A power grid equipment maintenance early warning device

Abstract

The application discloses a power grid equipment maintenance early warning device, and belongs to the technical field of power grid maintenance. The device comprises a data acquisition module, a real-time account updating module, a big data analysis and judgment module and an intelligent early warning and work order generation module to form a complete operation and maintenance system. A multi-dimensional state sensing system fuses multi-dimensional monitoring data of electrical quantity, temperature, partial discharge and vibration, breaks through the limitation of traditional monitoring of electrical quantity, represents the health state of equipment from the aspects of electrical performance, thermal state, insulation state and mechanical state, can more comprehensively identify hidden faults of equipment, especially early hidden dangers such as insulation deterioration and mechanical wear, significantly improves the identification accuracy, realizes dynamic management of equipment account and automatic closed loop of operation and maintenance process, greatly reduces the work load of dispatching personnel and operation and maintenance personnel, and improves the standardization and normalization level of operation and maintenance management.

Description

Technical Field

[0001] This invention relates to the field of maintenance early warning technology, specifically to a power grid equipment maintenance early warning device. Background Technology

[0002] With the continuous advancement of smart grid construction, the scale of power system grids and equipment capacity are growing exponentially. The safe and stable operation of grid equipment directly determines power supply reliability and socio-economic benefits. Currently, the value of equipment operation status monitoring in power equipment operation and maintenance management is increasingly prominent, driven by three core factors: First, power equipment failures can trigger unplanned power outages, causing not only losses in electricity sales for power supply companies but also economic losses for downstream users such as industrial production and residential life. Therefore, early warning of failures and planned maintenance have become core requirements. Second, operation and maintenance management departments hope to extend equipment maintenance cycles, shorten the duration of single maintenance, and reduce outage windows to minimize outage losses. Third, precise condition-based operation and maintenance can avoid excessive maintenance that damages equipment, effectively extending equipment lifespan and reducing total life-cycle costs.

[0003] For a long time, domestic power companies have generally adopted a preventative periodic maintenance system of "mandatory maintenance upon expiration." This system sets fixed maintenance cycles based on the service life of equipment and played a certain role in the early stages of power grid development. However, with the surge in the number of power grid equipment and its performance improvement, the drawbacks of periodic maintenance have gradually become apparent: On the one hand, some equipment with hidden faults may not have reached its maintenance cycle and cannot be inspected in a timely manner, resulting in "missed inspections." This can accelerate equipment performance degradation and shorten its service life, or even lead to equipment failures and power supply accidents. On the other hand, the forced power outages for maintenance of a large number of healthy operating equipment result in the ineffective consumption of human, material, and financial resources. At the same time, frequent power outages actually increase the risk of equipment failure.

[0004] At the level of dispatch automation, the OPEN3000 system has become the core support system for power grid dispatching in China, realizing real-time acquisition, monitoring, and alarm of power grid operation data, and possessing comprehensive SCADA (Supervisory Control and Data Acquisition) functions. However, there are still significant shortcomings in the current equipment status monitoring based on the OPEN3000 system: First, equipment ledger management is independent of the operation system, mostly relying on static manual input, which cannot be linked and updated with real-time equipment operation data, resulting in lag and bias in the equipment status information obtained by maintenance personnel; Second, equipment status screening relies on dispatch personnel manually filtering abnormal information from massive amounts of operation data. The large data volume and inconsistent screening standards not only cause serious waste of human resources but also easily lead to missed or misjudged anomalies; Third, the existing system only has the ability to provide "post-fault alarms" after a fault occurs, and cannot achieve "pre-fault warnings" based on operation trends, making it difficult to support the implementation of condition-based maintenance.

[0005] In recent years, technologies such as big data mining and edge computing have been gradually integrated into the power operation and maintenance field, providing a technical path to solve the aforementioned problems. By deeply correlating the operational data of the OPEN3000 system with equipment operation and maintenance data, an intelligent maintenance early warning mechanism can be built, enabling automatic assessment of equipment status, early fault warning, and automation of maintenance processes. This is the core direction for the transformation of power grid operation and maintenance from "periodic maintenance" to "condition-based maintenance." However, existing technical solutions often suffer from poor system compatibility, high hardware modification costs, insufficient accuracy of status assessment algorithms, and a lack of closed-loop operation and maintenance processes, making them difficult to directly adapt to the existing power grid dispatching system.

[0006] Based on the current realities and technological status of power grid operation and maintenance, the core deficiencies and unresolved technical problems of existing technologies can be summarized into four points: The maintenance model is extensive and the resource allocation is unbalanced: The traditional periodic maintenance system adopts a "one-size-fits-all" cycle management, which cannot match the actual operating status of different equipment. This results in both the safety risk of missing inspections of potentially hazardous equipment and the waste of resources due to over-inspection of healthy equipment, making it difficult to balance power supply reliability and operation and maintenance costs.

[0007] Static data in equipment ledgers leads to delayed status awareness: Existing equipment ledgers are mostly managed offline and statically, which is disconnected from real-time equipment operation data and maintenance records. Data updates rely on manual entry, resulting in insufficient timeliness and accuracy, and failing to provide accurate data support for maintenance decisions.

[0008] Manual screening is inefficient and the early warning mechanism is delayed: Equipment status screening based on the OPEN3000 system relies on manual work. With massive amounts of operational data, manual judgment is inefficient and inconsistent in standards. Furthermore, the system can only issue alarms after the fault triggers a threshold, and cannot identify early deterioration trends, resulting in a significant lag in early warning.

[0009] The operation and maintenance process is disconnected and the level of automation is low: the equipment status monitoring, anomaly analysis and maintenance work order generation are independent of each other and have not formed an automated closed loop. Maintenance work orders need to be manually compiled and filled out, which is cumbersome and prone to errors, making it difficult to improve operation and maintenance efficiency. Summary of the Invention

[0010] This invention addresses the aforementioned shortcomings of existing technologies by providing a power grid equipment maintenance early warning device. This device achieves dynamic management of equipment records and an automated closed-loop operation and maintenance process, significantly reducing the workload of dispatching and maintenance personnel and improving the standardization and normalization of operation and maintenance management. Simultaneously, the equipment's full lifecycle data archive provides data support for power grid equipment selection and operation and maintenance strategy formulation, enhancing the refined management level of power companies.

[0011] To achieve the above objectives, the present invention provides the following technical solution: a power grid equipment maintenance early warning device, comprising a data acquisition module, a real-time ledger update module, a big data analysis and judgment module, and an intelligent early warning and work order generation module, forming a complete operation and maintenance system; The data acquisition module is deployed at the edge layer of the substation, synchronously connecting to the OPEN3000 system to obtain electrical operation data, and simultaneously collecting non-electrical status data of the equipment. After preprocessing at the edge side, the data is encrypted and uploaded to the cloud. The real-time ledger update module is deployed in the cloud layer of the dispatch center, linking basic equipment parameters, real-time operating data and full maintenance records to build a dynamically updated equipment lifecycle data archive. The big data analysis and judgment module has a built-in combined weight fuzzy comprehensive evaluation model based on minimum relative entropy fusion, which quantifies the health of equipment and predicts the deterioration trend, thereby realizing the early identification of potential faults. The intelligent early warning and work order generation module executes tiered early warnings based on the health assessment results and automatically generates standardized maintenance work orders, forming a fully automated operation and maintenance closed loop of status monitoring, intelligent assessment, tiered early warning, work order execution, and effect feedback.

[0012] Furthermore, the data acquisition module includes an electrical quantity acquisition unit and a non-electrical quantity acquisition unit; the electrical quantity acquisition unit interfaces with the OPEN3000 system front-end unit through an IEC60870-5-104 standard protocol converter to acquire telemetry and telesignal electrical operation data; the non-electrical quantity acquisition unit includes temperature, partial discharge, and vibration sensor arrays to simultaneously acquire equipment thermal status, insulation status, and mechanical status data.

[0013] Furthermore, the real-time ledger update module is equipped with an RFID identification unit and a ledger synchronization server; the RFID identification unit is used to collect on-site inspection and maintenance operation information, and the ledger synchronization server synchronizes OPEN3000 operating data, on-site collected data and maintenance management system data from multiple sources, replacing the traditional static ledger and realizing real-time linkage update of equipment status data.

[0014] Furthermore, the combined weight fuzzy comprehensive evaluation model integrates the subjective weights of the analytic hierarchy process and the objective weights of the entropy weight method, solves the optimal combined weights based on the principle of minimum relative entropy, constructs a three-level evaluation system for equipment health status, and outputs a quantifiable equipment health score.

[0015] Furthermore, the intelligent early warning and work order generation module is equipped with a four-level early warning mechanism, which is divided into four levels: normal, attention, abnormal, and severe based on the equipment health score. Each level corresponds to a different early warning push path and alarm triggering method. All early warning information is synchronously transmitted back to the OPEN3000 system for unified display.

[0016] Furthermore, for abnormal and higher warning levels, the intelligent warning and work order generation module automatically matches the equipment type and fault type, retrieves equipment ledger parameters and historical maintenance records, generates standardized maintenance work orders and pushes them to the maintenance management system, replacing the manual filling process.

[0017] Furthermore, the substation edge layer is equipped with an industrial-grade edge computing gateway, which has the ability to cache local data and resume transmission after network outage; the dispatch center cloud layer adopts a distributed server cluster deployment and is set with the same security partition as the OPEN3000 system, meeting the requirements of the third level of power information security protection.

[0018] Furthermore, the overall hardware connection of the device follows a hierarchical logic of perception layer, edge layer, cloud layer, and application layer.

[0019] Furthermore, the specific topology is as follows: sensor, field acquisition terminal, edge computing gateway, power dedicated intranet, ledger synchronization server, big data analysis server cluster, intelligent early warning terminal, work order web server, dispatching screen, handheld terminal, printer. At the same time, the OPEN3000 system front-end unit communicates bidirectionally with the edge computing gateway through a protocol converter to realize the acquisition of operation data and the transmission of early warning information.

[0020] Furthermore, this device adopts an improved combined weight fuzzy comprehensive evaluation model, which integrates the subjective weights of the analytic hierarchy process and the objective weights of the entropy weight method through the principle of minimum relative entropy to obtain a combined weight that takes into account both expert experience and data characteristics. Then, the device health is calculated through fuzzy comprehensive evaluation, providing a quantitative basis for graded early warning.

[0021] Compared with the prior art, the beneficial effects of the present invention are: 1. This invention comprises a data acquisition module, a real-time ledger update module, a big data analysis and judgment module, and an intelligent early warning and work order generation module, forming a complete operation and maintenance system. The multi-dimensional status perception system integrates multi-dimensional monitoring data of electrical quantity, temperature, partial discharge, and vibration, breaking through the limitations of traditional methods that only monitor electrical quantity. It characterizes the health status of equipment from multiple dimensions, including electrical performance, thermal state, insulation state, and mechanical state, and can more comprehensively identify hidden equipment faults. In particular, it significantly improves the accuracy of identification for early hidden dangers such as insulation degradation and mechanical wear.

[0022] 2. This device enables early warning and precise assessment of equipment failures, effectively identifying early-stage potential problems and preventing power supply accidents caused by equipment deterioration, significantly improving the reliability and safety of power grid equipment operation. Calculations show that after applying this device, the unplanned outage rate of primary equipment in substations can be reduced by more than 40%, effectively ensuring the safe and stable operation of the power grid. On the one hand, precise condition-based maintenance replaces traditional periodic maintenance, extending the maintenance cycle of healthy equipment, avoiding the waste of manpower and resources caused by excessive maintenance, and reducing power outage maintenance time and losses. On the other hand, early detection of potential problems can prevent significant economic losses caused by equipment failures and extend equipment lifespan. Calculations show that after applying this device to a single 110kV substation, the average annual operation and maintenance cost can be reduced by more than 25%, and the average equipment lifespan can be extended by 3-5 years, resulting in significant economic benefits.

[0023] 3. The device of this invention realizes dynamic management of equipment ledgers and an automated closed loop of operation and maintenance processes, significantly reducing the workload of control and maintenance personnel and improving the standardization and normalization of operation and maintenance management. Simultaneously, the equipment's full lifecycle data archive can provide data support for power grid equipment selection and operation and maintenance strategy formulation, enhancing the refined management level of power companies. Attached Figure Description

[0024] Figure 1 This is a schematic diagram of the overall structure of a power grid equipment maintenance early warning device according to the present invention; Figure 2 This is a schematic diagram of the structure of a cloud analysis server cluster for a power grid equipment maintenance early warning device according to the present invention; Figure 3 This is a schematic diagram of the structure of an intelligent early warning module of a power grid equipment maintenance early warning device according to the present invention. Detailed Implementation

[0025] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0026] Example 1 This embodiment provides a power grid equipment maintenance early warning device. Based on an architecture design of edge data acquisition, cloud-based analysis, and closed-loop operation and maintenance, it is deeply compatible with the existing OPEN3000 dispatch automation system, requiring no large-scale modification of the existing dispatch system. The device consists of four core units: a data acquisition module, a real-time ledger update module, a big data analysis and judgment module, and an intelligent early warning and work order generation module. Through the full-process automation of multi-dimensional data acquisition, dynamic ledger synchronization, intelligent algorithm analysis, hierarchical early warning push, and automatic work order generation, it realizes early warning and intelligent operation and maintenance of power grid equipment status.

[0027] The core technical logic of this device is as follows: the data acquisition module synchronously acquires electrical operation data and newly added non-electrical sensor data of the OPEN3000 system, and transmits them to the cloud after preprocessing at the edge; the real-time ledger update module synchronizes with the intranet data through RFID technology to realize the linkage update of equipment ledgers with operation and maintenance data; the big data analysis and judgment module adopts an improved combined weight fuzzy comprehensive evaluation algorithm to quantitatively assess the health status of equipment and predict fault trends, and identify early hidden dangers; the intelligent early warning and work order generation module pushes early warning information in a hierarchical manner according to the judgment results, and automatically matches the maintenance strategy to generate standardized work tickets, forming a complete operation and maintenance closed loop of monitoring, judgment, early warning, maintenance, and feedback.

[0028] The hardware architecture of this device adopts a two-tier deployment architecture: the substation edge layer and the dispatch center cloud layer. This ensures both the real-time nature of on-site data acquisition and centralized analysis of big data in the cloud. All hardware meets power system electromagnetic compatibility and industrial-grade protection standards, making it adaptable to the complex operating environment of substations. The hardware composition, connection relationships, and installation locations of each module are as follows: The data acquisition module is the sensing layer of the device, responsible for acquiring all dimensions of the device's electrical and non-electrical operational data. It is divided into electrical quantity acquisition units and non-electrical quantity acquisition units, and the core hardware includes edge computing gateways, various sensors, and communication interface components.

[0029] The electrical quantity acquisition unit consists of: a voltage transmitter, a current transmitter, a switch quantity acquisition unit, and a power-specific communication protocol converter.

[0030] Connection relationship: The input terminals of the voltage and current transmitters are connected in parallel to the secondary output terminals of the existing PT (voltage transformer) and CT (current transformer) in the substation, and the output terminals are connected to the analog interface of the switch quantity acquisition device; the switch quantity acquisition device is connected to the edge computing gateway via RS485 bus; one end of the protocol converter is connected to the front-end communication interface of the OPEN3000 system via Ethernet, and the other end is connected to the edge computing gateway to realize bidirectional interaction of telemetry and remote signaling data. The communication protocol adopts the IEC60870-5-104 power standard protocol.

[0031] The hardware components of the non-electrical quantity acquisition unit include: an infrared array temperature sensor, an ultra-high frequency partial discharge sensor, a vibration acceleration sensor, and a data acquisition terminal.

[0032] Connection: The signal output terminals of various sensors are connected to the field data acquisition terminal via shielded cables. The acquisition terminal performs A / D conversion and preliminary filtering on the analog signals, and then transmits them to the edge computing gateway via industrial Ethernet. All sensors are powered by DC 24V, which is supplied by the DC power supply panel of the substation.

[0033] Edge computing gateway hardware selection: Adopting an industrial-grade ARM architecture edge gateway, equipped with 4 RS485 interfaces, 8 Ethernet interfaces, 1 dedicated power fiber interface, and built-in edge computing chip and 8GB local storage unit.

[0034] Connection relationship: It connects to all devices of electrical and non-electrical quantity acquisition units at the bottom, and connects to the big data analysis server and ledger server of the dispatch center at the top through the power dedicated fiber optic intranet. It also has local data caching capability and can store 72 hours of operation data when the network is interrupted.

[0035] Voltage and current transmitters: installed in the secondary instrument room of the substation switchgear, connected in parallel with the existing PT and CT secondary circuits, without changing the original wiring method.

[0036] Infrared array temperature sensor: For switchgear, it is installed on the inner wall of the circuit breaker contact chamber and busbar chamber, directly opposite the connection between the contact and the busbar; for transformers, it is installed on the bracket on the top of the transformer bushing and oil tank to monitor the bushing joint and oil temperature distribution; for GIS equipment, it is installed at the flange connection of the GIS air chamber.

[0037] Ultra-high frequency partial discharge sensor: It is built-in and installed inside the cable room of the switch cabinet or in the insulation basin of the GIS equipment to avoid external electromagnetic interference.

[0038] Vibration acceleration sensor: It is magnetically installed on the side wall of the transformer oil tank and the outer shell of the circuit breaker operating mechanism to monitor the vibration characteristics of the equipment during operation.

[0039] Edge computing gateway: Installed in the communication cabinet in the main control room of the substation, deployed in the same cabinet as the existing dispatch communication equipment, and connected to the DC power supply system in the station.

[0040] The real-time ledger update module is responsible for dynamically synchronizing basic equipment information, operation records, and maintenance records, solving the problem of static lag in traditional ledgers. The hardware consists of an RFID identification unit and a ledger synchronization server.

[0041] Hardware composition and connection relationship: RFID identification unit, hardware composition: anti-metal RFID electronic tag, handheld RFID reader / writer terminal, fixed RFID reader / writer.

[0042] Connection relationship: Fixed RFID readers are installed at the entrance and exit of the substation and access the local area network of the station via Ethernet, and communicate with the edge computing gateway; handheld RFID readers use the power wireless private network to access the system and can interact with the ledger server.

[0043] The ledger synchronization server uses a rack-mounted industrial server equipped with dual CPUs, 32GB of memory, a 4TB redundant disk array, and a dedicated Linux operating system for power applications.

[0044] Connection relationships: The system connects to the edge computing gateway, OPEN3000 system database, big data analysis server, and maintenance management system through the dedicated power intranet, enabling multi-source synchronization and sharing of ledger data.

[0045] RFID electronic tags: These are affixed to a prominent position on the outer casing of each primary device. The tags contain data such as the device's unique ID, basic parameters, and manufacturing information.

[0046] Fixed RFID reader: Installed on the pillars at the entrance and exit of the substation equipment area, 1.5m above the ground, covering the equipment access passage.

[0047] Handheld RFID reader / writer terminal: carried by maintenance and inspection personnel, with no fixed installation location.

[0048] Ledger synchronization server: Deployed in a standard rack in the dispatch center computer room of the municipal power supply company, in the same computer room as the dispatch system server.

[0049] The big data analysis and judgment module is the core computing unit of the device, responsible for data storage, feature mining and status judgment. It adopts a collaborative architecture of "centralized computing in the cloud + local preprocessing at the edge".

[0050] Hardware composition and connectivity: Edge preprocessing unit: integrated within the edge computing gateway, the gateway's built-in computing chip completes data cleaning, feature extraction, and initial threshold determination without requiring additional independent hardware.

[0051] The cloud analytics server cluster consists of: a 3-node distributed computing server, a storage disk array, and a 10 Gigabit data exchange switch.

[0052] Connection: Three computing servers form a distributed cluster via a 10 Gigabit Ethernet switch, sharing a storage disk array; the disk array adopts a RAID5 redundant architecture to store historical operating data, ledger data, and algorithm models; the cluster connects to the data acquisition module and ledger module through the power intranet, and simultaneously outputs analysis results to the early warning and work order modules.

[0053] The cloud analytics server cluster and disk array are both deployed in the information room of the municipal power supply company's dispatch center. They are deployed in Security Zone II in accordance with the three-level protection requirements for power information security, and are deployed in the same security zone as the OPEN3000 system to ensure the security of data interaction.

[0054] The intelligent early warning and work order generation module is responsible for issuing early warning information and automatically generating maintenance work orders. It is the output interaction unit of the device, realizing the automated issuance of operation and maintenance instructions.

[0055] The early warning release unit consists of the following hardware components: a large-screen early warning terminal for dispatching, a handheld smart terminal for maintenance personnel, and an industrial-grade audible and visual alarm.

[0056] Connection relationships: The dispatching large screen early warning terminal is connected to the dispatching center's large screen control system via an HDMI interface to synchronously display early warning information; the handheld smart terminal is connected to the system via the power mobile operation private network to receive pushed early warning and work order information; the audible and visual alarms are installed in the dispatching room and the maintenance team's duty room and are connected to the switch output interface of the early warning terminal.

[0057] The work order generation and output unit consists of the following hardware components: a web application server and a laser work order printer.

[0058] Connection structure: The web server is connected to the power intranet, obtains analysis results from the big data analysis module, and automatically generates standardized maintenance work tickets; the work order printer is deployed in the maintenance team and connects to the web server through the intranet to realize automatic printing of work orders.

[0059] Dispatch large screen early warning terminal: Deployed in the dispatch center control room and embedded in the existing dispatch large screen system.

[0060] Audible and visual alarms: installed on the walls of the dispatch room and the maintenance team's duty room respectively.

[0061] Web server and work order printer: The web server is deployed in the dispatch center's information room, and the printer is installed in the maintenance team's office.

[0062] The overall hardware connection of the device follows the hierarchical logic of the perception layer, edge layer, cloud layer, and application layer. The specific topology is as follows: various sensors, field acquisition terminals, edge computing gateway, power dedicated intranet, ledger synchronization server, big data analysis server cluster, intelligent early warning terminal, work order web server, dispatching screen, handheld terminal, and printer. At the same time, the OPEN3000 system front-end unit communicates bidirectionally with the edge computing gateway through a protocol converter to realize the acquisition of operation data and the transmission of early warning information.

[0063] The core judgment algorithm of this device is an improved combined weight fuzzy comprehensive evaluation model. Addressing the problems of subjective weight setting and large deviations in evaluation results in traditional equipment condition assessments, this model proposes to integrate the subjective weights of the Analytic Hierarchy Process (AHP) with the objective weights of the entropy weight method through the principle of minimum relative entropy. This results in a combined weight that takes into account both expert experience and data characteristics. Then, fuzzy comprehensive evaluation is used to calculate equipment health, providing a quantitative basis for graded early warning. This chapter provides a detailed derivation of the combined weight calculation and health assessment formulas.

[0064] First, a three-level evaluation system for the health status of power grid equipment is established: Target layer U: Overall equipment health; Criteria layer: Electrical operating status, temperature status, insulation status, and operating history status; Indicator layer: A total of 8 core indicators, including current deviation rate, voltage deviation rate, maximum temperature rise, temperature gradient, partial discharge amplitude, vibration acceleration, cumulative number of actions, and years of operation. Indicator standardization is necessary because the indicators have different dimensions and magnitudes. Standardization is required to map the indicators to the [0,1] interval. All indicators in this system are negative (larger values ​​indicate a worse state), and standardization is performed using the extreme value method.

[0065] In the formula, Let i be the standardized value of the i-th indicator for the j-th sample. The original value of the indicator. , These are the industry standard upper limit and ideal lower limit for this indicator, respectively. After standardization... The closer the value is to 1, the more severe the degradation of the indicator.

[0066] Traditional equipment condition assessments often rely on either subjective or objective weighting: subjective weighting depends on expert experience but is easily influenced by personal preferences, resulting in significant discrepancies in scores from different experts; objective weighting is entirely data-driven but tends to ignore industry maintenance experience, leading to weighting allocations that do not align with actual fault logic.

[0067] The core improvement of this algorithm lies in integrating subjective and objective weights based on the principle of minimum relative entropy. By minimizing the deviation between the two, a combined weight is obtained, which retains the expert experience of power operation and maintenance while fully reflecting the objective characteristics of the operation data. This effectively solves the bias problem of the single weight method and improves the accuracy of health assessment.

[0068] Let the subjective weight vector obtained through the analytic hierarchy process be... The objective weight vector obtained by the entropy weight method is: The combined weight vector to be solved is Satisfying the constraints and .

[0069] Step 1: Construct the objective function. Relative entropy (KL divergence) is used to measure the degree of difference between two probability distributions; the smaller the relative entropy, the higher the consistency between the two distributions. To ensure that the combined weights simultaneously align with both subjective and objective weights, the objective function is to minimize the sum of the relative entropies of the combined weights and the two types of weights.

[0070] Constraints:

[0071] Step 2: Construct the Lagrange function and solve the constrained optimization problem using the Lagrange multiplier method, introducing Lagrange multipliers. Construct the Lagrange function:

[0072] Step 3: Calculate the partial derivative and set it to 0. Then, with respect to the variables in the Lagrange function... Find the first-order partial derivative and set it to zero to obtain the extremum condition:

[0073] After simplification, we get:

[0074] Step 4: Refine and solve the weight expression by performing a logarithmic transformation on the above equation:

[0075]

[0076] Let the constant term Then the weight expression can be simplified to:

[0077] Step 5: Substitute the constraints to solve for the constant C, and then substitute the weight expression into the normalized constraints. We can obtain:

[0078]

[0079] Step 6: Obtain the final combined weight formula. Substitute the constant C back into the weight expression to obtain the final combined weight calculation formula:

[0080] This formula integrates subjective and objective weights through a geometric mean, and is rigorously derived using the principle of minimum relative entropy, achieving an optimal balance between subjective and objective information. Compared to traditional linear weighted fusion methods, this approach avoids the subjectivity of manually setting fusion coefficients, and the weight allocation more closely reflects the actual patterns of equipment degradation. This has been verified through actual testing. The accuracy of equipment health assessment can be improved by more than 15%.

[0081] Based on the combined weights, the fuzzy comprehensive evaluation method is used to calculate the final health score of the equipment. The specific process is as follows: Establish an evaluation set:

[0082] The corresponding score vector takes the median of each interval. .

[0083] The membership degree of each indicator to each evaluation level is calculated using the trapezoidal membership function, and a fuzzy relation matrix is ​​constructed. .

[0084] Combined weight vector Calculate the fuzzy evaluation vector .

[0085] The final overall health score of the equipment is:

[0086] The system uses a health score S to generate tiered alerts: S≥90 indicates a normal state; 70≤S<90 indicates a state of alert and triggers a yellow alert; 50≤S<70 indicates an abnormal state and triggers an orange alert; and S<50 indicates a severe state and triggers a red alert.

[0087] The complete workflow of this device can be divided into five stages, realizing full automation from data acquisition to maintenance closed loop. The specific process is as follows: After the device is started, the data acquisition module operates around the clock: the electrical quantity acquisition unit acquires telemetry and telecontrol data such as equipment current, voltage, power, and switch status in real time through the OPEN3000 system, with the sampling frequency consistent with the dispatch system; the non-electrical quantity acquisition unit acquires equipment temperature, partial discharge, and vibration data at a set frequency, with temperature sampling frequency of 1 time / minute and partial discharge and vibration sampling frequency of 1 time / 10 minutes. The acquired raw data is first preprocessed locally by the edge computing gateway, including outlier removal, filtering and noise reduction, and format standardization, and then cached in the local storage unit.

[0088] Meanwhile, the real-time ledger update module operates synchronously: when maintenance personnel conduct inspections, they read equipment tags and enter inspection information using handheld RFID terminals; after maintenance is completed, fixed readers automatically identify the equipment under maintenance and update the maintenance records synchronously; the ledger server obtains data such as the number of equipment actions and downtime from the OPEN3000 system in real time, and updates the equipment ledger information accordingly, ensuring that the ledger data is synchronized with the actual status of the equipment in real time, thus solving the problem of lag in traditional ledgers.

[0089] The edge computing gateway transmits pre-processed data encrypted through the dedicated power network to the big data analysis server in the dispatch center. Regular data transmission occurs every 5 minutes, while emergency and abnormal data is uploaded immediately. The cloud server categorizes and stores the data on a disk array by device ID, forming a database of device operation history. Simultaneously, it correlates and matches this data with the basic device information and maintenance history data from the ledger server to construct a complete lifecycle data archive for the equipment.

[0090] The big data analysis module calls the improved combined weight fuzzy comprehensive evaluation model every hour to calculate the health of all online devices: First, it extracts the average operating indicators of each device over the past hour and completes the standardization process; then, it calls the pre-stored combined weight parameters to calculate the current health score of the device; at the same time, it combines the health change trend over the past 7 days and uses linear regression to predict the health change over the next 7 days to identify potential deterioration trends.

[0091] When the equipment health score falls below the warning threshold, or is predicted to fall below the warning threshold within the next 7 days, the system automatically marks it as a potential hazard, generates an assessment report, and clarifies the abnormal indicators, causes of degradation, and fault risk level. Unlike traditional systems that can only identify real-time out-of-limit data, this device can achieve early prediction of potential faults, bringing the warning point forward to the early stages of equipment degradation.

[0092] The intelligent early warning module executes a tiered early warning strategy based on the analysis results: Normal state: No warning information is pushed; only the device health score is displayed on the scheduling screen.

[0093] Attention status (yellow warning): A yellow warning icon is displayed on the dispatch screen, and a warning message is pushed to the handheld terminals of the corresponding maintenance team, prompting them to increase the frequency of inspections.

[0094] Abnormal status (orange alert): An alert pop-up window appears on the dispatch screen, triggering the audible and visual alarm in the duty room, sending an alert SMS and APP message to the operations and maintenance manager, and generating maintenance suggestions simultaneously.

[0095] Severe situation (red alert): Immediately trigger the highest level audible and visual alarm, and simultaneously push emergency warning information to the dispatcher, operations and maintenance director, and supervisor in charge, prompting them to take immediate power outage measures.

[0096] All early warning information is synchronously transmitted back to the OPEN3000 system and displayed uniformly on the dispatch interface, eliminating the need for manual screening by dispatch personnel and significantly reducing workload.

[0097] For orange and higher alerts, the intelligent work order generation module automatically matches the equipment type, fault type, and standard maintenance procedures. It retrieves basic equipment parameters and historical maintenance records from the ledger server to automatically generate standardized maintenance work orders, including equipment information, fault description, maintenance items, safety measures, and personnel configuration. After automatic system verification, the work order is pushed to the maintenance management system, where maintenance personnel can directly print and execute it without manual data entry. This reduces work order generation time from over 30 minutes to less than 1 minute.

[0098] After the maintenance work is completed, the maintenance personnel enter the maintenance results, replaced parts, maintenance time and other information through handheld terminals, and update them to the ledger server and big data analysis module. The system recalculates the equipment health status, verifies the maintenance effect, and forms a complete closed loop of monitoring, analysis, early warning, maintenance and feedback, continuously optimizing the accuracy of the algorithm model.

[0099] Improved ledger management model: In response to the problems of static lag and inaccurate data in traditional equipment ledgers, a dynamic ledger update mechanism with RFID identification and multi-source data linkage is proposed. This mechanism links basic equipment information, operating data, maintenance records, and inspection information in real time, enabling dynamic and accurate management of ledger data and solving the pain point of the ledger being out of touch with the actual status.

[0100] Improved early warning mechanism: In response to the shortcomings of the traditional OPEN3000 system which only has post-alarm, a pre-alarm mechanism with trend prediction and hierarchical warning is constructed. Based on the trend of health changes, early deterioration risks of equipment are identified in advance, and warnings are pushed out according to risk level. The fault warning is upgraded from post-alarm to pre-judgment, which greatly reduces the failure rate.

[0101] Evaluation algorithm improvement: To address the problem of large subjective bias in traditional equipment status evaluation weights, a combined weight fusion algorithm based on minimum relative entropy is proposed. This algorithm combines the expert experience of the analytic hierarchy process with the data characteristics of the entropy weight method. Compared with the single weight method, the accuracy of health evaluation is improved by more than 15%, effectively reducing missed detections and false detections.

[0102] Architectural Innovation: The OPEN3000-compatible edge-cloud collaborative hardware architecture device adopts a two-level architecture of edge acquisition + cloud analysis. It can be seamlessly connected to the existing OPEN3000 system through a standard protocol converter without modifying the original scheduling system. It reuses existing PT / CT and other acquisition equipment, and only adds non-electrical quantity sensors and edge gateways. The transformation cost is low and the deployment cycle is short, realizing a smooth upgrade from traditional scheduling system to intelligent early warning.

[0103] Model Innovation: The fully automated operation and maintenance closed loop constructs a fully automated operation and maintenance closed loop that includes data collection, status assessment, hierarchical early warning, work order generation, and maintenance feedback. It integrates the scattered equipment monitoring, manual assessment, and manual invoicing into an integrated process, completely replacing the manual screening and manual work order mode. The operation and maintenance efficiency is improved by more than 70%, realizing the digital and intelligent transformation of power grid equipment operation and maintenance.

[0104] Application Innovation: The multi-dimensional state perception system integrates monitoring data from multiple dimensions, including electrical quantities, temperature, partial discharge, and vibration. It breaks through the limitations of traditional methods that only monitor electrical quantities, and characterizes the health status of equipment from multiple dimensions, including electrical performance, thermal condition, insulation condition, and mechanical condition. This allows for a more comprehensive identification of hidden equipment faults, especially for early-stage problems such as insulation degradation and mechanical wear, with a significantly improved accuracy rate.

[0105] This device enables early warning and precise assessment of equipment failures, effectively identifying early-stage potential problems, preventing power supply accidents caused by equipment deterioration, and significantly improving the reliability and safety of power grid equipment operation. Calculations show that after applying this device, the unplanned outage rate of primary equipment in substations can be reduced by more than 40%, effectively ensuring the safe and stable operation of the power grid.

[0106] On the one hand, precise condition-based maintenance, replacing traditional periodic maintenance, can extend the maintenance cycle of healthy equipment, avoid the waste of manpower and resources caused by over-maintenance, and reduce power outage maintenance time, thus reducing power outage losses. On the other hand, early detection of potential hazards can avoid significant economic losses caused by equipment failures and extend the service life of equipment. Calculations show that after applying this device to a single 110kV substation, the average annual operation and maintenance cost can be reduced by more than 25%, and the average service life of equipment can be extended by 3-5 years, resulting in significant economic benefits.

[0107] The device enables dynamic management of equipment records and an automated closed-loop operation and maintenance process, significantly reducing the workload of dispatching and maintenance personnel and improving the standardization and normalization of operation and maintenance management. Simultaneously, the equipment's full lifecycle data archive provides data support for power grid equipment selection and operation and maintenance strategy formulation, enhancing the refined management level of power companies.

[0108] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A power grid equipment maintenance early warning device, characterized in that, The complete operation and maintenance system consists of a data acquisition module, a real-time ledger update module, a big data analysis and judgment module, and an intelligent early warning and work order generation module. The data acquisition module is deployed at the edge layer of the substation, synchronously connecting to the OPEN3000 system to obtain electrical operation data, and simultaneously collecting non-electrical status data of the equipment. After preprocessing at the edge side, the data is encrypted and uploaded to the cloud. The real-time ledger update module is deployed in the cloud layer of the dispatch center, linking basic equipment parameters, real-time operating data and full maintenance records to build a dynamically updated equipment lifecycle data archive. The big data analysis and judgment module has a built-in combined weight fuzzy comprehensive evaluation model based on minimum relative entropy fusion, which quantifies the health of equipment and predicts the deterioration trend, thereby realizing the early identification of potential faults. The intelligent early warning and work order generation module executes tiered early warnings based on the health assessment results and automatically generates standardized maintenance work orders, forming a fully automated operation and maintenance closed loop of status monitoring, intelligent assessment, tiered early warning, work order execution, and effect feedback.

2. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The data acquisition module includes an electrical quantity acquisition unit and a non-electrical quantity acquisition unit. The electrical quantity acquisition unit connects to the OPEN3000 system front-end unit through an IEC60870-5-104 standard protocol converter to acquire telemetry and telesignal electrical operation data. The non-electrical quantity acquisition unit includes temperature, partial discharge, and vibration sensor arrays to simultaneously acquire equipment thermal status, insulation status, and mechanical status data.

3. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The real-time ledger update module is equipped with an RFID identification unit and a ledger synchronization server. The RFID identification unit is used to collect on-site inspection and maintenance operation information. The ledger synchronization server synchronizes OPEN3000 operating data, on-site collected data and maintenance management system data from multiple sources, replacing the traditional static ledger and realizing real-time linkage update of equipment status data.

4. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The combined weight fuzzy comprehensive evaluation model integrates the subjective weights of the analytic hierarchy process and the objective weights of the entropy weight method. Based on the principle of minimum relative entropy, it solves for the optimal combined weights, constructs a three-level evaluation system for equipment health status, and outputs a quantifiable equipment health score.

5. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The intelligent early warning and work order generation module is equipped with a four-level early warning mechanism, which is divided into four levels: normal, attention, abnormal and serious, based on the equipment health score. Each level has a different early warning push path and alarm triggering method. All early warning information is synchronously transmitted back to the OPEN3000 system for unified display.

6. The power grid equipment maintenance early warning device according to claim 5, characterized in that, For abnormal and higher warning levels, the intelligent warning and work order generation module automatically matches the equipment type and fault type, retrieves equipment ledger parameters and historical maintenance records, generates standardized maintenance work orders and pushes them to the maintenance management system, replacing the manual filling process.

7. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The substation edge layer is equipped with an industrial-grade edge computing gateway, which has the ability to cache local data and resume transmission after network outage; the dispatch center cloud layer adopts a distributed server cluster deployment and is set with the same security partition as the OPEN3000 system, meeting the requirements of the third level of power information security protection.

8. The power grid equipment maintenance early warning device according to claim 1, characterized in that, The overall hardware connection of the device follows the hierarchical logic of the perception layer, edge layer, cloud layer, and application layer.

9. The power grid equipment maintenance early warning device according to claim 8, characterized in that, The specific topology is as follows: sensor, field acquisition terminal, edge computing gateway, power dedicated intranet, ledger synchronization server, big data analysis server cluster, intelligent early warning terminal, work order web server, dispatching screen, handheld terminal, printer. At the same time, the OPEN3000 system front-end unit communicates bidirectionally with the edge computing gateway through a protocol converter to realize the acquisition of operation data and the transmission of early warning information.

10. The power grid equipment maintenance early warning device according to claim 1, characterized in that, This device employs an improved combined weight fuzzy comprehensive evaluation model, which integrates the subjective weights of the analytic hierarchy process (AHP) with the objective weights of the entropy weight method through the principle of minimum relative entropy. This results in a combined weight that takes into account both expert experience and data characteristics. The device's health is then calculated through fuzzy comprehensive evaluation, providing a quantitative basis for graded early warning.

Images