Massive power grid safety grounding and intelligent dispatching system and method based on big data

Abstract

The present application relates to a kind of large-scale power grid safety grounding and intelligent scheduling system and method based on big data, solve the independent operation of three links of grounding decision, dispatching cooperation and safety protection, lack unified data support and linkage mechanism problem;System contains: set and integrate multi-source data, construct fault knowledge graph, false rule library and equipment health state baseline, convert unstructured experience into structured knowledge base;According to dynamic working condition, select the grounding position with minimum comprehensive cost, generate distribution network reconstruction and load transfer scheme and issue dispatching instruction;With knowledge graph, match fault reason and output processing scheme, realize closed-loop linkage through data transmission and feedback between each step;The present application is applied to the live maintenance scene of distribution ring network cabinet load switch electric operating mechanism;The present application makes maintenance plan and distribution network reconstruction load transfer be executed according to preset timing sequence in cooperation rather than independently formulated.

Description

Technical Field

[0001] This invention relates to a large-scale power grid safety grounding and intelligent dispatching system and method based on big data, belonging to the field of power grid safety grounding technology. Background Technology

[0002] Distribution ring main units are core equipment in distribution networks, responsible for load distribution and line connection. Their load switch electric operating mechanisms, due to long-term energized operation, are prone to failures such as coil aging, contact wear, and mechanism jamming. Therefore, maintenance and repair must be carried out without power outages or with minimal power interruptions. With the expansion of distribution network scale and the increasing demands for power supply reliability, live-line maintenance has become a routine operation method for distribution network maintenance.

[0003] The following problems exist in the current live-line maintenance of ring main units: Firstly, the determination of the necessity of safe grounding and the selection of grounding points mainly rely on the experience of operating personnel or fixed procedures. Whether to install grounding wires before maintenance, and at which point to install them, is usually carried out according to general operating guidelines, without fully considering dynamic factors such as the historical failure rate of equipment, current load level, status of nearby equipment, and weather conditions. When distributed power sources are connected, the power flow distribution of the distribution network becomes more complex, and the safety margin of fixed grounding strategies is difficult to accurately assess.

[0004] Secondly, the live-line maintenance plan and the distribution network dispatch instructions are formulated independently. After the maintenance department submits the power outage window, the dispatch department arranges load transfer and network reconfiguration separately, lacking an automatic coordination mechanism between the two. When a certain section of the ring main unit is shut down for maintenance, if the load is not transferred to an adjacent feeder in advance, it may lead to a power outage of the entire station or a large-scale power loss for users.

[0005] Third, the anti-misoperation interlocking logic and fault diagnosis methods are separated. Existing ring main unit anti-misoperation functions rely on mechanical or electrical interlocks, with fixed interlocking rules that cannot be dynamically updated with operational data. If an abnormality occurs in the operating mechanism during maintenance, troubleshooting relies on manually reviewing drawings and maintenance records, which is time-consuming and prone to missing related information. Summary of the Invention To address the problems existing in the prior art, this invention proposes a large-scale power grid safety grounding and intelligent dispatching system and method based on big data.

[0006] The technical solution of the present invention is as follows: On the one hand, this invention provides a large-scale power grid safety grounding and intelligent dispatching system based on big data, including: The distribution ring network cabinet data integration component is used to collect and integrate distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; Construct a fault knowledge graph, a rule base for preventing errors, and a baseline for equipment health status; The historical fault and maintenance records include the equipment number, model, and operating mechanism type at the time of the fault, a description of the fault phenomenon, an analysis of the fault cause, maintenance measures and post-repair test data, the time of the fault and the interval since the last maintenance, the sequence of operating instructions before and after the fault and their execution results, the operating conditions such as load current and bus voltage at the time of the fault, and environmental conditions such as temperature and humidity. The fault knowledge graph stores the entities and their relationships of fault phenomena, causes, components, etc., from historical fault records, maintenance cases, and expert experience texts in a pre-set graph database. This is used to quickly match and output the fault cause and handling plan by inputting fault phenomena during live-line maintenance. The anti-misoperation rule base is used to convert historical misoperation characteristics into structured judgment conditions for the real-time monitoring module to perform millisecond-level comparisons, so as to automatically trigger interlocking when abnormal operation instructions are identified. The equipment health status baseline includes insulation resistance value, operating mechanism action time, coil current peak value, etc. This is used to provide a normal range reference benchmark for real-time monitoring parameters. When the measured data exceeds the baseline threshold, a mechanism fault warning is triggered, and the knowledge graph is activated to troubleshoot the fault. Adjust the scheduling instruction component to calculate the necessity and optimal location of safe grounding based on the equipment's historical failure rate, current load, status of nearby equipment, and weather conditions; Generate power grid reconfiguration and load transfer plans, and coordinate the issuance of maintenance plans and dispatch instructions; An abnormal signal identification component is used to continuously monitor changes in electrical quantities and position status of the operating mechanism, perform millisecond-level comparisons with the anti-misoperation rule base, and trigger local and remote dual interlocks after identifying abnormal signals; If a mechanical failure is detected, the user can input the symptoms via a mobile terminal, and the system will use a knowledge graph to match the cause of the failure and output a solution for handling the energized system.

[0007] Preferably, the fault knowledge graph is used in live-line maintenance to quickly match fault symptoms and output fault causes and solutions. The error prevention rule base is used to transform historical error operation characteristics into structured judgment conditions, which are then compared by the real-time monitoring module at the millisecond level to automatically trigger the interlock when an abnormal operation command is detected. The equipment health status baseline is used to provide a reference benchmark for the normal range of real-time monitoring parameters. When the measured data exceeds the baseline threshold, a mechanism fault warning is triggered, and the knowledge graph is activated to troubleshoot the fault.

[0008] Preferably, the optimal location is selected from all candidate grounding points of the ring main unit using a three-layer nested nonlinear cost model; The first layer of the model transforms grounding impedance, load influence range and operation time into grounding impedance risk factor, load influence factor and operation exposure factor through saturation mapping, respectively. The load influence factor is determined by the minimum outage section load defined by the adjacency table and the residual load estimated by the reliable transfer margin of the adjacent feeder. The second layer integrates the equipment's historical failure rate, current load, status of neighboring equipment, and lightning strike probability into a working condition stress index, and couples the working condition stress index with the three risk factors into three sub-costs in different nonlinear forms. The third layer uses logarithmic-exponential aggregation to calculate the comprehensive cost based on the adaptive change of the aggregation order with the stress exponent under the operating conditions, and selects the point with the minimum comprehensive cost as the target location for grounding operation; the distribution network reconfiguration and load transfer scheme is a combination of switching strategies that minimize the power outage range affected by grounding operation by adjusting the switch status and transferring the load after determining the grounding point location.

[0009] Preferably, the adjustment scheduling instruction component includes: The module for calculating the necessity and optimal location of grounding is used to collect historical failure rates, current load values, operating status of nearby equipment, and weather conditions of the ring main unit equipment. The module for generating a distribution network reconfiguration scheme is used to read the current distribution network topology; generate a load transfer scheme, identify the load segments that need to be de-energized due to grounding operations, calculate the transferable capacity of adjacent feeders, traverse all transferable feeders, sort them by remaining capacity from largest to smallest, and then allocate the loads in sequence until all loads to be transferred have been allocated. The dispatch instruction module is used to merge grounding operation instructions, load transfer schemes, and distribution network reconfiguration schemes into a dispatch instruction set; the dispatch instruction set is time-sequenced, with load transfer operations executed first, followed by distribution network reconfiguration, and then grounding operations; The sorted instruction set is sent to the on-site maintenance personnel terminal and the remote dispatch control system, and execution confirmation feedback is received; the dispatch instruction component is adjusted to reassess the current distribution network topology based on the interlocking event information and determine whether the issued load transfer path needs to be updated.

[0010] Preferably, if the current load is greater than a preset percentage of the rated value, the historical failure rate is higher than a preset threshold, and there is an overload in the nearby equipment, then it is determined that grounding is required; otherwise, it is also determined that grounding is required when the probability of lightning strikes exceeds a preset percentage in the meteorological conditions. The system iterates through all groundable points in the ring main unit and calculates the comprehensive cost of each groundable point using a three-layer nested nonlinear cost model. The groundable point with the minimum comprehensive cost is then selected for output. After the power grid reconfiguration and load transfer scheme is executed, the path attenuation parameters used for residual load estimation are corrected by feedback using the measured residual load.

[0011] Preferably, if the remaining capacity of a single feeder is insufficient, multiple feeders are selected simultaneously to share the transferred load, and the load rate of each feeder after allocation is verified; the feeders that pass the verification are used to form a load transfer path; a distribution network reconfiguration scheme is generated; the on / off states of the sectionalizing switches and tie switches in the topology are adjusted; and the reconfigured switch state sequence is output.

[0012] Preferably, the module for generating distribution network reconfiguration schemes includes: The module for constructing a switch state matrix is ​​used to read the current distribution network topology and construct the switch state matrix; obtain the opening and closing positions of all sectional switches and tie switches in the current distribution network, as well as the current load rate of each feeder; and establish an adjacency table to record the connection relationships between ring main unit nodes and switch station nodes. The transfer path submodule is used to identify feeder sections affected by grounding operations; mark the bus or feeder trunk section where the optimal grounding point is located; list all directly connected sectionalizing switches; and determine the minimum power outage section that needs to be isolated if a grounding operation is performed. The optimal reconfiguration scheme submodule is selected to determine the target open / closed state of the switch after reconfiguration. It simulates each power transfer path, changes the sectionalizing switch at the boundary of the power outage section from closed to open, and changes the target tie switch from open to closed.

[0013] Preferably, the search is conducted for available transfer paths for the power outage load section. Starting from the boundary switch of the power outage load section, the search proceeds outward along the currently closed tie switches, traversing all tie switches that are electrically connected to the power outage section and are currently in the open state. For each candidate tie switch, read the real-time load rate of its target feeder and calculate its remaining transferable capacity; filter tie switches whose remaining capacity is greater than the capacity required by the outage load; the minimum outage section is the outage load.

[0014] Preferably, the simulated transfer path is verified; if a path passes the verification, the new state of the switch corresponding to that path is recorded as part of the reconfiguration scheme; multiple transfer paths are compared and the optimal reconfiguration scheme is selected. If multiple tie switches meet the conditions, compare the total number of reconfiguration operations under each path and the maximum load rate of each feeder after reconfiguration, and select the path with the fewest operations and the lowest maximum load rate as the final distribution network reconfiguration scheme. Output the switch numbers that need to change state and their target state reconstruction schemes; read the reconstruction schemes and extract the list of all switches that need to be opened or closed.

[0015] On the other hand, the present invention also provides a method for large-scale power grid security grounding and intelligent dispatching based on big data, including the following steps: Collect and integrate power distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; construct a fault knowledge graph, a misoperation prevention rule base, and an equipment health status baseline. Based on the equipment's historical failure rate, current load, status of neighboring equipment, and weather conditions, calculate the necessity and optimal location of safe grounding; generate power distribution network reconfiguration and load transfer schemes, and coordinate the issuance of maintenance plans and dispatch instructions; The system continuously monitors changes in electrical quantities and position status of the operating mechanism, performs millisecond-level comparisons with the anti-misoperation rule base, and triggers local and remote dual interlocks after identifying abnormal signals. If a mechanism malfunction is detected, the system inputs the phenomenon through a mobile terminal, uses a knowledge graph to match the cause of the malfunction, and outputs a live handling solution.

[0016] The present invention has the following beneficial effects: 1. This invention structures unstructured historical maintenance texts, operational data, and scheduling logs into three different forms of knowledge representation, providing a unified data reference benchmark for subsequent grounding decisions, scheduling coordination, and anomaly identification. Specifically, the fault knowledge graph stores entities and relationships in a graph structure, constructing navigation paths between fault phenomena, causes, components, and handling measures; the error prevention rule base transforms historical erroneous operation characteristics into trigger conditions that can be compared in real time; and the equipment health status baseline calibrates the multi-dimensional distribution range of normal operating parameters. 2. This invention determines whether safety grounding is required based on a combination of criteria including historical equipment failure rate, current load, status of neighboring equipment, and meteorological conditions. It selects the grounding point with the lowest overall cost based on impedance, load impact range, and operating time, and uses the selected grounding point to drive the generation of subsequent distribution network reconfiguration and load transfer schemes. This ensures that the load is allocated to feeders with sufficient remaining capacity, generates a switch state sequence containing the action targets of sectionalizing switches and tie switches, and merges the grounding operation command, load transfer scheme, and distribution network reconfiguration scheme into a scheduling command set, issuing them in the order of transfer, reconfiguration, and grounding, thus linking maintenance plans with scheduling execution.

[0017] 3. During maintenance, this invention continuously collects electrical quantities and location status at a preset sampling frequency and compares them with the anti-misoperation rule base at the millisecond level. When an abnormal signal matching the status condition, electrical quantity threshold, timing condition, or instruction consistency condition is identified, the on-site output relay circuit is cut off and the remote dispatch master station equipment is locked simultaneously, so that physical blocking and system blocking take effect at the same time. When the measured parameters exceed the preset range of the health baseline, the current abnormal characteristics are traversed and matched with the knowledge graph along the defined relationships such as cause and manifestation, and the corresponding processing solution is output, so that the fault cause retrieval and the processing solution output are linked. Attached Figure Description

[0018] Figure 1 This is a block diagram of a large-scale power grid security grounding and intelligent dispatching system based on big data, as provided in Embodiment 1 of the present invention. Figure 2 This is a block diagram of the power distribution ring network cabinet data integration module provided in Embodiment 2 of the present invention; Figure 3 This is a block diagram of the adjustment scheduling instruction component provided in Embodiment 6 of the present invention; Figure 4 This is a block diagram of the abnormal signal identification component provided in Embodiment 9 of the present invention; Figure 5 This is a flowchart of a large-scale power grid security grounding and intelligent scheduling method based on big data, as provided in Embodiment 10 of the present invention. Figure 6 A block diagram of the electronic device provided by the present invention; Figure 7 A block diagram of a computer-readable storage medium provided for this invention.

[0019] Reference numerals in the attached diagram: 1. Distribution ring main unit data integration component; 2. Adjustment and dispatch instruction component; 3. Abnormal signal identification component; 4. Central processing unit / microprocessor / main control chip; 5. Storage medium; 6. Data bus; 7. Input / output bus / external bus / device bus; 8. Display; 9. Input / output device; 10. Computer-readable instructions; 11. Non-transitory computer-readable storage medium. Detailed Implementation

[0020] 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.

[0021] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.

[0022] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.

[0023] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.

[0024] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.

[0025] Example 1: See Figure 1 This invention provides a large-scale power grid security grounding and intelligent dispatching system based on big data, comprising: Distribution ring network cabinet data integration component 1 is used to collect and integrate distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; and to build a fault knowledge graph, a fault prevention rule base, and a device health status baseline. The historical fault and maintenance records include the equipment number, model, and operating mechanism type at the time of the fault, a description of the fault phenomenon, an analysis of the fault cause, maintenance measures and post-repair test data, the time of the fault and the interval since the last maintenance, the sequence of operating instructions before and after the fault and their execution results, the operating conditions such as load current and bus voltage at the time of the fault, and environmental conditions such as temperature and humidity. The fault knowledge graph stores the entities and their relationships of fault phenomena, causes, components, etc., from historical fault records, maintenance cases, and expert experience texts in a pre-set graph database. This is used to quickly match and output the fault cause and handling plan by inputting fault phenomena during live-line maintenance. The anti-misoperation rule base is used to convert historical misoperation characteristics into structured judgment conditions for the real-time monitoring module to perform millisecond-level comparisons, so as to automatically trigger interlocking when abnormal operation instructions are identified. The equipment health status baseline includes insulation resistance value, operating mechanism action time, coil current peak value, etc. This is used to provide a normal range reference benchmark for real-time monitoring parameters. When the measured data exceeds the baseline threshold, a mechanism fault warning is triggered, and the knowledge graph is activated to troubleshoot the fault. Adjust the dispatch instruction component 2 to calculate the necessity and optimal location of safe grounding based on the equipment's historical failure rate, current load, status of nearby equipment, and weather conditions; generate distribution network reconfiguration and load transfer schemes, and coordinate the issuance of maintenance plans and dispatch instructions; Among them, the optimal location is the grounding operation target location with the lowest comprehensive cost selected from all candidate grounding points of the ring main unit by weighted comparison of grounding impedance, load impact range and operation time; the distribution network reconfiguration and load transfer scheme is a combination of switching strategies that minimize the power outage range affected by grounding operation by adjusting switch status and transferring load after determining the grounding point location. The abnormal signal identification component 3 is used to continuously monitor the changes in electrical quantities and position status of the operating mechanism, compare them with the anti-misoperation rule base at the millisecond level, and trigger local and remote dual interlocks after identifying abnormal signals; if a mechanism fault is detected, the phenomenon is input through a mobile terminal, the fault cause is matched using a knowledge graph and a live handling solution is output. Among them, the real-time monitoring data of electrical quantity changes and position status sampled at a preset sampling frequency during maintenance execution is used to compare with the anti-misoperation rule base to trigger interlocking, and to compare with the health baseline to determine the mechanism fault; the dual interlocking includes redundant safety protection measures that simultaneously cut off the execution circuit on site and send a blocking command to the dispatch master station after abnormal operation signal is detected.

[0026] In the above embodiments, the distribution ring network cabinet data integration module transforms unstructured experience into a structured knowledge base by constructing a knowledge graph, an anti-misoperation rule base, and a health baseline, providing a unified data reference benchmark for subsequent decision-making. The scheduling instruction adjustment module determines grounding requirements based on multi-source status data and selects the optimal location, simultaneously generating reconfiguration and load transfer schemes to minimize the impact of grounding operations on the power supply range, achieving coordinated cooperation between maintenance tasks and distribution network operation modes. The abnormal signal identification module continuously compares real-time electrical quantities and location status during maintenance execution. When an anomaly is detected, it simultaneously executes local circuit disconnection and remote blocking to prevent misoperation; at the same time, it triggers fault warnings based on the health baseline and completes rapid matching of fault causes through the knowledge graph. This embodiment connects data preparation, scheme decision-making, and protection execution in a sequential manner, shifting the determination of grounding necessity from fixed rules to dynamic calculation based on real-time operating conditions, expanding the selection of grounding location from single impedance consideration to comprehensive cost assessment, upgrading the anti-misoperation mechanism from isolated threshold judgment to multi-dimensional rule comparison and dual execution, and transforming fault handling from manual data review to structured matching. The data integration module provides equipment status baseline and graph support for the scheduling instruction module. The scheduling instruction module outputs the expected operation timing benchmark for the abnormal signal identification module. The maintenance data of the abnormal signal identification module is fed back to the data integration module to update the baseline. The whole realizes the integration of grounding decision-making, scheduling coordination and safety protection based on multi-source data-driven. In this embodiment, the knowledge graph constructed by the distribution ring network cabinet data integration component 1 provides fault cause classification labels for the anti-misoperation rule base, and the health baseline provides the normal parameter range reference for the anti-misoperation rule base, so that the judgment conditions of the anti-misoperation rule base have the ability to adapt to equipment status. The operation timing benchmark generated by the scheduling instruction component 2 is adjusted and output to the abnormal signal identification component 3, so that the judgment basis of the abnormal signal includes not only the static comparison of electrical quantities with the rule base, but also the deviation comparison with the expected operation timing. The interlocking trigger record and fault matching result during the maintenance execution of the abnormal signal identification component 3 are fed back to the distribution ring network cabinet data integration component 1 to update the health baseline threshold and supplement the anti-misoperation rule base entries. The three components form a data-driven closed-loop iteration, which enables the knowledge graph, error prevention rule base, and health baseline to be continuously optimized as actual operational data accumulates, distinguishing them from existing systems where each module operates independently.

[0027] Example 2: Figure 2 As shown, based on Embodiment 1, the power distribution ring network cabinet data integration module 1 provided in this embodiment of the invention includes: A fault knowledge graph module 11 is constructed to extract entities and relationships from historical faults and maintenance records, maintenance cases, and expert experience texts; perform keyword matching and dependency parsing on maintenance cases and historical faults and maintenance records to extract entity and relationship triples; store entity and relationship triples in a preset graph database; and establish a fault graph indexing and retrieval mechanism. The entities include six categories: fault phenomena, fault causes, equipment components, operating mechanism types, environmental conditions, and temporal characteristics. Relationships include four categories: causing, manifesting as, occurring at, and related operations. "Cause" refers to fault phenomena pointing to fault causes; "manifesting as" refers to equipment components pointing to fault phenomena; "occurring at" refers to fault causes pointing to environmental conditions or temporal characteristics; and "related operations" refers to fault phenomena pointing to operating command sequences. Entities and relationships together constitute the core elements of the knowledge graph, used for triple extraction and storage. The fault graph indexing and retrieval mechanism is used during live-line maintenance to quickly locate relevant parameters in the preset graph database by inputting fault phenomena or equipment status keywords. The fault causes and handling solutions are described; the fault graph index and retrieval mechanism includes a text index for fault phenomena and fault cause node name attributes; a precise index for equipment component node names; an enumerated index for operating mechanism type node classification attributes; and a general index for all node equipment numbers and record time attributes; during retrieval, candidate nodes are located through the index, and a graph traversal algorithm is used to expand along the cause and manifestation of equal relationships, returning the associated subgraph and outputting it in sorted order by historical hit frequency; relation triples are the basic data units that constitute the fault knowledge graph, using a structured format of head entity, relation type, and tail entity to store the logical relationships between entities; A rule base module 12 for preventing malfunctions is constructed to collect time-series data of voltage, current, and position status of the operating mechanism during historical live-line maintenance processes, as well as personnel operation instructions during the same period; the data is time-series aligned, and the electrical quantity change characteristics and operation instruction sequence before and after the malfunction occurs are marked; judgment conditions are extracted from the marked samples; each judgment condition is converted into a structured rule; the structured rules are classified and stored according to equipment type and operation type, and rule version identifiers and timestamps are established; The judgment conditions include state conditions, electrical quantity conditions, timing conditions, and command consistency conditions. State conditions include the switch position and mechanism travel position when the operation occurs. Electrical quantity conditions include the amplitude range, rate of change threshold, and duration of over-limit of voltage or current. Timing conditions include the time interval between the issuance of the command and the actual action. Command consistency conditions include whether the intention of the operation command conflicts with the current state of the equipment. The equipment health status baseline module 13 is used to filter historical data of the ring main unit during the period of no fault and no abnormal operation in the operation data; extract the characteristics of each monitoring parameter; establish a multi-dimensional joint distribution of the key parameters of the operating mechanism and determine the upper and lower limits of the preset range; store the threshold range, characteristic curve and update timestamp of the baseline into the baseline database, and set the baseline update mechanism. The monitored parameters include voltage characteristics, current characteristics, electrical characteristics of the operating mechanism, mechanical characteristics of the operating mechanism, and environmental characteristics. Voltage characteristics include the average, effective, maximum, minimum, and fluctuation range of the three-phase voltage. Current characteristics include the average, peak, effective, and load rate distribution of the three-phase load current. Electrical characteristics of the operating mechanism include the peak value of the trip coil current, the peak value of the closing coil current, the rise time and duration of the coil current, and the area of ​​the coil current waveform. Mechanical characteristics of the operating mechanism include the tripping action time, the closing action time, the peak value of the action stroke displacement, and the cumulative number of operations. The tripping action time refers to the time from the issuance of the command to the switch... The time of measurement; environmental characteristics include the average and rate of change of ambient temperature, the average and rate of change of ambient humidity, and the frequency of condensation; the multidimensional joint distribution is a joint probability distribution established for the peak value of the coil current of the operating mechanism, the action time, and the stroke displacement, used to determine the combined range of multiple parameters when they are simultaneously in a normal state; the baseline database stores the threshold range, characteristic curves, and update timestamps of the device health status baseline, including the normal value range of parameters such as voltage, current, peak coil current, and action time, as well as the confidence interval of the multidimensional joint distribution, for the real-time monitoring module to call and compare during maintenance to determine whether the measured parameters deviate from the normal range.

[0028] In the above embodiments, the fault knowledge graph module 11 extracts entities and relationships from fault records and maintenance cases to form a graph database storage structure, supporting the rapid retrieval of causes and handling measures based on fault phenomena; the anti-misoperation rule base module 12 extracts electrical quantities, timing, and consistency judgment conditions from historical malfunction events, transforming them into categorized and stored rule entries for real-time comparison; and the equipment health status baseline module 13 extracts parameter features from fault-free operation data and establishes multi-dimensional joint distribution intervals to form a normal range benchmark. These three modules transform the raw data into three technical forms: graph retrieval, rule comparison, and baseline reference, respectively, collectively constituting a unified data foundation required for subsequent grounding decisions, scheduling coordination, and anomaly identification.

[0029] Example 3: Based on Example 2, the fault knowledge graph construction module 11 provided in this embodiment of the invention includes: The Relationship Constraints submodule is used to summarize entity categories and define relationship categories from historical fault records, maintenance cases, and expert experience texts; it also specifies attribute fields and relationship constraints for each type of entity. The defined relation categories include four types: cause, manifestation, occurrence, and association. These are used to associate entities extracted from historical fault records, maintenance cases, and expert experience texts to form structured triples that are stored in a pre-defined graph database. Relation constraints refer to the rule restrictions imposed on the application of the four types of relations: cause, manifestation, occurrence, and association. These constraints are used to verify the validity of triples when storing nodes and edges in the pre-defined graph database and to limit the validity of traversal paths during retrieval. Cause relations allow only fault phenomenon entities as head entities and fault cause entities as tail entities. One fault phenomenon can correspond to multiple fault causes, but a fault cause cannot be reversed. This can lead to multiple fault phenomena; the relationship is such that only equipment component entities are allowed as head entities and fault phenomenon entities are allowed as tail entities; a single equipment component can exhibit multiple fault phenomena, but a fault phenomenon can only be attributed to one equipment component; the occurrence relationship includes allowing only fault cause entities as head entities and environmental condition or time characteristic entities as tail entities; a fault cause can occur under multiple environmental or time conditions, but environmental conditions do not occur inversely to the fault cause; the associated operation relationship includes allowing only fault phenomenon entities as head entities and operation instruction sequences as tail entities; a fault phenomenon can be associated with multiple operation instruction sequences, but the same operation instruction sequence does not associate with different fault phenomena; The triple generation submodule is used to perform text cleaning, word segmentation, and terminology unification on maintenance records and fault reports; it uses a pre-built entity dictionary in conjunction with regular expressions to match entity mentions; it uses dependency parsing for parts not covered by the dictionary or ambiguous parts; it generates triples based on relation categories and adds contextual attributes; and it performs deduplication and conflict detection on the extracted triples. The entity dictionary includes a predefined terminology library for identifying and matching entity references from unstructured text. Deduplication and conflict detection of extracted triples refers to the process of eliminating redundancy and verifying logical consistency of triples extracted from different maintenance records or fault reports. Deduplication involves retaining one of the duplicate triples for the same equipment, the same fault phenomenon, and the same fault cause, even if they originate from different sources, and marking the source of redundancy. Conflict detection involves resolving conflicts by retaining the triple with the highest frequency in historical records or the highest weight in expert annotations when multiple contradictory fault causes are extracted for the same equipment and the same fault phenomenon. The candidate node submodule is used to store triples into a preset graph database, with entities as nodes and relations as directed edges, and attribute fields stored on the nodes; it creates indexes for entity categories and attributes; and it locates candidate nodes by inputting fault phenomena or equipment status keywords through the index. The process of locating candidate nodes by indexing utilizes a preset graph database indexing mechanism to quickly filter out initial entity nodes that match the query conditions based on input keywords. After receiving input keywords for fault phenomena or equipment status, the keywords are fuzzily matched with the text indexes of the fault phenomenon node and the "fault cause" node, respectively. At the same time, attribute indexes such as equipment number and operating mechanism type are used for filtering, and all nodes with a matching degree higher than a set threshold are returned as a candidate node list. If the number of candidate nodes exceeds a preset limit, the scope is further narrowed down by using the equipment number or operating mechanism type attribute index.

[0030] In the above embodiments, the relation constraint submodule, triple generation submodule, and candidate node submodule collaboratively constitute the complete technical path for constructing the fault knowledge graph module 11. The relation constraint submodule imposes head and tail entity type restrictions and cardinality limitations on four types of relations, serving as the basis for legality verification in subsequent storage and retrieval stages. The triple generation submodule, based on the definition of relation constraints, extracts triples that meet the constraints from unstructured text through entity dictionary matching, dependency parsing, deduplication, and conflict detection. The candidate node submodule stores the triples in the form of nodes and directed edges into the graph database, establishes multiple indexes, and achieves rapid selection of candidate nodes through index location. The sequential connection of the three submodules—defining constraints, extracting triples, storing in the database, and establishing indexes—transforms unstructured fault text into searchable graph structure data, supporting subsequent rapid location of associated fault causes and handling solutions through keywords.

[0031] Example 4: Based on Example 2, the steps in this embodiment of the invention to construct the error prevention rule base module include: The syntax tree generation submodule is used to locate the core predicate in the syntax tree obtained through dependency parsing based on the relation category; extract the subject part of the predicate as the candidate head entity and the object part as the candidate tail entity; if the subject or object is a compound phrase, the core noun is identified as the entity head word and the modifiers are identified as entity attribute values. The syntax tree is used to parse unstructured fault description text that cannot be covered by the entity dictionary or is ambiguous; the relation type is determined by locating the root node, the subject is extracted as the head entity and the object as the tail entity, and the modifiers are used as entity attribute values; thus, the triples are extracted; the entity attribute values ​​provide an additional filtering dimension during retrieval matching. The basic triplet format generation submodule is used to match candidate graph entities, candidate tail entities, and core predicates with relation categories; if the core predicate matches, a triplet is generated; if the core predicate matches the representation, the head entity, representation, and tail entity are generated; for predicates that cannot be directly matched, they are mapped to the closest relation category based on the predicate's directionality; and the basic triplet format is generated. The basic triplet format consists of a first entity, a relation type, and a second entity; The final triplet submodule is used to attach preset attributes to each generated triplet; the modifiers extracted from the head and tail entities are stored as attribute values ​​in the attribute fields of the corresponding entities; after the attribute attachment is completed, the triplet is input into the queue to be stored. The preset attributes include the record source, record time, device number, and operator identification.

[0032] In the above embodiments, the syntax tree generation submodule, the basic triplet format generation submodule, and the final triplet submodule collaboratively constitute the complete execution path for triplet extraction. The syntax tree generation submodule performs syntactic structure parsing on text not covered by the entity dictionary, locating the core predicate and extracting the subject as a candidate head entity, the object as a candidate tail entity, and the modifiers as entity attribute values. The basic triplet format generation submodule matches candidate entities and core predicates with predefined relation categories, mapping them to the closest relation category based on predicate directionality to generate basic triples. The final triplet submodule adds preset attributes such as record source, record time, device number, and operator identifier to each triplet, and stores the modifiers in the corresponding entity attribute fields before inputting them into the storage queue. The three submodules are sequentially linked according to the order of parsing text, matching relations, adding attributes, and outputting, transforming unstructured fault description text not covered by the entity dictionary into complete triples with contextual attributes for subsequent storage in the graph database.

[0033] Example 5: Based on Example 2, the module for constructing a device health status baseline provided in this embodiment of the invention includes the following steps: The module for constructing nodes and directed edges is used to sequentially read each triple from the queue to be stored; if the node corresponding to the first entity does not exist in the preset graph database, a new node is created, and the node label is the entity category; if the node already exists, it is reused; the same applies to the second entity; a directed edge is created from the first entity node to the second entity node, and the edge type is set to relation type; the context attributes attached to the triple are stored in the attribute fields of the first entity node; the modifiers are stored in the attribute fields of the second entity node. A multi-index submodule is established to create indexes for node labels in the preset graph database; to create a text index supporting fuzzy matching for the name attribute of fault phenomenon label nodes; to create a text index for the name attribute of fault cause label nodes; to create an exact index for the name of equipment component label nodes; to create an enumerated index for the classification attribute of operating mechanism type label nodes; and to create a general index for the equipment number and record time attribute of all nodes. The node location and retrieval output submodule receives input keywords, performs fuzzy matching between the keywords and the text indices of fault phenomenon nodes and fault cause nodes to obtain a list of candidate nodes; if the number of matching results exceeds a set threshold, it filters the results using the device number or operating mechanism type attribute index; it outputs the located candidate nodes and returns the node's name, label, and context information in the attribute fields; if it is necessary to continue traversing along the relationship, it uses the candidate node as the starting point and returns all outgoing target nodes and relationship types.

[0034] In the above embodiments, the node and directed edge construction submodule, the multi-index building submodule, and the node location and retrieval output submodule together constitute a complete technical path for graph database storage and retrieval. The node and directed edge construction submodule sequentially reads triples from the queue to be stored, creates or reuses entity nodes in the graph database, and establishes directed edges with relational types. Simultaneously, it stores contextual attributes and modifiers in the attribute fields of the corresponding nodes. The multi-index building submodule builds text indexes, exact indexes, enumeration indexes, and ordinary indexes for nodes with different labels, providing a multi-dimensional and fast filtering channel for subsequent retrieval. The node location and retrieval output submodule receives input keywords, locates candidate nodes through text indexes, filters and outputs node information using attribute indexes, and supports traversing along relational edges to return associated subgraphs. The three submodules are connected sequentially according to the data entry, index building, and retrieval output, transforming triple data into a graph structure that can be quickly queried, supporting the location of fault causes and solutions through keywords.

[0035] Example 6: As Figure 3 As shown, based on Embodiment 1, the adjustment scheduling instruction component 2 provided in this embodiment of the invention includes: The grounding necessity and optimal location module 21 is used to collect the historical failure rate, current load value, operating status of adjacent equipment, and meteorological conditions of the ring main unit equipment. If the current load is greater than a preset percentage of the rated value, the historical failure rate is higher than a preset threshold, and there is an overload in the adjacent equipment, it is determined that grounding is required. Otherwise, it is also determined that grounding is required when the probability of lightning strikes exceeds a preset percentage in the meteorological conditions. All groundable points in the ring main unit are traversed, the comprehensive cost of each groundable point is calculated, and the groundable point with the minimum comprehensive cost is selected for output. Among them, the historical failure rate includes the number of failures per unit time; the current load value includes current and power; the adjacent equipment is other ring main units on the same bus or line; the operating status of the adjacent equipment is whether it is under maintenance or overloaded; the meteorological conditions are temperature, humidity, rainfall and lightning strike probability; the grounding points are the terminals of incoming and outgoing cables and the bus grounding switch, etc.; the comprehensive cost is the grounding impedance, the impact range on load transfer and the operation time. The distribution network reconfiguration scheme generation module 22 is used to read the current distribution network topology; generate a load transfer scheme, identify load segments that need to be de-energized due to grounding operations, calculate the transferable capacity of adjacent feeders, traverse all transferable feeders, sort them by remaining capacity from largest to smallest, and allocate the loads sequentially until all loads to be transferred are allocated; if the remaining capacity of a single feeder is insufficient, multiple feeders are selected simultaneously to share the transferred loads, and the load rate of each feeder after allocation is verified; the verified feeders form a load transfer path, and the distribution network reconfiguration scheme is generated; the on / off states of sectionalizing switches and tie switches in the topology are adjusted; and the reconfigured switch state sequence is output. The current distribution network topology includes the status of tie switches and the load rate of each feeder; The dispatch instruction module 23 is used to merge grounding operation instructions, load transfer schemes, and distribution network reconfiguration schemes into a dispatch instruction set; it performs time-sequence sequencing on the dispatch instruction set, executing load transfer operations first, then distribution network reconfiguration, and finally grounding operations; it sends the sequenced instruction set to the on-site maintenance personnel terminal and the remote dispatch control system, and receives execution confirmation feedback; after executing the interlocking operation, it outputs the timestamp of the interlocking event, the trigger rule number, and the affected switch number to the adjustment dispatch instruction component 2; the adjustment dispatch instruction component 2 re-evaluates the current distribution network topology based on the interlocking event information to determine whether the issued load transfer paths need to be updated; Among them, the grounding operation instructions include the grounding point and grounding sequence; the load transfer scheme includes the list of switches to be operated and the action sequence; and the distribution network reconfiguration scheme includes the tie switch adjustment instructions.

[0036] In the above embodiments, the grounding necessity and optimal location module 21, the distribution network reconfiguration scheme generation module 22, and the dispatch command issuance module 23 collaboratively constitute the complete technical path of the adjustment dispatch command component 2. The grounding necessity and optimal location module 21 collects historical equipment failure rates, current load, adjacent equipment status, and weather conditions. Based on the load and failure rate combination criterion or lightning strike probability criterion, it determines whether grounding is required. Then, it iterates through candidate grounding points, calculates the comprehensive cost of grounding impedance, load impact range, and operation time, and selects the minimum point for output. After the grounding point is determined, the distribution network reconfiguration scheme generation module 22 reads the distribution network topology, identifies the load segments that need to be de-energized, calculates the transferable capacity of adjacent feeders, and uses a greedy algorithm to allocate the transferred load in descending order of remaining capacity. The module then assigns the load rate to each feeder after allocation. Verification is performed, and a reconfiguration scheme is formed by adjusting the states of sectionalizing switches and tie switches, outputting a switch state sequence. The dispatch instruction module 23 merges the grounding operation instruction, load transfer scheme, and distribution network reconfiguration scheme into a dispatch instruction set, and sends it to the field terminal and dispatch control system in the order of executing load transfer first, then distribution network reconfiguration, and finally grounding operation. After executing the interlocking operation, the timestamp of the interlocking event, the trigger rule number, and the affected switch number are fed back to the adjustment dispatch instruction component 2 for reassessing the distribution network topology and determining whether the issued load transfer path needs to be updated. The three modules are connected in the order of determining the selection point, generating the scheme, issuing and executing, and dynamically adjusting, so that the grounding decision, load transfer, and dispatch execution form a closed loop linkage.

[0037] Example 6-1: Based on Example 6, in the calculation module 21 for grounding necessity and optimal location provided in this embodiment of the invention, the comprehensive cost is calculated through a three-layer nested nonlinear cost model, including: The first layer is the characteristic saturation mapping layer, which transforms three physical quantities—grounding impedance, load influence range, and operating time—into dimensionless risk factors through nonlinear saturation mapping. The load influence range is quantified by the residual load prediction sub-model, and the grounding impedance is corrected by meteorological conditions. The second layer is the operating condition coupling layer, which integrates the equipment's historical failure rate, current load, status of neighboring equipment, and lightning strike probability into an operating condition stress index. It then uses three different nonlinear forms—exponential amplification, power-law modulation, and secondary self-coupling—to couple the operating condition stress with the three risk factors output by the first layer one by one, generating three sub-costs. The third layer is the adaptive aggregation decision layer, which adopts log-exponential aggregation with the aggregation order adaptively changing with the stress exponent of the working condition. This makes the comprehensive cost under high-risk working conditions dominated by the worst sub-cost, and tends to be evaluated in equilibrium under low-risk working conditions, outputting the comprehensive cost and selecting the minimum point.

[0038] Specifically, the first layer is a feature saturation mapping layer: (a) Grounding impedance risk factor Record all grounding points of the ring main unit as follows Candidate point number (Grounding points include incoming and outgoing cable terminals and busbar grounding switches, etc.)

[0039] Considering the impact of soil moisture content on grounding impedance, a meteorological correction is first applied to the measured grounding impedance in the ledger: ; In the formula, For point The equivalent grounding impedance (Ω); For point The most recent grounding impedance test value (Ω) in the ledger; The current ambient humidity (%) in the environmental data; The humidity decay scale constant (%) is determined by the soil type of the region. This is a drying correction factor, dimensionless, with a typical value range of 0.1 to 0.5. The drier the environment, the higher the equivalent grounding impedance.

[0040] The grounding impedance risk factor is then obtained through S-type saturation mapping: ; In the formula, For point The grounding impedance risk factor is dimensionless. The grounding impedance safety threshold (Ω) is set according to the grounding technical regulations; Steepness coefficient (Ω) This controls the steepness of the transition of the risk factor near the threshold. When the equivalent grounding impedance is below the threshold, the risk factor tends to 0, and when it is above the threshold, it rapidly saturates and tends to 1, thus avoiding the ineffective resolution of linear normalization for intervals far from the threshold.

[0041] (II) Residual load prediction sub-model and load influence factors During the site selection phase, the distribution network reconfiguration and load transfer schemes have not yet been generated, and the load impact range cannot be given by the complete reconfiguration solution. This sub-model only uses the accessibility of the adjacency table and the real-time margin of the adjacent feeders to make a conservative estimate of the residual load that cannot be transferred after the grounding is performed.

[0042] Set point The corresponding minimum power outage section load is (kW), from the adjacency list along the points The sectionalizing switch directly connected to the busbar or feeder trunk section is defined; (Note: The original text appears to be incomplete and contains several typographical errors. A more accurate translation would require the full context.) This refers to the group of adjacent feeders that are electrically connected to the de-energized section via a tie switch that is in the open state. For each feeder... Its reliable supply margin is: ; In the formula, For feeder Reliable supply margin (kW); For feeder The current load rate, dimensionless; This is the upper limit allowed for load rate verification; it is dimensionless and typically ranges from 0.8 to 0.9. For feeder Rated capacity (kW).

[0043] Considering that the deeper the supply path and the more switching operations involved, the lower the margin of reliability, a path attenuation coefficient is introduced: ; In the formula, For feeder The path attenuation coefficient is dimensionless. From the boundary switch of the power outage section to the feeder along the adjacency table The number of switches required to be operated; This is the path depth scale constant, with an initial value set by operational experience and corrected online by the feedback mechanism described later.

[0044] Points The estimated residual load is: ; In the formula, This represents the residual load (kW) that cannot be transferred after a grounding operation, based on a conservative estimate. In the point selection phase, this formula uses an adjacency list traversal instead of a complete reconstruction solution, and the computational complexity is linearly related to the number of candidate points.

[0045] The load impact factor is composed of residual load and transferred disturbance, and is obtained through exponential saturation mapping: ; In the formula, For point The load influence factor is dimensionless; The transfer disturbance coefficient is dimensionless and typically ranges from 0.05 to 0.2. It characterizes the disturbance to power supply continuity caused by the load transfer operation itself, even when the residual load is zero. The reference load (kW) is calibrated according to the typical load level of the area supplied by the ring main unit.

[0046] (iii) Operational exposure factors The operational risk in live-line working environments increases superlinearly with exposure time, and the operation time is transformed into an operational exposure factor through a power-law saturation mapping: ; In the formula, For point The operational exposure factor is dimensionless. For point The estimated operation time (min) is determined by the standard operating time quota and the field path coefficient corresponding to the point type (incoming and outgoing cable terminal, bus grounding switch, etc.); For reference, the operation time (min); To expose the risk index, dimensionless and This reflects the superlinear cumulative nature of exposure risk over time.

[0047] Specifically, the second layer is the operating condition coupling layer: (a) Normalization of dynamic operating condition components ; ; ; ; In the formula, To normalize the historical failure rate, This represents the number of failures per unit time for the ring main unit (times / year). The failure rate scaling constant (times / year); To normalize the load level, The ratio of the current load value to the rated value. As the load warning threshold, These are load steepness coefficients, all dimensionless; The proportion of risk associated with nearby equipment. This refers to the total number of adjacent ring main units on the same busbar or line. This refers to the number of units that are under maintenance or overloaded. To normalize the probability of a lightning strike, This represents the probability of a lightning strike under meteorological conditions. This represents the upper limit of the probability range for lightning strikes.

[0048] (ii) Stress index under working conditions The four normalized components are combined using probabilistic operations to form the stress index for the working condition: ; In the formula, For point The stress index of the ring main unit is dimensionless, and its physical meaning is the equivalent probability that at least one type of risk source is in an excited state.

[0049] (III) Three-way heterogeneous nonlinear coupling The stress index is coupled to the three risk factors in the first layer in three distinct nonlinear forms: ; ; ; In the formula, For the sake of safety, The stress amplification factor (dimensionless, calibration range 0.5 to 1.5) is used to exponentially amplify the impedance risk due to the working stress. To supply electronic costs, The tolerance attenuation coefficient (dimensionless) The higher the working stress, the smaller the power exponent, and the heavier the penalty for the same load impact, reflecting the nonlinear contraction of the tolerance to the impact of power outages under high stress conditions. For the cost of the task, The lightning strike coupling coefficient is dimensionless. The lightning strike probability and the operation exposure factor form a second-order self-coupling term. The longer the operation time and the higher the lightning strike probability, the more the operation sub-cost increases in a product-like manner.

[0050] Specifically, the third layer is the adaptive aggregation decision layer. (I) Stress-adaptive polymerization stage ; In the formula, For point The aggregation order is dimensionless; As the baseline aggregation level; This is the stress sensitivity coefficient.

[0051] (ii) Logarithmic-exponential aggregation of comprehensive costs ; In the formula, For point The overall cost is dimensionless. This polymerization has the following properties: hour It degenerates into the arithmetic mean of three sub-costs. hour Approaching As can be seen from the stress adaptive aggregation order calculation formula, the higher the stress exponent under the working condition, the larger the aggregation order, and the more the overall cost is dominated by the worst-case cost.

[0052] (III) Selection of the optimal location ; In the formula, The target location for grounding operation with the lowest overall cost is selected. If the difference in overall cost between multiple locations is less than the resolution threshold... Then, within this set of equivalent points, the safety sub-cost is applied. Minimum number of selections is used to ensure safety as a priority.

[0053] After the distribution network reconfiguration and load transfer scheme is actually implemented, the actual residual load is transmitted back to the network, and the path depth scale constant is corrected online. ; ; In the formula, The measured residual load (kW) after the reconfiguration scheme was implemented; Dimensionless for estimating relative deviation; To correct the step size coefficient (dimensionless, When the actual residual load is higher than the estimate (the transfer capacity is overestimated). , Narrowing the path attenuation coefficient tightens it, leading to a more conservative prediction in the next iteration; conversely, widening it loosens it. This mechanism creates a data loop between the point selection layer and the reconstruction layer, which, together with the feedback from the blocking event of the abnormal signal identification component, constitutes the evidence chain for the "cross-linkage closed loop" of this application.

[0054] Example 7: Based on Example 6, the module for generating a distribution network reconfiguration scheme provided in this embodiment of the invention includes: The module for constructing a switch state matrix is ​​used to read the current distribution network topology and construct the switch state matrix; obtain the opening and closing positions of all sectional switches and tie switches in the current distribution network, as well as the current load rate of each feeder; and establish an adjacency table to record the connection relationships between ring main unit nodes and switch station nodes. The power transfer path submodule is used to identify feeder sections affected by grounding operations; mark the busbar or feeder trunk section where the optimal grounding point is located; list all directly connected sectionalizing switches; determine the minimum outage section that needs to be isolated after a grounding operation; search for available power transfer paths for the outage load section; starting from the boundary switch of the outage load section, search outwards along the currently closed tie switches, traversing all tie switches that are electrically connected to the outage section and are currently in the open state; for each candidate tie switch, read the real-time load rate of its target feeder and calculate its remaining transferable capacity; and filter tie switches whose remaining capacity is greater than the capacity required by the outage load. The minimum power outage section is the section of load that is out of power; The optimal reconfiguration scheme submodule is used to determine the target open / closed state of the switches after reconfiguration. It simulates each power transfer path, changing the sectionalizing switches at the power outage boundary from closed to open, and changing the target tie switches from open to closed. The simulated power transfer paths are verified. If a path passes verification, the new state of the corresponding switch is recorded as part of the reconfiguration scheme. Multiple power transfer paths are compared, and the optimal reconfiguration scheme is selected. If multiple tie switches meet the conditions, the total number of reconfiguration operations and the maximum load rate of each feeder after reconfiguration are compared for each path. The path with the fewest operations and the lowest maximum load rate is selected as the final distribution network reconfiguration scheme. All switch numbers requiring state changes and their target state reconfiguration schemes are output. The reconfiguration scheme is read, and a list of all switches requiring open or close operations is extracted.

[0055] In the above embodiments, this embodiment constructs a complete technical path for generating a distribution network reconfiguration scheme by collaboratively forming a switch state matrix submodule, a power transfer path submodule, and an optimal reconfiguration scheme selection submodule. The switch state matrix submodule reads the current distribution network topology, obtains the opening and closing positions of all sectionalizing switches and tie switches, and the load rate of each feeder, and establishes an adjacency table to record the connection relationships between nodes. After determining the grounding point, the power transfer path submodule marks the bus or feeder trunk section where the optimal grounding point is located, lists the directly connected sectionalizing switches to determine the minimum outage section that needs to be isolated after grounding, and searches outwards along the closed tie switches, starting from the boundary switch of the outage load section. It traverses tie switches that are electrically connected to the outage section and are in the open state, calculates their remaining transferable capacity, and selects those whose capacity meets the requirements. The module selects candidate tie switches for the required reconfiguration scheme. It then simulates each power transfer path, changing the sectionalizing switch from closed to open and the target tie switch from open to closed. The simulated power transfer paths are validated, and those that pass validation are recorded as components of the reconfiguration scheme. When multiple tie switches meet the conditions, the total number of reconfiguration operations and the maximum load rate of each feeder after reconfiguration are compared. The path with the fewest operations and the lowest maximum load rate is selected as the final scheme. The module outputs the switch numbers and target states that need to be changed, and extracts a list of switches requiring either opening or closing operations. These three sub-modules are sequentially connected according to the topology construction, path search, simulation comparison, and scheme output, completing the full calculation from the current state of the distribution network to the optimal reconfiguration scheme.

[0056] Example 8: Based on Example 7, the sub-module for selecting the optimal reconstruction scheme provided in this embodiment of the invention includes: The status change entry unit is used to read the reconfiguration scheme and extract a list of all switches that need to be opened or closed; each switch in the list is recorded as a status change entry, and the entry format is uniform. The status change entry grouping unit is used to initially group status change entries according to action type, into a tripping operation group and a closing operation group; according to the distribution network operation safety regulations, all operations in the tripping group are executed first, and then all operations in the closing group are executed. A time-sequential switch state sequence unit is formed, which is used to sort the switches from near to far in the electrical distance from the power source within the opening group; and to sort the switches from far to near in the electrical distance from the backup power source within the closing group; and to merge the entries of the two groups to form a time-sequential switch state sequence. In the above embodiments, the state change entry unit, state change entry grouping unit, and time-sequential switch state sequence formation unit collaboratively constitute a complete technical path for converting the reconfiguration scheme into time-sequential operation instructions. The state change entry unit reads the selected distribution network reconfiguration scheme, extracts a list of all switches requiring opening or closing operations, and records each switch as a state change entry in a uniform format. The state change entry grouping unit divides the state change entries into opening operation groups and closing operation groups according to the action type, setting the inter-group order based on distribution network operation safety regulations: first, all operations within the opening group are executed, then all operations within the closing group are executed. The time-sequential switch state sequence formation unit sorts the switches in the opening group according to their electrical distance from the power source from near to far, and in the closing group according to their electrical distance from the backup power source from far to near. It then merges the two sorted groups of entries to output a complete time-sequential switch state sequence. The three units, through the sequential connection of entry extraction, grouping and sequencing, and sorting and merging within groups, transform the reconfiguration scheme into time-sequential operation instructions that comply with safety regulations, for execution by the dispatch instruction module.

[0057] Example 9: As Figure 4 As shown, based on Embodiment 1, the abnormal signal identification component 3 provided in this embodiment of the invention includes: The error prevention rule base comparison module 31 is used to collect the voltage, current, and position status time-series data of the operating mechanism during live maintenance; and compare the real-time data with the structured rules stored in the error prevention rule base at the millisecond level to determine whether there is any abnormal risk in the current operation command; The structured rules include state conditions, electrical quantity thresholds, timing conditions, and instruction consistency conditions. The abnormal signal identification module 32 is used to execute local interlocking and remote interlocking simultaneously if the real-time collected electrical quantity data and position status meet the triggering conditions of any rule in the anti-misoperation rule base, and record the triggering rule number and timestamp and store them in the operation log. The triggering conditions include, for example, issuing a closing command in the split state and the current mutation rate exceeding the threshold; local interlocking is cutting off the output relay circuit of the operating mechanism; remote interlocking is sending an interlocking signal to the dispatch master station and marking the equipment as interlocked. The output processing module 33 is used to determine that there is a fault in the mechanism if the measured electrical quantity parameters exceed the baseline threshold range of the equipment health status; match the current abnormal parameter characteristics with the constructed knowledge graph, traverse the preset graph database along the relationships such as cause and manifestation, and output the fault cause with the highest matching degree and the corresponding live handling operation guidance to the mobile terminal. Among them, the electrical parameters include the peak value of the coil current and the operating time.

[0058] In the above embodiments, the error prevention rule base comparison module 31, the abnormal signal identification module 32, and the output processing scheme module 33 collaboratively constitute the complete technical path of the abnormal signal identification component 3. During live-line maintenance, the error prevention rule base comparison module 31 collects the voltage, current, and position status time-series data of the operating mechanism. It compares the real-time data with the state conditions, electrical quantity thresholds, timing conditions, and instruction consistency conditions stored in the error prevention rule base at the millisecond level to determine whether the current operation instruction has any abnormal risk. When the real-time collected electrical quantity data and position status meet the triggering conditions of any rule in the error prevention rule base, the abnormal signal identification module 32 simultaneously executes local and remote interlocking, records the triggering rule number and timestamp, and stores it in the operation log. When the measured electrical quantity parameters exceed the baseline threshold range of the equipment health status, the output processing scheme module 33 determines that the mechanism has a fault. It matches the current abnormal parameter characteristics with the knowledge graph, traverses the graph database along relationships such as cause and manifestation, and outputs the fault cause with the highest matching degree and the corresponding live-line processing operation guidance to the mobile terminal. The three modules are connected in a time sequence of real-time comparison, triggering interlocking, and matching output schemes to achieve a full-link response from anomaly identification to interlocking execution and fault handling.

[0059] Example 10: As Figure 5 As shown, based on Examples 1-9, the large-scale power grid security grounding and intelligent dispatching method based on big data provided in this invention specifically includes the following steps: Step S100: Collect and integrate power distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; construct a fault knowledge graph, a fault prevention rule base, and a device health status baseline; The historical fault and maintenance records include the equipment number, model, and operating mechanism type at the time of the fault, a description of the fault phenomenon, an analysis of the fault cause, maintenance measures and post-repair test data, the time of the fault and the interval since the last maintenance, the sequence of operating instructions before and after the fault and their execution results, the operating conditions such as load current and bus voltage at the time of the fault, and environmental conditions such as temperature and humidity. The fault knowledge graph stores the entities and their relationships of fault phenomena, causes, components, etc., from historical fault records, maintenance cases, and expert experience texts in a pre-set graph database. This is used to quickly match and output the fault cause and handling plan by inputting fault phenomena during live-line maintenance. The anti-misoperation rule base is used to convert historical misoperation characteristics into structured judgment conditions for the real-time monitoring module to perform millisecond-level comparisons, so as to automatically trigger interlocking when abnormal operation instructions are identified. The equipment health status baseline includes insulation resistance value, operating mechanism action time, coil current peak value, etc. This is used to provide a normal range reference benchmark for real-time monitoring parameters. When the measured data exceeds the baseline threshold, a mechanism fault warning is triggered, and the knowledge graph is activated to troubleshoot the fault. Step S200: Based on the equipment's historical failure rate, current load, status of nearby equipment, and weather conditions, calculate the necessity and optimal location of safe grounding; generate a power distribution network reconfiguration and load transfer scheme, and coordinate the issuance of maintenance plans and dispatch instructions; Among them, the optimal location is the grounding operation target location with the lowest comprehensive cost selected from all candidate grounding points of the ring main unit by weighted comparison of grounding impedance, load impact range and operation time; the distribution network reconfiguration and load transfer scheme is a combination of switching strategies that minimize the power outage range affected by grounding operation by adjusting switch status and transferring load after determining the grounding point location. Step S300: Continuously monitor the changes in electrical quantities and position status of the operating mechanism, compare them with the anti-misoperation rule base at the millisecond level, and trigger local and remote dual interlocks after identifying abnormal signals; if a mechanism fault is found, input the phenomenon through the mobile terminal, use the knowledge graph to match the fault cause and output a live handling solution. Among them, the real-time monitoring data of electrical quantity changes and position status sampled at a preset sampling frequency during maintenance execution is used to compare with the anti-misoperation rule base to trigger interlocking, and to compare with the health baseline to determine the mechanism fault; the dual interlocking includes redundant safety protection measures that simultaneously cut off the execution circuit on site and send a blocking command to the dispatch master station after abnormal operation signal is detected.

[0060] In the above embodiments, this embodiment constitutes a complete method for safe grounding and intelligent scheduling by sequentially connecting data construction, scheme decision-making, and protection execution. Step S100 transforms unstructured fault text and operational data into three structured knowledge forms: knowledge graph, anti-misoperation rule base, and health baseline, providing a unified data reference for subsequent steps; Step S200 determines grounding requirements based on multi-source status data and selects the point with the lowest comprehensive cost, simultaneously generating a distribution network reconfiguration and load transfer scheme, and issuing maintenance plans and dispatch instructions in a coordinated manner to minimize the impact of grounding operations on the power supply range; Step S300 continuously compares real-time electrical quantities and location status during maintenance execution, identifies abnormal signals, and simultaneously executes local circuit disconnection and remote blocking, while triggering fault warning mechanisms based on the baseline, and matching fault causes and outputting processing schemes through graph matching. Step S100 provides equipment status baseline and map support for step S200. Step S200 outputs the expected operation timing benchmark for step S300. The interlocking trigger record and fault matching result of step S300 are fed back to step S100 to update the anti-misoperation rule base and health baseline, forming a data-driven closed-loop iteration to realize the integration of grounding decision, scheduling coordination and safety protection.

[0061] Figure 6 A block diagram of an exemplary electronic device suitable for implementing embodiments of the present invention is shown.

[0062] The electronic device may include a central processing unit / microprocessor / main control chip 4; and a storage medium 5 coupled to the central processing unit / microprocessor / main control chip 4 and storing computer-executable instructions therein for performing the steps of various methods of embodiments of the present invention when executed by the processor.

[0063] The central processing unit / microprocessor / main control chip 4 may include, but is not limited to, one or more processors or microprocessors.

[0064] Storage medium 5 may include, but is not limited to, random access memory (RAM), read-only memory (ROM), flash memory, EPROM memory, EEPROM memory, registers, computer storage media (e.g., hard disk, floppy disk, solid-state drive, removable disk, CD-ROM, DVD-ROM, Blu-ray disc, etc.).

[0065] In addition, the electronic device may include (but is not limited to) a data bus 6, an input / output bus / external bus / device bus 7, a display 8, and input / output devices 9 (e.g., keyboard, mouse, speaker, etc.).

[0066] The central processing unit / microprocessor / main control chip 4 can communicate with external devices (8, 9, etc.) via wired or wireless networks (not shown) through the input / output bus / external bus / device bus 7.

[0067] The storage medium 5 may also store at least one computer-executable instruction for performing the steps of various functions and / or methods in the embodiments described herein when run by the central processing unit / microprocessor / main control chip 4.

[0068] In one embodiment, the at least one computer-executable instruction may also be compiled into or comprise a software product, wherein one or more computer-executable instructions are executed by a processor to perform the steps of the various functions and / or methods in the embodiments described herein.

[0069] Figure 7 A schematic diagram of a computer-readable storage medium according to an embodiment of the present invention is shown.

[0070] like Figure 7 As shown, the non-transitory computer-readable storage medium 11 stores instructions, such as computer-readable instructions 10. When the computer-readable instructions 10 are executed by a processor, the various methods described above can be performed. The non-transitory computer-readable storage medium includes, but is not limited to, volatile memory and / or non-volatile memory. Volatile memory may include, for example, random access memory (RAM) and / or cache memory. Non-transitory non-volatile memory may include, for example, read-only memory (ROM), hard disk, flash memory, etc. For example, the non-transitory computer-readable storage medium 11 can be connected to a computing device such as a computer, and then, when the computing device executes the computer-readable instructions 10 stored on the non-transitory computer-readable storage medium 11, the various methods described above can be performed.

[0071] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.

[0072] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.

[0073] Those skilled in the art will understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0074] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0075] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.

Claims

1. A large-scale power grid safety grounding and intelligent dispatching system based on big data, characterized in that, include: The distribution ring network cabinet data integration component is used to collect and integrate distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; Construct a fault knowledge graph, a rule base for preventing errors, and a baseline for equipment health status; Adjust the scheduling instruction component to calculate the necessity and optimal location of safe grounding based on the equipment's historical failure rate, current load, status of nearby equipment, and weather conditions; Generate power grid reconfiguration and load transfer plans, and coordinate the issuance of maintenance plans and dispatch instructions; The abnormal signal identification component is used to continuously monitor the changes in electrical quantities and position status of the operating mechanism, compare them with the anti-misoperation rule base in milliseconds, and trigger local and remote dual interlocking after identifying abnormal signals; If a mechanical failure is detected, the user can input the symptoms via a mobile terminal, and the system will use a knowledge graph to match the cause of the failure and output a solution for handling the energized system.

2. The large-scale power grid security grounding and intelligent dispatching system based on big data as described in claim 1, characterized in that, Fault knowledge graphs are used in live-line maintenance to quickly match fault symptoms with fault causes and output solutions. The error prevention rule base is used to transform historical error operation characteristics into structured judgment conditions, which are then compared by the real-time monitoring module at the millisecond level to automatically trigger the interlock when an abnormal operation command is detected. The equipment health status baseline is used to provide a reference benchmark for the normal range of real-time monitoring parameters. When the measured data exceeds the baseline threshold, a mechanism fault warning is triggered, and the knowledge graph is activated to troubleshoot the fault.

3. The large-scale power grid safety grounding and intelligent dispatching system based on big data as described in claim 1, characterized in that, The optimal location was selected from all candidate grounding points of the ring main unit using a three-layer nested nonlinear cost model. The first layer of the model transforms grounding impedance, load influence range and operation time into grounding impedance risk factor, load influence factor and operation exposure factor through saturation mapping, respectively. The load influence factor is determined by the minimum outage section load defined by the adjacency table and the residual load estimated by the reliable transfer margin of the adjacent feeder. The second layer integrates the equipment's historical failure rate, current load, status of neighboring equipment, and lightning strike probability into a working condition stress index, and couples the working condition stress index with the three risk factors into three sub-costs in different nonlinear forms. The third layer uses logarithmic-exponential aggregation to calculate the comprehensive cost based on the adaptive change of the aggregation order with the stress exponent under the operating conditions, and selects the point with the minimum comprehensive cost as the target location for grounding operation; the distribution network reconfiguration and load transfer scheme is a combination of switching strategies that minimize the power outage range affected by grounding operation by adjusting the switch status and transferring the load after determining the grounding point location.

4. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 3, characterized in that, Adjustments to the scheduling instruction components include: The module for calculating the necessity and optimal location of grounding is used to collect historical failure rates, current load values, operating status of nearby equipment, and weather conditions of the ring main unit equipment. The module for generating a distribution network reconfiguration scheme is used to read the current distribution network topology; generate a load transfer scheme, identify the load segments that need to be de-energized due to grounding operations, calculate the transferable capacity of adjacent feeders, traverse all transferable feeders, sort them by remaining capacity from largest to smallest, and then allocate the loads in sequence until all loads to be transferred have been allocated. The dispatch instruction module is used to merge grounding operation instructions, load transfer schemes, and distribution network reconfiguration schemes into a dispatch instruction set; the dispatch instruction set is time-sequenced, with load transfer operations executed first, followed by distribution network reconfiguration, and then grounding operations; The sorted instruction set is sent to the on-site maintenance personnel terminal and the remote dispatch control system, and execution confirmation feedback is received; the dispatch instruction component is adjusted to reassess the current distribution network topology based on the interlocking event information and determine whether the issued load transfer path needs to be updated.

5. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 4, characterized in that, If the current load is greater than a preset percentage of the rated value, the historical failure rate is higher than a preset threshold, and there is an overload in the nearby equipment, then it is determined that grounding is required; otherwise, it is also determined that grounding is required when the probability of lightning strikes exceeds a preset percentage in the meteorological conditions. The system iterates through all groundable points in the ring main unit and calculates the comprehensive cost of each groundable point using a three-layer nested nonlinear cost model. The groundable point with the minimum comprehensive cost is then selected for output. After the power grid reconfiguration and load transfer scheme is executed, the path attenuation parameters used for residual load estimation are corrected by feedback using the measured residual load.

6. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 4, characterized in that, If the remaining capacity of a single feeder is insufficient, multiple feeders are selected simultaneously to share the transferred load, and the load rate of each feeder after allocation is verified; the feeders that pass the verification are used to form a load transfer path; a distribution network reconfiguration scheme is generated; the open and closed states of the sectionalizing switches and tie switches in the topology are adjusted; and the reconfigured switch state sequence is output.

7. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 4, characterized in that, The module for generating distribution network reconfiguration schemes includes: The module for constructing a switch state matrix is ​​used to read the current distribution network topology and construct the switch state matrix; obtain the opening and closing positions of all sectional switches and tie switches in the current distribution network, as well as the current load rate of each feeder; and establish an adjacency table to record the connection relationships between ring main unit nodes and switch station nodes. The transfer path submodule is used to identify feeder sections affected by grounding operations; mark the bus or feeder trunk section where the optimal grounding point is located; list all directly connected sectionalizing switches; and determine the minimum power outage section that needs to be isolated if a grounding operation is performed. The optimal reconfiguration scheme submodule is selected to determine the target open / closed state of the switch after reconfiguration. It simulates each power transfer path, changes the sectionalizing switch at the boundary of the power outage section from closed to open, and changes the target tie switch from open to closed.

8. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 7, characterized in that, Search for available transfer paths for the power outage load section; starting from the boundary switch of the power outage load section, search outwards along the currently closed tie switches, traversing all tie switches that are electrically connected to the power outage section and are currently in the open state; For each candidate tie switch, read the real-time load rate of its target feeder and calculate its remaining transferable capacity; filter tie switches whose remaining capacity is greater than the capacity required by the outage load; the minimum outage section is the outage load.

9. The large-scale power grid safety grounding and intelligent dispatching system based on big data according to claim 7, characterized in that, Verify the simulated power transfer paths; if a path passes the verification, record the new state of the switch corresponding to that path as part of the reconfiguration scheme; compare multiple power transfer paths and select the optimal reconfiguration scheme. If multiple tie switches meet the conditions, compare the total number of reconfiguration operations under each path and the maximum load rate of each feeder after reconfiguration, and select the path with the fewest operations and the lowest maximum load rate as the final distribution network reconfiguration scheme. Output the switch numbers that need to change state and their target state reconstruction schemes; read the reconstruction schemes and extract the list of all switches that need to be opened or closed.

10. A method for large-scale power grid security grounding and intelligent dispatching based on big data, used to implement the system as described in any one of claims 1 to 9, characterized in that, Includes the following steps: Collect and integrate power distribution ring network cabinet operation data, historical fault and maintenance records, environmental data, and dispatch operation logs; construct a fault knowledge graph, a misoperation prevention rule base, and an equipment health status baseline. Based on the equipment's historical failure rate, current load, status of neighboring equipment, and weather conditions, calculate the necessity and optimal location of safe grounding; generate power distribution network reconfiguration and load transfer schemes, and coordinate the issuance of maintenance plans and dispatch instructions; The system continuously monitors changes in electrical quantities and position status of the operating mechanism, performs millisecond-level comparisons with the anti-misoperation rule base, and triggers local and remote dual interlocks after identifying abnormal signals. If a mechanism malfunction is detected, the system inputs the phenomenon through a mobile terminal, uses a knowledge graph to match the cause of the malfunction, and outputs a live handling solution.

Images