Realization method of regional cooperative anti-misoperation locking and medium

Abstract

The invention discloses a regional collaborative anti-misoperation locking implementation method and a medium, and the method comprises the steps: collecting regional power grid data, carrying out the preprocessing of the regional power grid data, obtaining a multi-source fusion data set, and constructing a dynamic anti-misoperation rule base based on the multi-source fusion data set; constructing an electrical island set based on the multi-source fusion data set, wherein each electrical island comprises all connected equipment nodes in the region; performing risk verification on the specific operation based on the live node data, the electrical island set and the dynamic anti-error rule base to generate a verification result, the verification result including operation permission or operation prohibition; and generating an operation instruction based on the verification result, wherein the operation instruction comprises an unlocking instruction or a locking instruction. Multi-source power grid data are collected and preprocessed, a dynamic anti-error rule base and an electrical island set are constructed, operation risk verification is completed in combination with multi-dimensional information, a corresponding instruction is generated, and global collaboration and dynamic pre-control can be achieved.

Description

Technical Field

[0001] This invention belongs to the field of power system safe operation technology, specifically relating to the implementation method and medium of regional collaborative anti-misoperation interlocking. Background Technology

[0002] With the expansion of power grid scale and the improvement of intelligence level, the unmanned operation and maintenance mode of substations combined with centralized monitoring is gradually becoming more widespread. However, existing substation anti-misoperation technologies such as microcomputer-based five-prevention systems, intelligent lock control systems, and access control locks generally have the following limitations:

[0003] 1. Strong system independence, insufficient collaboration: Each system operates independently without business association, making it impossible to achieve cross-site and cross-regional collaborative error prevention.

[0004] 2. The anti-misoperation logic is too simple and the ability to deal with complex scenarios is weak: It mainly targets the switching operation of primary equipment, and the anti-misoperation technology is insufficient for scenarios such as secondary equipment operation, temporary grounding wire operation, and forced interlocking of inter-station operation.

[0005] Therefore, there is an urgent need to develop a technology system for preventing errors that can achieve full-domain collaboration and dynamic pre-control. Summary of the Invention

[0006] The purpose of this invention is to obtain a multi-source fusion dataset by collecting regional power grid data, and to construct a method and medium for implementing regional collaborative anti-misoperation interlocking that can achieve full-domain collaboration and dynamic pre-control by combining it with a dynamic anti-misoperation rule base.

[0007] On the one hand, to achieve the above objectives, this invention proposes a method for implementing regional collaborative anti-misoperation interlocking, comprising: collecting regional power grid data and preprocessing the regional power grid data to obtain a multi-source fusion dataset, wherein the regional power grid data includes equipment operating status data, operation and maintenance information, and environmental data; constructing a dynamic anti-misoperation rule base based on the multi-source fusion dataset; constructing an electrical island set based on the multi-source fusion dataset, wherein each electrical island contains all connected equipment nodes within the region; performing risk verification on specific operations based on energized node data, the electrical island set, and the dynamic anti-misoperation rule base, generating verification results, wherein the verification results include permitted or prohibited operations; and generating operation instructions based on the verification results, wherein the operation instructions include unlocking instructions or interlocking instructions.

[0008] In one optional implementation, regional power grid data is collected and preprocessed to obtain a multi-source fusion dataset. Specifically, this includes: cleaning the regional power grid data to obtain valid data; performing spatiotemporal alignment on the valid data to generate structured data tuples, the structured data tuples including device addresses, topological coordinates, and timestamps; and standardizing the structured data tuples to obtain the multi-source fusion dataset.

[0009] In one optional implementation, a dynamic error prevention rule base is constructed based on the multi-source fusion dataset, specifically including: constructing a local rule base based on the multi-source fusion dataset, wherein the local rule base is set with a maximum cache capacity; constructing a rule priority model based on the historical trigger count, recent access time, and risk level of the local rule base; and if the current cache capacity reaches the maximum cache capacity, performing rule elimination and updating based on the rule priority model to obtain the dynamic error prevention rule base.

[0010] In one optional implementation, constructing an electrical island set based on the multi-source fusion dataset specifically includes: constructing an undirected topology graph of the power grid based on the multi-source fusion dataset, wherein the undirected topology graph of the power grid includes a set of device nodes and a set of electrical connection edges between nodes; dividing the undirected topology graph of the power grid into electrical islands using a disjoint-set data structure algorithm; and obtaining the electrical island set based on the electrical islands.

[0011] In one optional implementation, risk verification is performed on specific operations based on the acquired energized node data, the set of electrical islands, and the dynamic anti-misoperation rule base, and verification results are generated. Specifically, this includes: setting the set of energized nodes before operation; and performing specific operations on the operating equipment based on the dynamic anti-misoperation rule base.

[0012] Obtain the energized node data after the operation; obtain the node state after the operation based on the energized node data; re-divide the electrical island based on the node state after the operation to obtain the electrical island where the operating device is located; if the electrical island where the operating device is located has no intersection with the set of energized nodes before the operation, generate a verification result that allows the operation; if the electrical island where the operating device is located has an intersection with the set of energized nodes before the operation, generate a verification result that prohibits the operation.

[0013] In one optional implementation, generating an operation instruction based on the verification result specifically includes: constructing spatiotemporal constraint equations based on different time scales; constructing a dynamic adaptation model of the operation sequence and the power grid operating state based on the spatiotemporal constraint equations; using the dynamic adaptation model of the operation sequence and the power grid operating state to score the legality of the verification result, thereby obtaining a legality score result; and generating the operation instruction based on the legality score result, wherein the operation instruction further includes pushing an optimization scheme.

[0014] In one optional implementation, the spatiotemporal constraint equations include second-level operation constraint equations, minute-level mode adjustment constraint equations, and hour-level maintenance constraint equations.

[0015] In an optional implementation, the method for implementing the regional collaborative anti-misoperation interlocking further includes: constructing a global electrical topology map, which includes regional equipment nodes and cross-site connection edges; identifying topological risks of loop closing or loop opening operations based on the global electrical topology map, and calculating loop current risks and verifying protection mismatch risks to obtain risk identification results; and generating the operation command based on the risk identification results.

[0016] In an optional implementation, the regional collaborative anti-misoperation interlocking method further includes: judging the operation intent of the operation command through a multimodal behavioral intent recognition algorithm; if the intent is clear, outputting an instruction to allow entry to the next operation step; if the intent is ambiguous, triggering secondary confirmation; calculating the permission matching degree based on personnel qualifications, task type, and equipment risk level, and obtaining the matching degree calculation result; generating an authorization permission or insufficient permission instruction based on the matching degree calculation result; if an ambiguous intent or insufficient permission is detected, issuing an electronic interlocking instruction to lock the equipment and prohibit operation.

[0017] On the other hand, the present invention also proposes a computer storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any one of the regional cooperative anti-misoperation interlocking methods.

[0018] The beneficial effects of this invention are as follows: by collecting and preprocessing multi-source power grid data, constructing a dynamic anti-misoperation rule base and an electrical island set, and combining multi-dimensional information to complete operational risk verification and generate corresponding instructions, it can comprehensively cover risk factors such as equipment, operation, and environment, accurately intercept electrical misoperations, effectively ensure the safe operation of the regional power grid, and at the same time improve the scientific nature of operational decisions and the level of intelligence and collaboration in power grid operation and maintenance. Attached Figure Description

[0019] Figure 1 A flowchart illustrating a regional collaborative anti-misoperation interlocking method according to an embodiment of the present invention;

[0020] Figure 2 The flowchart illustrates a method for implementing a regional collaborative anti-misoperation interlocking mechanism, as provided in another embodiment of the present invention. Detailed Implementation

[0021] The present invention will be further described in detail below with reference to the accompanying drawings and specific embodiments.

[0022] like Figure 1As shown, according to an embodiment of the present invention, in one aspect, a method for implementing regional cooperative anti-misoperation interlocking is provided, comprising the following steps:

[0023] Step S101: Collect regional power grid data and preprocess the regional power grid data to obtain a multi-source fusion dataset. The regional power grid data includes equipment operating status data, operation and maintenance information, and environmental data.

[0024] Step S103: Construct a dynamic error prevention rule base based on the multi-source fusion dataset.

[0025] Step S105: Construct an electrical island set based on the multi-source fusion dataset, where each electrical island contains all connected device nodes within the region.

[0026] Step S107: Perform risk verification on specific operations based on energized node data, electrical island set, and dynamic anti-misoperation rule base, and generate verification results, including allowed or prohibited operations.

[0027] Step S109: Generate operation instructions based on the verification results. The operation instructions include unlocking instructions or locking instructions.

[0028] In this embodiment, a multi-dimensional data acquisition system is constructed for core assets such as substations, distribution lines, and switchgear covered by the regional power grid. By deploying various sensors, video monitoring equipment, and handheld terminals for operation and maintenance personnel in key equipment such as circuit breakers, disconnectors, and transformers within the substations and in operation and maintenance work areas, real-time data on equipment operating status, operation and maintenance information, and environmental data are collected.

[0029] Among them, equipment operation status data is acquired in real time through sensors deployed on key equipment such as transformers, circuit breakers, and disconnect switches, covering 12 core operating parameters including voltage, current, power factor, equipment temperature, and insulation status; operation and maintenance information is synchronously entered through the power grid operation and maintenance management system, including key information such as work plan, work order number, operator qualifications, operation steps, and power outage scope, realizing full traceability of the operation process; environmental data is collected through meteorological monitoring equipment and temperature and humidity sensors deployed along the power grid and in substations, covering environmental parameters such as wind speed, rainfall, temperature, humidity, and icing thickness, providing environmental impact basis for subsequent risk assessment.

[0030] The specific operations include:

[0031] Switching operations: opening and closing operations of circuit breakers and load switches.

[0032] Disconnecting switch operation: opening and closing operations of disconnecting switches (knife switches) (including grounding disconnecting switches).

[0033] Grounding operation: Installation and removal of temporary grounding wires.

[0034] Equipment status transition operations: such as a series of operations to change equipment such as transformers and busbars from "operation" status to "maintenance", "hot standby", or "cold standby" status (which need to be completed in conjunction with the operation of switches, disconnect switches, etc.).

[0035] These operations directly affect the energized state and topology of the power grid, and are the core control objects for preventing misoperation. They need to be verified through energized nodes, electrical island topology, and misoperation prevention rules to avoid dangerous operations such as opening and closing disconnecting switches under load or closing switches with the ground wire connected.

[0036] Regional power grid data is preprocessed to form a structured multi-source fusion dataset.

[0037] A rule library for error prevention is built based on a two-tier architecture of basic rules and dynamically adjusted rules, constructed using a multi-source fusion dataset. The basic rules cover mandatory safety regulations in the power industry, such as core safety operation principles like "prohibiting the operation of disconnecting switches under energized conditions," "prohibiting the operation of circuit breakers under load," and "power outage operations must disconnect the load side before disconnecting the power supply side." The dynamically adjusted rules are generated dynamically based on real-time data. For example, a temporary rule "prohibiting operation under load" is automatically triggered based on equipment operating status data such as "equipment temperature exceeds a threshold (e.g., transformer oil temperature ≥ 85℃)." An environmental adaptation rule "prohibiting live-line work on outdoor lines" is triggered based on environmental data such as "ice thickness ≥ 10mm." And an operation association rule "prohibiting entry into the area for operation" is triggered based on operation and maintenance information such as "there are un-energized devices within the work area."

[0038] By collecting data from multiple sources and using a dynamic anti-error rule base, and through composite logic verification, it can achieve full-dimensional prediction and real-time interception of operational risks, breaking through the limitations of traditional static rule bases. It supports cross-site and multi-device collaborative anti-error measures, effectively reducing the misjudgment rate in complex operation scenarios and improving decision response speed.

[0039] The collected data undergoes outlier removal, data cleaning, and time-series and spatial calibration of multi-source data.

[0040] Further, step S101: Collect regional power grid data and preprocess the regional power grid data to obtain a multi-source fusion dataset, specifically including the following steps:

[0041] Step S1011: Perform data cleaning on the regional power grid data to obtain valid data.

[0042] Step S1013: Perform spatiotemporal alignment on the valid data to generate structured data tuples, which include device address, topology coordinates, and timestamp.

[0043] Step S1015: Standardize the structured data tuples to obtain a multi-source fusion dataset.

[0044] In this embodiment, to address the issues of "high noise, disordered timing, and inconsistent format" in the collected data, a data cleaning and preprocessing workflow is executed. Key steps require quantification and precision control using formulas, as follows:

[0045] use The principle is to identify and remove invalid data. Let a certain type of data sample set be... Calculate the mean Standard deviation Where X is a sample set of current phasor or voltage data. For the sample set A single sample, The total number of samples in the sample set. The arithmetic mean of the sample set. The standard deviation of the sample set This is the threshold for identifying outliers. If... These are identified as outliers and removed to ensure data integrity.

[0046] By employing the 3σ principle to identify and eliminate invalid data, noise data caused by sensor malfunctions, communication interference, and other factors can be effectively filtered out, significantly improving the accuracy of data samples.

[0047] Spatiotemporal alignment includes temporal alignment and spatial alignment.

[0048] Time alignment: Achieve time synchronization, assuming the PTP master clock time is... , No. The local timestamp of each sensor is Define time deviation The synchronization accuracy is controlled by the following formula:

[0049] ;

[0050] in, , is the maximum permissible time deviation; if This will trigger sensor time recalibration.

[0051] Spatial alignment: Achieved through power grid GIS topology coordinates, binding each data entry with "device ID - topology coordinates". -timestamp ", forming structured data tuples ,in For equipment coordinates in the power grid GIS, Spatial alignment ensures spatial correlation of data across devices for device on / off status or current values.

[0052] By generating structured data tuples containing device addresses, topological coordinates, and timestamps, previously scattered multi-source data is precisely correlated in both time and space. For example, in power grid topology analysis, it can accurately correspond the operating status of different devices at the same point in time, providing spatiotemporally consistent multi-dimensional information for electrical island division and energized node identification, thus avoiding analysis errors caused by spatiotemporal deviations.

[0053] The cleaned and aligned data is encapsulated into JSON format according to the IEC 61850 standard, defining a general field structure: D_{\text{std}}=\{\text{type: data type,device_id: device ID,ts: timestamp,param: core parameter,topo: topology coordinates}\}. For example, device status data is encapsulated as: {"type":"switch_status","device_id":"CB-101","ts":"2025-XX-XX10:00:00.001","param":"shutdown","topo":"(100, 200,30)"}, ensuring that the edge layer can directly parse and call it.

[0054] The cleaned and aligned data is encapsulated in JSON format and sent to the edge intelligence processing module. The JSON format defines the field structure as follows: type: data type, device_id: device ID, ts: timestamp, param: core parameters, topo: topology coordinates}, for example, device status data is encapsulated as: {"type":"switch_status", "device_id":"CB-101","ts":"2025-XX-XX 10:00:00.001","param":"shutdown","topo":"(100, 200, 30)"}, ensuring that the edge layer can directly parse and call it.

[0055] Standardizing structured data tuples unifies the format, units, and encoding rules of multi-source data, enabling direct fusion analysis of data from different sources (such as equipment operation, maintenance, and environmental monitoring). For example, standardized equipment temperature data (°C) and ambient humidity data (%) can be used for joint calculations across data types in a dynamic error prevention rule base, improving the accuracy of the rules.

[0056] Further, step S103 involves constructing a dynamic error prevention rule base based on the multi-source fusion dataset, specifically including the following steps:

[0057] Step S1031: Build a local rule base based on the multi-source fusion dataset, and set the maximum cache capacity of the local rule base.

[0058] Step S1033: Construct a rule priority model based on the historical trigger count, recent access time, and risk level of the local rule base.

[0059] Step S1035: If the current cache capacity reaches the maximum cache capacity, rule elimination and update are performed based on the rule priority model to obtain a dynamic anti-error rule base.

[0060] Dynamic caching adapts to the site's domain's error prevention rules, supports incremental updates of scheduling layer policies, and ensures that basic interlocking is available when communication is interrupted.

[0061] Set the local rule base cache capacity A set of rules Each rule It contains 3 attributes:

[0062] Historical trigger counts; high-frequency trigger rules have higher priority.

[0063] (Recent access time, recent access rules have higher priority).

[0064] Risk level, high-risk rules Regular rules .

[0065] Define rule priority The weighted sum of the three is calculated using the following formula:

[0066] ;

[0067] in , , All of these are weighting coefficients, determined through training with historical data, and satisfy the following conditions: =1; This represents the maximum number of times all rules can be triggered. The rule is the duration of the last unaccessed session. Hours is the longest cache period for a rule.

[0068] The local rule base has a cache capacity of up to 1,000 rules, which can store a large number of error prevention rules, covering a variety of operation scenarios and risk types. At the same time, through a priority model and eviction mechanism, it can still ensure the efficiency of rule calling and updating on the basis of large capacity, avoid system lag caused by too many rules, and achieve a balance between comprehensiveness and efficiency.

[0069] By constructing a rule priority model based on historical trigger counts, recent access times, and risk levels, rules that are frequently triggered, recently accessed, and high-risk can be given higher priority. For example, rules such as "operating a disconnect switch while the circuit is energized" in a substation, which are high-risk and frequently triggered, will be prioritized for verification, ensuring that key risks are controlled first and effectively improving the accuracy of operational risk identification.

[0070] The dynamic rule base for preventing errors employs a rule elimination and update mechanism.

[0071] When the scheduling layer pushes incremental rules , The maximum cache capacity of the LRU rule base, if the current cache capacity is... Then the rule with the lowest elimination priority will be eliminated. Ensure that caching rules are always adapted to high-frequency and high-risk scenarios within this site domain.

[0072] The formula for the time consumption of incremental update of a single rule is as follows:

[0073] ;

[0074] To meet the requirement of a total computation time of ≤100ms for 1000 rules and to adapt to the lightweight computing capabilities of industrial-grade processors, the computation time for single rule priority is specified. Elimination or addition of time-consuming Substituting these values, we obtain the total update time for a single incremental rule. This meets the incremental update response requirements.

[0075] When the cache capacity reaches its limit, the lowest priority rules are evicted and incremental rules are updated, ensuring that the rule base always adapts to high-frequency, high-risk scenarios within the domain. For example, after power grid equipment upgrades or changes in operational scenarios, new incremental rules can be promptly integrated into the rule base, while old, low-priority, and poorly adaptable rules are evicted, guaranteeing the timeliness and relevance of the rule base.

[0076] The formula for incremental update time for a single rule ensures that the total computation time is ≤100ms, adapting to the lightweight computing capabilities of industrial-grade processors. Taking 1000 rules as an example, the total update time can be controlled within a reasonable range, enabling rapid response to changes in the power grid's operating status, achieving real-time rule updates, and ensuring the timeliness of operation verification.

[0077] Dynamic updates and priority management of the rule base ensure that high-risk and high-frequency operations receive continuous and focused attention, effectively reducing the probability of erroneous operations. For example, in complex power grid switching operations, a dynamic anti-misoperation rule base can accurately identify operations that violate safety regulations and promptly lock them out, providing a higher rate of erroneous operation interception compared to a static rule base.

[0078] Basic interlocking is available during communication interruptions, ensuring that even in extreme situations such as network failures, basic anti-misoperation functions can still be achieved by relying on the local rule base, avoiding operational loss of control due to communication problems, and improving the overall reliability of power grid operation.

[0079] Further, step S105, constructing an electrical island set based on the multi-source fusion dataset, specifically includes the following steps:

[0080] Step S1051: Construct an undirected topology graph of the power grid based on the multi-source fusion dataset. The undirected topology graph of the power grid includes a set of device nodes and a set of electrical connection edges between nodes.

[0081] Step S1053: Divide the undirected topology graph of the power grid into electrical islands using the disjoint-set data structure algorithm;

[0082] Step S1055: Obtain the set of electrical islands based on the electrical islands.

[0083] In this embodiment, a lightweight topology analysis algorithm is used to quickly verify the anti-misoperation rules of the local domain at the edge, with the response time controlled within 200ms.

[0084] Abstracting the power grid topology as an undirected graph Where V is the set of device nodes, and each node status (0 = open / isolate, 1 = close / connect); where E is the set of electrical connection edges between nodes (such as the connection between the busbar and the disconnecting switch). For nodes and The weight of the connecting edges, for each edge weight This indicates a valid connection. This indicates a disconnection.

[0085] The electrical islands (connected regions) are partitioned using the Union-Find algorithm, and the merging operation formula is as follows:

[0086] ;

[0087] in, For unjoined set union operations, , It is a node The real-time status of opening or closing the circuit breaker ultimately yields the electrical island assembly. Each electrical island It includes all connected nodes in the region.

[0088] By partitioning the undirected topology graph of the power grid using the disjoint-set data structure algorithm, it is possible to accurately identify sets of electrically connected equipment nodes (electrical islands). For example, it can accurately determine whether a certain line and a substation busbar are located on the same electrical island, providing a core basis for identifying energized nodes and defining the operational scope, avoiding misjudgments caused by errors in topology analysis, and thus improving the accuracy of operational risk verification.

[0089] The opening and closing operations of power grid equipment change the topology in real time. This solution can dynamically update the electrical island partitioning results based on the real-time status (open / closed) of equipment nodes and the weights of connecting edges. For example, when a disconnector is closed, it can quickly integrate related equipment into the same electrical island, ensuring that the electrical island set is completely synchronized with the actual operating state of the power grid, providing a real-time and accurate topology basis for subsequent operation verification.

[0090] Further, in step S107, based on the acquired energized node data, electrical island set, and dynamic anti-misoperation rule base, a risk verification is performed on the specific operation to generate a verification result, which specifically includes the following steps:

[0091] Step S1071: Set the set of charged nodes before operation.

[0092] Step S1073: Perform specific operations on the operating equipment based on the dynamic anti-misoperation rule base.

[0093] Step S1075: Obtain the data of the charged node after the operation.

[0094] Step S1077: Obtain the node state after the operation based on the charged node data.

[0095] Step S1079: Re-divide the electrical island based on the node status after the operation to obtain the electrical island where the operating device is located.

[0096] Step S1081: If the electrical island where the operating device is located has no intersection with the set of energized nodes before the operation, then generate a verification result that allows the operation.

[0097] Step S1083: If the electrical island where the operating device is located intersects with the set of energized nodes before the operation, a verification result prohibiting the operation is generated.

[0098] In this embodiment, the set of charged nodes before operation is assumed to be: The data is obtained from the SCADA system, and the operating device is... After the operation, the node state becomes The electrical islands were re-divided to obtain If the operational objective is "to open the isolating switch (isolate and de-energize)," the following conditions must be met. ,in, Represents an electric island State of being charged ), Represents an electric island Power outage status. If the condition is met... If the operation is not allowed, output "Operation permitted"; otherwise, output "Operation prohibited" and indicate the reason for the risk.

[0099] After path compression optimization, the time complexity of a single merge and search operation in the disjoint-set data structure algorithm is O(log n). ,in This is the inverse function of the Ackermann function, approximately a constant. Traversal... Each node and Strip edge (in power grid topology) The total time complexity is:

[0100] ;

[0101] like If the calculation time is ≤3ms when there are multiple nodes, then the requirement of single rule verification ≤5ms is met.

[0102] This rule is a verification standard for "power outage isolation safety" in power operations. Its core is to determine the intersection between the electrical island and the original energized area to ensure that the target area is truly de-energized after the "isolation power outage" operation, thus avoiding the risks of live operation.

[0103] Based on step S109, an operation instruction is generated based on the verification result, which specifically includes the following steps:

[0104] Step S1091: Construct spatiotemporal constraint equations based on different time scales.

[0105] Step S1093: Construct a dynamic adaptation model of operation sequence and power grid operation state based on spatiotemporal constraint equations.

[0106] Step S1095: Use a dynamic adaptation model of operation sequence and power grid operation status to score the legality of the verification result and obtain the legality score result.

[0107] Step S1097: Generate operation instructions based on the legality score results. The operation instructions also include pushing optimization solutions.

[0108] By constructing spatiotemporal constraint equations and dynamic adaptation models, the validity of verification results is scored and operation instructions with optimized solutions are generated. This enables precise and dynamic generation of operation instructions, improves the accuracy of operation validity judgment, and pushes optimized solutions to assist decision-making, effectively ensuring the safe and efficient operation of the power grid and promoting the intelligent and refined upgrading of power grid operation management.

[0109] Among them, differentiated constraint equations are designed for different time scales to achieve mathematical matching between operation and grid status. The spatiotemporal constraint equations include second-level operation constraint equations, minute-level mode adjustment constraint equations, and hour-level maintenance constraint equations.

[0110] Second-level operation constraint equation: with "device operation timing and topological connectivity" as the core variables, the expression is:

[0111] ;

[0112] in, Score the legality of the operation ( For it to be valid, T=100). Timestamps for operations performed in seconds; Let Topo(t) be the real-time status of the device at time t (opening = 0, closing = 1), and let Topo(t) be the topological connectivity at time t (isolation = 0, connectivity = 1). and These are the weighting coefficients (values ​​obtained during training based on historical operation data). , ).

[0113] By quantitatively scoring the real-time status of equipment and topology connectivity, the legality of operations can be accurately determined on a second-level timescale. For example, in circuit breaker opening and closing operations, it is possible to verify in real time whether the operation sequence conforms to the topology connectivity rules, avoiding accidents such as short circuits and load tripping caused by misoperation, and improving the accuracy of second-level operation legality judgment.

[0114] Adjusting constraint equations at the minute-level: Using "power flow parameter and operating mode compatibility" as the core variable, the expression is:

[0115] ;

[0116] in, Adjust the risk value according to the method (≤L is safe, L=80); ; for Flow deviation rate during the period ( =5 minutes); It is a compatibility score for the operating mode (match = 0, conflict = 1); Weighting coefficients ( =0.7、 =0.3).

[0117] Quantitative analysis based on power flow deviation rate and operating mode compatibility can identify risks arising from operating mode adjustments within a minute-level time window. For example, during power flow adjustments, it can promptly detect excessive power flow deviations or operating mode conflicts, preventing power grid stability issues caused by improper mode adjustments.

[0118] Hourly maintenance constraint equation: with "maintenance scope, time window, and safety measures" as core variables, the expression is:

[0119] ;

[0120] in, The maintenance plan is scored for compliance (≤K indicates feasibility, K=90); K is the feasibility threshold for the compliance score. The isolation degree between the maintenance scope and the energized area within the time period ΔT of the hourly maintenance plan. =1 hour); Time( This refers to the matching degree between maintenance time and the off-peak period of power grid load; For the completeness of safety measures; , , Weighting coefficients ( =0.4、 =0.3、 =0.3).

[0121] By comprehensively evaluating the isolation of the maintenance scope, the matching degree between the time window and the off-peak load, and the completeness of safety measures, the system can ensure that the maintenance plan does not affect the normal power supply of the grid while guaranteeing maintenance safety. For example, it can accurately match the maintenance time with the off-peak load period of the grid, reducing the impact of maintenance on users' electricity consumption. At the same time, the completeness score of safety measures reduces the personal and equipment risks during the maintenance process.

[0122] Second-level operation: After receiving the device operation command, the scheduling layer calls the second-level data and constraint equation in real time to calculate the operation legality score S(t). If S(t)≤T, the operation is allowed, and the locking or unlocking command is sent to the edge layer synchronously. The response time is ≤200ms.

[0123] Minute-level adjustment: When the EMS system pushes a new power flow section, it automatically substitutes the constraint equations to calculate M(Δt). If M(Δt)≤L, it generates a mode adjustment suggestion (such as "the neutral point grounding switch of the main transformer is closed"). If M(Δt)>L, it marks the risk point (such as "the output of the distributed power source is too high, causing the power flow to exceed the limit") and pushes the optimization solution.

[0124] Hourly maintenance plan: After the maintenance plan is submitted, the power grid status during the ΔT period is simulated using a digital twin model. H(ΔT) is calculated by substituting it into the constraint equation. If H(ΔT)≤K, the plan is confirmed to be feasible. If H(ΔT)>K, the maintenance scope (such as reducing the power outage interval) or time window (such as avoiding peak load periods) is adjusted.

[0125] Differentiated constraint equations were designed for operation scenarios at different time scales of seconds, minutes, and hours, realizing digital constraints for the entire process from real-time operation to planned maintenance. This transformed power grid operation management from a traditional experience-driven model to a data-driven and model-driven intelligent management model, improving the overall level of management precision.

[0126] By transforming the matching relationship between operations and grid conditions into mathematical equations, the legality and risk value of operations are quantitatively scored, and the accuracy of decision-making is greatly improved, reducing the operational risks caused by human judgment errors.

[0127] Furthermore, the implementation method of regional collaborative anti-misoperation interlocking also includes the following steps:

[0128] Step S201: Construct a global electrical topology diagram, which includes regional device nodes and cross-site connection edges.

[0129] Step S203: Identify the topology risks of loop closing or loop opening operations based on the global electrical topology diagram, calculate the loop current risk and verify the protection mismatch risk, and obtain the risk identification results.

[0130] Step S205: Generate operation instructions based on the risk identification results.

[0131] Construct a global electrical topology diagram ,in For all equipment nodes in the region, To connect edges across stations, a graph neural network (GNN) algorithm is used to identify topological risks associated with loop merging or de-looping operations.

[0132] The variable definitions are shown in Table 1.

[0133] Table 1 Variable Definitions

[0134] Variable symbol Variable meaning Value range / type Total impedance of the cross-domain loop (Ω) [0.1,10] Voltage difference across the loop (kV) [0,10] Expected loop current (kA) [0,1.5] Rated current of the line (kA) [0.5,1.2] Action time (s) of one-sided protection device [0.1,1] The activation time (s) of the protection device on the other side [0.1,1]

[0135] Loop current risk calculation. Before closing the loop, calculate the expected loop current to determine if there is an overload. The formula is as follows:

[0136] ;

[0137] in, For the current threshold, if The network was identified as a "high-risk electromagnetic ring network," and ring-closing operations were prohibited.

[0138] Protection mismatch risk calibration: Verify the time-limit coordination of cross-site protection devices to avoid protection failure to operate. The formula is as follows:

[0139] ,in It is the minimum protection time limit difference.

[0140] like If the time limit is met, then the calculation result is considered satisfactory. If the protection is mismatched, the setting needs to be adjusted.

[0141] Furthermore, the implementation method of regional collaborative anti-misoperation interlocking also includes the following steps:

[0142] Step S301: The operation intent of the operation command is judged by the multimodal behavior intent recognition algorithm. If the intent is clear, the command to allow the next operation is output; if the intent is ambiguous, a second confirmation is triggered.

[0143] Step S303: Calculate the permission matching degree based on personnel qualifications, task type, and equipment risk level, and obtain the matching degree calculation result.

[0144] Step S305: Generate authorization or insufficient authorization instructions based on the matching degree calculation results.

[0145] Step S307: If the intent is ambiguous or the permissions are insufficient, an electronic interlock command is issued to lock the device and prohibit operation.

[0146] The system executes decisions issued by the edge intelligent processing module and records feedback to the edge intelligent processing module to form a closed-loop optimization, including multimodal behavior intent recognition function and execution control feedback function.

[0147] 1) Multimodal behavioral intent recognition algorithm:

[0148] After the scheduling decision module outputs the decision instruction, the human-computer interaction module accurately judges the operation intention through a multimodal behavior intention recognition algorithm, thereby providing effective operation guidance.

[0149] The confidence score of the operation intent is calculated by weighted summation and fusion of multimodal features, as shown in the following formula:

[0150] ;

[0151] in, This is a location characteristic value; the closer to the target device, the larger the value; it is 0 when the distance is ≥2m. The action feature value is the cosine similarity to the standard operation action, with 100 indicating a perfect match; The three values ​​are: speech feature value, text matching degree with preset command, 100 for perfect match, and the range of values ​​for all three is [0, 100]. These are feature weight coefficients, trained based on operation logs from the past three years, prioritizing action and speech features. Confidence of operational intent This is the intent determination threshold, with a value of 95.

[0152] like If the intent is deemed "clear", the output will be "allow proceed to the next step". If the intent is deemed "ambiguous," a second confirmation is triggered to prevent accidental operation.

[0153] Dynamic permission adaptation algorithm:

[0154] Based on personnel qualifications, task type, and equipment risk, the system calculates permission matching degree to prevent "unqualified personnel from operating high-risk equipment".

[0155] The permission matching degree is calculated by weighted fusion of qualifications, task adaptability, and device risk, using the following formula:

[0156] ;

[0157] in For personnel qualification levels, the work supervisor = 100, operator = 80, supervisor = 60, and no qualification = 0. For task type fit, exact match = 100, partial match = 50, no match = 0; For equipment risk coefficients, main transformer and busbar = 100, line = 80, distribution network equipment = 60, and no risk = 0. This is a weighted coefficient for access permissions, prioritizing qualifications and equipment risk. =0.4, =0.3, =0.3. The permission matching degree is calculated based on three dimensions; the higher the value, the more suitable the permission is. This is the access threshold; operations are prohibited below this value. The threshold for high-risk devices is 80, and for regular devices it is 60. To ensure that high-risk equipment is operated by highly qualified personnel, a risk-matching mechanism is established between qualifications and equipment.

[0158] If the target of the operation is a high-risk device ( ),need Only grant permissions if the target device is a regular device ( ), need Only grant permissions if; If the system fails to grant sufficient permissions, the electronic interlock will be triggered, prohibiting operation.

[0159] 3) Execution feedback submodule:

[0160] Based on the core algorithm output, the entire process of "guidance-execution-feedback" is controlled. If "ambiguous intent" or "insufficient permissions" are detected, an electronic lockout (e-Lock) command is issued to lock the device and prohibit operation. The execution log is archived according to the content of "user_id-operation time-device ID-intent confidence-permission matching degree-execution result". Every quarter, a series of key optimization information is extracted from the log, including the deviation between the topology verification result and the actual state, the deviation of the device state timing synchronization, and the deviation between the actual trigger frequency of the rule and the cache priority. This forms an edge algorithm optimization dataset, which is input into the algorithm optimization unit of the edge intelligent processing module to correct the topology partitioning parameters of the lightweight topology analysis algorithm and the priority weight of the LRU dynamic rule adaptation algorithm in the edge intelligent processing module. The optimized algorithm parameters are redeployed to the edge layer to support subsequent local logic verification and rule caching, forming a closed loop of execution feedback, algorithm optimization, and accuracy improvement.

[0161] The system of this invention comprises a region perception module, an edge intelligence processing module, a scheduling decision module, and a human-computer interaction module. The method flow based on the system composition is as follows: Figure 2 As shown.

[0162] The regional sensing module, through distributed sensing and communication technologies, is responsible for collecting a series of secondary system information, including the real-time switching status of primary equipment, signals from integrated protection devices, the communication status of smart terminals (GOOSE), and sampled values ​​from merging units. It establishes a mathematical correlation model between the power grid topology and anti-misoperation blocking rules, and analyzes the system operation mode.

[0163] The edge intelligent processing module, deployed in the regional control center, uses a lightweight topology analysis algorithm to achieve multimodal data fusion, standardized processing, and local logical verification. It also establishes a local library of error prevention rules based on the LRU algorithm, caches dynamic rules, and supports dynamic loading and incremental updates of global policies in the scheduling layer, ensuring the availability of basic interlocking functions in communication interruption scenarios.

[0164] The scheduling decision module, deployed in the main scheduling center, is based on the global collaboration engine of the cloud platform. It constructs a multi-time-scale anti-misoperation coordination mechanism, dynamically models the electrical connection relationships of substations, transmission lines, distribution networks and power plants, realizes cross-regional collaborative decision-making and risk identification, and can integrate quantum secure communication.

[0165] The human-computer interaction module includes a standard operating procedure generator for guiding operations, multimodal sensors for behavior perception, an electronic interlocking device for command execution, a digital twin simulator for pre-rehearsal, and a knowledge graph library, supporting collaborative decision-making and human-computer interaction.

[0166] Existing technologies typically lack cross-site coordination capabilities, resulting in low correlation between different substations and an inability to achieve coordinated interlocking of cross-site operations. Furthermore, they fail to perceive and control the impact of one substation's operation on other substations. This invention, through a regional sensing module, breaks down data silos, aggregates the real-time operating status of multiple substations and various types of equipment, and achieves "global coordinated interlocking," thereby improving regional safety redundancy.

[0167] Existing technologies have simple and static logic, with error prevention rules mainly based on fixed operating procedures and hard-coded equipment structures. This makes them inflexible in responding to frequent changes in power grid operation modes, the integration of new energy sources, and the application of new equipment. This invention, through dynamic rule caching in the edge processing module and dynamic loading of global strategies in the scheduling layer, enables the error prevention logic to adaptively adjust according to real-time power grid operation modes and topology changes.

[0168] Existing technologies have failed to effectively compensate for human limitations. This invention, by monitoring the behavior of operators in real time and providing early warnings, prompts, or even mandatory dynamic locking interventions at critical points, nipps potential risks in the bud.

[0169] On the other hand, the present invention also proposes an electronic device for setting up a data processing module. The electronic device includes: at least one processor; a memory communicatively connected to the at least one processor; wherein the memory stores instructions executable by the at least one processor, the instructions being executed by the at least one processor to enable the at least one processor to execute any one of the regional cooperative anti-misoperation interlocking implementation methods.

[0170] On the other hand, the present invention also proposes a computer storage medium storing a computer program, wherein the computer program, when executed by a processor, implements any one of the regional cooperative anti-misoperation interlocking methods.

[0171] Those skilled in the art will understand that implementing all or part of the processes in the above embodiments can be accomplished by a computer program instructing related hardware, and can be stored in a non-volatile computer-readable storage medium. When executed, the program can include the processes of the embodiments of the above methods. Any references to memory, storage, databases, or other media used in the embodiments provided in this application can include non-volatile and / or volatile memory. Non-volatile memory may include read-only memory (ROM), programmable ROM (PROM), electrically programmable ROM (EPROM), electrically erasable programmable ROM (EEPROM), or flash memory. Volatile memory may include random access memory (RAM) or external cache memory. By way of illustration and not limitation, RAM is available in various forms, such as Static RAM (SRAM), Dynamic RAM (DRAM), Synchronous DRAM (SDRAM), Dual Data SDRAM (DDRSDRAM), Enhanced SDRAM (ESDRAM), Synchlink DRAM (SLDRAM), Rambus Direct RAM (RDRAM), Direct Memory Bus Dynamic RAM (DRDRAM), and Memory Bus Dynamic RAM (RDRAM). The various embodiments described in this specification are presented in a progressive manner, and similar or identical parts between embodiments can be referred to interchangeably. Each embodiment focuses on describing the differences from other embodiments. In particular, for embodiments of apparatus, devices, and non-volatile computer storage media, since they are substantially similar to the method embodiments, the description is relatively simple, and relevant parts can be referred to the description of the method embodiments.

[0172] The above embodiments are merely illustrative examples and are not intended to limit the implementation. Those skilled in the art will recognize that other variations or modifications can be made based on the above description. It is neither necessary nor possible to exhaustively list all possible implementations. However, obvious variations or modifications derived therefrom are still within the scope of protection of this invention.

Claims

1. A method for implementing regional collaborative anti-misoperation interlocking, characterized in that, include: Collect regional power grid data and preprocess the regional power grid data to obtain a multi-source fusion dataset. The regional power grid data includes equipment operating status data, operation and maintenance information, and environmental data. A dynamic error prevention rule base is constructed based on the aforementioned multi-source fusion dataset; Based on the multi-source fusion dataset, an electrical island set is constructed, and each electrical island contains all connected device nodes within the region; Based on the charged node data, the set of electrical islands and the dynamic anti-misoperation rule base, risk verification is performed on specific operations to generate verification results, which include allowed or prohibited operations. An operation instruction is generated based on the verification result, and the operation instruction includes an unlock instruction or a lock instruction.

2. The method for implementing regional cooperative anti-misoperation interlocking according to claim 1, characterized in that, Collect regional power grid data and preprocess the regional power grid data to obtain a multi-source fusion dataset, specifically including: The regional power grid data is cleaned to obtain valid data; The valid data is spatiotemporally aligned to generate structured data tuples, which include device address, topology coordinates and timestamp; The structured data tuples are standardized to obtain the multi-source fusion dataset.

3. The method for implementing regional cooperative anti-misoperation interlocking according to claim 1, characterized in that, A dynamic error prevention rule base is constructed based on the aforementioned multi-source fusion dataset, specifically including: A local rule base is constructed based on the multi-source fusion dataset, and the local rule base is set with a maximum cache capacity. A rule priority model is constructed based on the historical trigger count, recent access time, and risk level of the local rule base. If the current cache capacity reaches the maximum cache capacity, rules are eliminated and updated based on the rule priority model to obtain the dynamic error prevention rule base.

4. The method for implementing regional cooperative anti-misoperation interlocking according to claim 1, characterized in that, An electrical island set is constructed based on the aforementioned multi-source fusion dataset, specifically including: A power grid undirected topology graph is constructed based on the multi-source fusion dataset. The power grid undirected topology graph includes a set of device nodes and a set of electrical connection edges between nodes. The undirected topology graph of the power grid is divided into electrical islands using a disjoint-set data structure algorithm. The set of electrical islands is obtained based on the electrical islands.

5. The method for implementing regional cooperative anti-misoperation interlocking according to any one of claims 1 to 4, characterized in that, Based on the acquired energized node data, the electrical island set, and the dynamic anti-misoperation rule base, risk verification is performed on specific operations to generate verification results, specifically including: Set the set of energized nodes before operation; Specific operations are performed on the operating equipment based on the aforementioned dynamic anti-misoperation rule base; Acquire the data of the energized nodes after the operation; The node state after the operation is obtained based on the charged node data; Based on the node status after the operation, the electrical island is re-divided to obtain the electrical island where the operating device is located; If the electrical island where the operating device is located has no intersection with the set of energized nodes before the operation, a verification result indicating that the operation is permitted is generated. If the electrical island where the operating device is located intersects with the set of energized nodes before the operation, a verification result prohibiting the operation is generated.

6. The method for implementing regional cooperative anti-misoperation interlocking according to claim 5, characterized in that, Based on the verification result, an operation instruction is generated, specifically including: Spatiotemporal constraint equations are constructed based on different time scales; A dynamic adaptation model between the operation sequence and the power grid operating state is constructed based on the aforementioned spatiotemporal constraint equations. The validity of the verification result is scored using a dynamic adaptation model between the operation sequence and the power grid operating state, resulting in a validity score. The operation instruction is generated based on the legality score result, and the operation instruction also includes pushing optimization scheme.

7. The method for implementing regional cooperative anti-misoperation interlocking according to claim 6, characterized in that, The spatiotemporal constraint equations include second-level operation constraint equations, minute-level mode adjustment constraint equations, and hour-level maintenance constraint equations.

8. The method for implementing regional cooperative anti-misoperation interlocking according to any one of claims 1 to 4, characterized in that, Also includes: Construct a global electrical topology diagram, which includes regional device nodes and cross-site connection edges; Based on the global electrical topology diagram, topology risks of loop closing or loop opening operations are identified, loop current risks and protection mismatch risks are calculated, and risk identification results are obtained. The operation instructions are generated based on the risk identification results.

9. The method for implementing regional cooperative anti-misoperation interlocking according to any one of claims 1 to 4, characterized in that, Also includes: The operation intent of the operation instruction is determined by a multimodal behavior intent recognition algorithm. If the intent is clear, an instruction to proceed to the next operation step is output. If the intent is ambiguous, a second confirmation is triggered; Based on personnel qualifications, task type, and equipment risk level, the permission matching degree is calculated to obtain the matching degree calculation result. Based on the matching degree calculation result, an authorization permission or insufficient permission instruction is generated; If the intent is ambiguous or the permissions are insufficient, an electronic interlock command is issued to lock the device and prevent operation.

10. A computer storage medium, characterized in that, The system contains a computer program that, when executed by a processor, implements the regional cooperative anti-misoperation interlocking method as described in any one of claims 1 to 9.

Images