries can be set to 2. User roles can include different categories such as maintenance personnel, team leaders, and administrators, and each role can be configured with different data access permissions and function menus.

[0036] Data display content can be dynamically loaded based on user identity. After logging in with a maintenance personnel account, the main display shows a list of pending and processed work orders. Pending work orders can be sorted by urgency. Team leader accounts can additionally view work order statistics reports and equipment status overviews for their teams. Announcements can be differentiated into regular notices and emergency announcements; emergency announcements can be pinned to the top and remain displayed for 24 hours. Knowledge base documents support online preview in multiple formats such as PDF and Word, with a single file size limit of 50MB. Policy and regulation documents can be sorted by publication date, with the latest version automatically pinned to the top.

[0037] User authentication can use JWT tokens, and the session validity period can be set to 8 hours. Access control can be implemented using the RBAC model, allowing each user to be assigned multiple roles, each with different data access scopes. The front-end can adopt a responsive design, adapting to different terminal devices such as computers, tablets, and mobile phones. Data loading can use pagination, displaying 20 records per page. Important notification messages can be pushed in real-time via the WebSocket protocol, and message read status can be synchronously updated in the database.

[0038] Compared to existing technologies, traditional management systems typically use a unified interface to display all information, requiring users to frequently switch between different functional modules to obtain the information they need. This implementation achieves personalized, aggregated information display through a personal space module, providing targeted information views based on the responsibilities of different personnel. This design reduces the time users spend searching for information, improves work efficiency, and ensures data security through strict access control. The system also offers flexible configurability, allowing for easy adjustment of the displayed content and layout according to actual usage needs.

[0039] In another technical solution, a system management module is also included; the system management module is used to configure and manage the departmental organization, basic operating parameters, and type classification of electromechanical equipment within the system. In the above technical solution, the system management module can be deployed on a dedicated configuration management server, implemented using a B / S architecture. This module can be developed based on the Java EE platform, using the Spring MVC framework to provide web services. Department organization management can be displayed in a tree structure, supporting multi-level department settings. Each department can have fields such as department code, department name, and person in charge. Basic operating parameter configurations can include system timeout settings, log retention days, and data backup strategies. The session timeout can be set to 30 minutes, and operation logs can be retained for 180 days. Device type management can use a classification coding system, dividing devices into major categories such as monitoring devices, billing devices, and communication devices, with further subdivisions within each category.

[0040] In terms of hardware configuration, a 2U rack-mount server equipped with an Intel Xeon Silver series processor and 64GB of memory can be selected. The database can be MySQL 8.0, using the InnoDB storage engine. Network connectivity can utilize a Gigabit Ethernet interface, isolated from external networks by a firewall. The system can be equipped with an uninterruptible power supply (UPS) system, ensuring continuous operation for 4 hours in the event of a power outage. Data backup can employ a RAID 5 disk array with a daily incremental backup strategy, allowing critical configuration data to be synchronized to a backup server in real time.

[0041] The system management interface can employ role-based access control, with the system administrator account having the highest privileges. Department information maintenance supports batch import and export functions; import templates can be in Excel format. Equipment type codes can use an 8-digit encoding rule: the first two digits represent the major category, the middle three digits represent the minor category, and the last three digits represent the subcategory. The parameter configuration page provides default value setting and reset functions; all modification operations must be logged. The system provides configuration validation functionality to check the integrity and validity of data before saving.

[0042] Compared to existing technologies, traditional management systems often employ a decentralized configuration approach, resulting in inconsistent parameter settings across departments and varying equipment classification standards. This implementation method achieves unified management and maintenance of configuration parameters through a centralized system management module. A standardized equipment classification coding system ensures the standardization and consistency of equipment information. The system provides a robust data backup and recovery mechanism, guaranteeing the security of configuration data. Fine-grained access control prevents unauthorized configuration modifications, improving system stability and reliability.

[0043] In another technical solution, the maintenance management module is also used for: Receive fault repair requests sent by the user and generate a fault repair work order for the request; The maintenance work orders are automatically or manually assigned to the operation and maintenance unit. Receive maintenance feedback information filled in by maintenance personnel through the handheld terminal APP; After receiving confirmation from the person reporting the fault that the fault has been repaired, the fault repair work order is closed.

[0044] In the above technical solution, the fault reporting and handling function in the maintenance management module can be deployed on the application server, adopting a distributed architecture design. Fault request reception can be achieved through multiple channels, including web forms, mobile APP interfaces, and telephone recording transcription systems. The fault reporting request form can include required fields such as device number, fault description, and urgency level, where urgency can be divided into three levels: general, important, and urgent. After receiving a request, the system can generate a repair work order within 30 seconds. The work order number can be generated using a rule of year, month, day plus a serial number, such as 20250822-0001. Work order allocation can be automatically executed based on preset rules, prioritizing allocation to maintenance teams within 5 kilometers of the fault location and with fewer than 3 current work orders.

[0045] The work order allocation engine can integrate with a map API to calculate the optimal dispatch plan, taking into account factors such as the repair personnel's current location, skill matching, and traffic conditions. If automatic allocation fails, it can be transferred to manual dispatch. The dispatcher's interface displays all pending work orders and the repair personnel's status dashboard. Repair feedback information can be collected through forms on a handheld app, including fields such as fault cause analysis, replacement parts list, and repair time, accurate to the minute. Repair requester confirmation can be achieved via SMS link or app push message; the confirmation interface displays before-and-after photos and repair instructions. Closing a work order requires two conditions: the repair personnel submitting complete feedback information and the repair requester clicking "confirm."

[0046] The system allows for setting up timeout monitoring for work orders. For urgent work orders, if processing hasn't begun within 2 hours, the system will automatically escalate the notification. All work order workflow records can be stored in a MySQL database, supporting queries and statistics by multiple dimensions such as time, device, and work group. Message notifications can be implemented using a RabbitMQ message queue to ensure reliable delivery. Work order satisfaction ratings can be provided by inviting the person who requested the service to rate the work order within 24 hours of its closure, using a 1-5 star rating system.

[0047] Compared to existing fault reporting systems, traditional methods rely heavily on telephone communication and paper records, which are prone to information omissions and transmission errors. This implementation method ensures complete collection and efficient transmission of fault information through standardized processes and electronic processing. An automatic dispatch algorithm improves dispatch efficiency and reduces the workload of manual dispatching. A double-confirmation mechanism guarantees the traceability of repair quality, providing data support for service quality improvement. The entire process forms a complete closed-loop management system, contributing to the continuous improvement of the standardization of repair services.

[0048] Another technical solution also includes multiple edge computing nodes deployed along the highway or at the station end; The edge computing node is communicatively connected to a group of electromechanical devices within its geographical area, and is used to receive and preprocess the operating status data of the devices in real time; the preprocessing includes preliminary judgment of abnormal states and data filtering based on preset rules; The edge computing node is also connected to the statistical analysis module, which is used to upload the preprocessed data and preliminary judgment results; The predictive early warning information generated by the statistical analysis module for devices in the area is preferentially sent to the corresponding edge computing nodes, which then provide an immediate early warning response.

[0049] In the above technical solution, the edge computing nodes can be industrial-grade edge computing gateways, and the deployment spacing can be set at one node every 2-5 kilometers according to the device density. These nodes can be installed in all-weather protected enclosures, with an operating temperature range that can be controlled between -40℃ and 75℃, and a protection level of IP67. Each node can connect to the electromechanical equipment along the line via Ethernet or fiber optic, supporting industrial protocols such as Modbus TCP and OPC UA, and can connect up to 256 devices. The preprocessing rules built into the nodes can include threshold judgment and rate of change monitoring, such as triggering an anomaly flag when the current fluctuation exceeds the rated value by 15% or the temperature rises by more than 2℃ per minute. The preprocessing includes: marking an anomaly when the current fluctuation exceeds the rated value ±15% or the temperature rises by more than 2℃ per minute, and using a moving average filter (window size = 10 sampling points) to eliminate instantaneous noise.

[0050] The edge computing node is equipped with a multi-mode communication module, including a primary 4G / 5G channel and a backup LoRa / Wi-Fi Mesh self-organizing network channel. When the primary network is interrupted, it automatically switches to the backup channel and enables local data caching (storage capacity ≥ 72 hours of data), automatically resuming transmission after the network is restored. For extremely remote road sections, the node supports data relay via patrol vehicles or drones as mobile relays.

[0051] Data preprocessing can employ lightweight algorithms running on ARM architecture processors, with processing latency controlled to within 100 milliseconds. Data filtering can utilize a moving average algorithm, with a window size set to 10 sampling points. Anomaly detection results can generate concise JSON messages, including device ID, anomaly type, and timestamp. Preprocessed data can be uploaded in batches to the cloud-based statistical analysis module every 5 minutes, achieving a data compression rate of up to 60%. In the event of a network outage, edge nodes can locally store 72 hours of data, automatically resuming transmission once the network is restored.

[0052] The early warning response mechanism can be configured with multiple trigger conditions. When an edge node receives a red alert, it can immediately activate the on-site audible and visual alarms and send an SMS to the area manager. Yellow alerts can be displayed on the node's local human-machine interface and generate a daily report. Nodes can be equipped with a 4G / 5G cellular network module as the primary communication channel, while using a wired network as a backup link. Each node can be equipped with an independent battery, maintaining operation for 8 hours in the event of an external power outage.

[0053] Compared to existing pure cloud processing models, traditional methods require remote transmission of all device data to a central server, placing high demands on network bandwidth and resulting in significant response latency. This implementation deploys edge nodes near the devices, enabling localized data preprocessing and rapid response, significantly reducing network transmission load. Edge nodes can instantly assess and initially address sudden anomalies, providing filtered, high-value data for in-depth analysis by the central system. This architecture is particularly suitable for applications like highways, where devices are widely distributed and network conditions are unstable, improving system reliability and real-time response.

[0054] In another technical solution, the maintenance management module further includes an intelligent early warning routing unit; the intelligent early warning routing unit is used to receive predictive early warning information issued by the statistical analysis module, and to perform dynamic risk assessment and classification on the early warning information to generate risk level labels; The intelligent early warning routing unit is pre-set with processing strategies corresponding to different risk levels. The processing strategies include at least the notification target, the urgency of the notification, and the expected response time. Based on the risk level label and the corresponding processing strategy, the maintenance management module automatically routes the early warning information to different user terminals, superior management units, or external emergency service systems.

[0055] In the above technical solution, the intelligent early warning routing unit can be deployed on an application server cluster, using a microservice architecture for independent deployment. This unit can receive early warning information from the statistical analysis module via a message queue. The message format can adopt the JSON standard, including fields such as device ID, early warning type, and confidence level. Dynamic risk assessment and grading can use a multi-factor weighted scoring model. Scoring factors can include device criticality coefficient, early warning severity index, and historical failure frequency, with weights set to 50%, 30%, and 20%. Risk levels can be divided into three levels: red, orange, and yellow. A total score of 85 or above is rated as a red early warning, 70-84 as an orange early warning, and 55-69 as a yellow early warning.

[0056] The processing strategy library can be stored using a relational database. Each strategy record includes fields such as risk level, list of notification recipients, notification method, and response time limit. Notification recipients can be configured according to an organizational structure tree. A red alert requires simultaneous notification to the maintenance foreman, technical supervisor, and section manager; an orange alert requires notification to the maintenance foreman and technical supervisor; and a yellow alert only requires notification to on-site maintenance personnel. The urgency level of the notification can be divided into three levels: immediate, urgent, and routine. A red alert uses the immediate level, requiring double confirmation via phone and SMS within 5 minutes; an orange alert uses the urgent level, requiring notification via app push and SMS within 15 minutes; and a yellow alert uses the routine level, requiring notification via app push within 30 minutes. Expected response times can be set to 30 minutes, 2 hours, and 8 hours, respectively.

[0057] Warning information routing and distribution can be achieved through multiple communication channels. High-risk warnings, in addition to internal system notifications, can also utilize API interfaces to call SMS platforms and external emergency service systems. The routing engine can be driven by a rule engine, supporting online updates of routing policies without system downtime. All warning distribution records need to be persistently stored, including information such as reception time, sending channel, and confirmation status. The system can be configured with rules for escalating unconfirmed warnings; if a red warning is not confirmed within 30 minutes, it will automatically escalate and notify the next higher-level supervisor.

[0058] Compared to existing early warning processing methods, traditional systems typically use fixed rules for early warning pushes, failing to differentiate processing based on the actual risk level of the warning. This implementation method, by establishing a dynamic risk assessment model and a tiered processing strategy, achieves intelligent routing and distribution of early warning information, ensuring that high-risk warnings are quickly delivered to key decision-makers. A multi-channel notification mechanism improves the reliability of early warning delivery, and tiered response requirements allow for a more rational allocation of operational resources. This refined early warning management approach significantly improves the efficiency of emergency response to high-risk events and reduces operational risks caused by untimely information transmission.

[0059] In another technical solution, the system further includes a model self-optimization module; The model self-optimization module is connected to the maintenance management module and the statistical analysis module; The model self-optimization module is used to continuously collect on-site maintenance feedback information transmitted back by the handheld terminal APP, and to obtain closed maintenance work order records from the maintenance management module. The model self-optimization module is also used to correlate and compare the maintenance feedback information with the early warning information that initially triggered the work order to generate maintenance effect evaluation indicators. The model self-optimization module, based on the maintenance effect evaluation indicators and maintenance result data, performs parameter tuning or algorithm iteration on the electromechanical maintenance analysis and prediction model in the statistical analysis module.

[0060] In the above technical solution, the model self-optimization module can be deployed on a data analysis server, developed using Python and integrated with the TensorFlow machine learning framework. This module collects maintenance feedback information from the handheld terminal APP hourly via a database interface. Data fields include structured data such as maintenance measures, replaced parts, and maintenance duration. Simultaneously, this module extracts closed work order records daily from the maintenance management module's work order database, obtaining fields including work order number, equipment type, and fault classification. Data association and comparison can employ a matching algorithm based on work order number, maintaining a matching success rate above 99.5%. During the initial system deployment, the model self-optimization module reuses pre-trained models from other road sections using transfer learning technology and synthesizes supplementary training data using a Generative Adversarial Network (GAN). For real-time collected data, a sliding window Z-score algorithm is used for outlier filtering to ensure input data quality. The first three months after system launch are a data accumulation period, during which the prediction model primarily uses a rule engine (with preset threshold alarms) while simultaneously collecting equipment operation data. After three months, when the effective data volume reaches 100,000 records, the machine learning model is automatically enabled and the rule engine is disabled.

[0061] The maintenance effectiveness evaluation index can adopt a multi-dimensional scoring system, including indicators such as maintenance time ratio (the ratio of actual maintenance time to estimated time), recurrence rate (the recurrence of the same fault in the same equipment), and parts matching degree. Each indicator can be assigned a corresponding weight coefficient, ultimately generating a comprehensive evaluation score from 0 to 100. The evaluation threshold can be set to below 60 points, indicating poor maintenance effectiveness, requiring in-depth analysis of the causes. The model optimization process can adopt an automated pipeline, with a model retraining task starting every Sunday morning. The training data can include the maintenance data from the most recent 180 days.

[0062] Parameter tuning can employ a grid search approach, focusing on optimizing the confidence threshold and feature weights of the prediction model. Algorithm iteration can utilize A / B testing, validating the new model alongside the current production model in parallel, with a testing period of up to 7 days. Model performance evaluation can use metrics such as accuracy, recall, and F1 score. The new model should only be deployed to the production environment when it outperforms the existing model by more than 3 percentage points across all metrics. Each model update requires the retention of version snapshots and training data backups, supporting rollback to previous versions when necessary.

[0063] Compared to existing technologies, traditional predictive maintenance systems often rely on periodic manual updates, lacking automated optimization mechanisms based on actual maintenance results. This implementation establishes a continuous feedback learning loop, enabling the predictive model to continuously adjust and optimize based on actual maintenance outcomes. This self-optimization mechanism ensures continuous improvement in model prediction accuracy, allowing the system to adapt to dynamic factors such as equipment aging and environmental changes. The automated pipeline design reduces the need for manual intervention, improving the efficiency of model maintenance. A version management mechanism guarantees the security and traceability of system updates, ensuring long-term stable operation.

[0064] The number of devices and processing scale described herein are for simplification. Applications, modifications, and variations of the intelligent highway electromechanical equipment fault early warning and maintenance system of this invention will be readily apparent to those skilled in the art.

[0065] Although the embodiments of the present invention have been disclosed above, they are not limited to the applications listed in the specification and embodiments. They can be applied to various fields suitable for the present invention. For those skilled in the art, other modifications can be easily made. Therefore, without departing from the general concept defined by the claims and their equivalents, the present invention is not limited to the specific details.

Claims

1. An intelligent highway electromechanical equipment fault early warning and maintenance system, characterized in that, It includes an equipment management module, a maintenance management module, a statistical analysis module, a GIS service module, an inspection management module, and a handheld terminal APP; The equipment management module is used to establish and maintain the full life cycle ledger information of all electromechanical equipment within the managed road section, and to receive real-time operating status data, fault data and location data from the electromechanical equipment; The statistical analysis module, connected to the equipment management module and the maintenance management module, is used to receive and store the equipment operating status data and historical maintenance data from the maintenance management module; the statistical analysis module analyzes and processes the received data, generates predictive early warning information and health assessment results, and sends the early warning information and assessment results to the maintenance management module and the equipment management module. The maintenance management module is connected to the equipment management module and the statistical analysis module, and is used to receive equipment fault reporting information and predictive early warning information from the statistical analysis module; The maintenance management module generates maintenance work orders based on the received information, and performs work order assignment, tracking, and closed-loop management. The handheld terminal APP interacts with the maintenance management module and the inspection management module to receive work orders from the maintenance management module. This allows on-site personnel to perform operations and transmits on-site maintenance feedback information back to the maintenance management module. The GIS service module is connected to the equipment management module and is used to obtain the geographical location information and real-time status information of electromechanical equipment from the equipment management module, and to visualize and query the data on the map. The inspection management module is connected to the handheld terminal APP and is used to issue inspection tasks to the handheld terminal APP and receive inspection records uploaded by the handheld terminal APP.

2. The intelligent highway electromechanical equipment fault early warning and maintenance system as described in claim 1, characterized in that, The inspection management module is also used to formulate and publish inspection plans; The handheld terminal APP receives inspection tasks from the inspection management module and allows on-site personnel to sign in and record inspection results by scanning the RFID tags on the device. The inspection records, which include sign-in information and potential problems, are then sent back to the inspection management module.

3. The intelligent highway electromechanical equipment fault early warning and maintenance system as described in claim 1, characterized in that, It also includes a personal space module; The personal space module is connected to the maintenance management module and the device management module. It is used to obtain and centrally display announcement information, pending work orders, processed work orders, knowledge base and policies from the connected modules according to the user's role and permissions.

4. The intelligent highway electromechanical equipment fault early warning and maintenance system as described in claim 3, characterized in that, It also includes a system management module; the system management module is used to configure and manage the departmental organization, basic operating parameters and types of electromechanical equipment within the system.

5. The intelligent highway electromechanical equipment fault early warning and maintenance system as described in claim 1, characterized in that, The maintenance management module is also used for: Receive fault repair requests sent by the user and generate a fault repair work order for the request; The maintenance work orders are automatically or manually assigned to the operation and maintenance unit. Receive maintenance feedback information filled in by maintenance personnel through the handheld terminal APP; After receiving confirmation from the person reporting the fault that the fault has been repaired, the fault repair work order is closed.

6. The intelligent highway electromechanical equipment fault early warning and maintenance system according to claim 1, characterized in that, It also includes multiple edge computing nodes deployed along the highway or at the station end; The edge computing node is communicatively connected to a group of electromechanical devices within its geographical area, and is used to receive and preprocess the operating status data of the devices in real time; the preprocessing includes preliminary judgment of abnormal states and data filtering based on preset rules; The edge computing node is also connected to the statistical analysis module, which is used to upload the preprocessed data and preliminary judgment results; The predictive early warning information generated by the statistical analysis module for devices in the area is preferentially sent to the corresponding edge computing nodes, which then provide an immediate early warning response.

7. The intelligent highway electromechanical equipment fault early warning and maintenance system according to claim 1 or 6, characterized in that, The maintenance and management module also includes an intelligent early warning routing unit; the intelligent early warning routing unit is used to receive predictive early warning information issued by the statistical analysis module, and to perform dynamic risk assessment and classification on the early warning information to generate risk level labels; The intelligent early warning routing unit is pre-set with processing strategies corresponding to different risk levels. The processing strategies include at least the notification target, the urgency of the notification, and the expected response time. Based on the risk level label and the corresponding processing strategy, the maintenance management module automatically routes the early warning information to different user terminals, superior management units, or external emergency service systems.

8. The intelligent highway electromechanical equipment fault early warning and maintenance system according to claim 1 or 6, characterized in that, The system also includes a model self-optimization module; The model self-optimization module is connected to the maintenance management module and the statistical analysis module; The model self-optimization module is used to continuously collect on-site maintenance feedback information transmitted back by the handheld terminal APP, and to obtain closed maintenance work order records from the maintenance management module. The model self-optimization module is also used to correlate and compare the maintenance feedback information with the early warning information that initially triggered the work order to generate maintenance effect evaluation indicators. The model self-optimization module, based on the maintenance effect evaluation indicators and maintenance result data, performs parameter tuning or algorithm iteration on the electromechanical maintenance analysis and prediction model in the statistical analysis module.