A power grid emergency report consistency checking and correcting method and related device

Abstract

This application discloses a method and related apparatus for consistency verification and correction of power grid emergency reports, which is applied in the field of intelligent operation and maintenance of power grids. By acquiring power grid operation data and emergency reports generated by large models, the method extracts the power grid operation events or operation status elements described in the reports and constructs a structured representation. Based on a preset verification dimension, the method performs consistency verification between the structured representation and the actual operation data. When there is an inconsistency between the report and the operation data, the verification results are used to correct the report. This method can effectively eliminate factual illusions and logical deviations in the reports generated by large models, achieve a high degree of consistency between the report content and the actual power grid status, and improve the accuracy and security of emergency reports in unmanned operation or assisted decision-making environments.

Description

Technical Field

[0001] This invention pertains to intelligent operation and maintenance technology for power grids, and particularly relates to a method and related apparatus for verifying and correcting the consistency of power grid emergency reports. Background Technology

[0002] With the continuous advancement of the construction of new power systems, the scale and structure of power grids are becoming increasingly complex, and their operating status is exhibiting greater dynamism and uncertainty. In the event of equipment failure or abnormal operating conditions, dispatchers need to quickly analyze collected remote signaling, telemetry, relay protection information, and on-site feedback to compile emergency reports and upload or publish them. These emergency reports serve as crucial evidence for emergency command and dispatch decisions, and their accuracy and logical consistency directly impact the safety of power grid operation.

[0003] Currently, the compilation of power grid emergency reports is still mainly done manually, requiring the comparison, judgment, and integration of multi-source heterogeneous operational data. This is not only labor-intensive and inefficient, but also prone to problems such as inaccurate equipment status descriptions, numerical omissions, or flawed logical reasoning under high emergency response pressure. To improve document generation efficiency, recent research has attempted to introduce Large Language Models (LLMs) into the field of power grid dispatching to assist in generating emergency reports. However, LLMs mainly rely on linguistic statistical rules for text generation, and their output lacks an intrinsic understanding of the real-time operating status and physical constraints of the power grid, easily leading to inconsistencies with actual operating data. This poses potential application risks in high-safety scenarios such as emergency command.

[0004] While most data verification techniques are effective in improving the reliability of structured numerical data, their application is mainly limited to numerical data itself. They lack the ability to parse and understand natural language text content, making it difficult to effectively verify the authenticity and logical consistency of emergency report content. Summary of the Invention

[0005] Based on this, the present invention aims to propose a method and related device for consistency verification and correction of power grid emergency reports. The method performs correlation analysis between the generated emergency report text and the actual operating status of the power grid, comprehensively verifies the content of the emergency report from multiple dimensions, and introduces real operating data as a basis to correct the generated report content, thereby improving the authenticity, logical correctness and application security of the emergency report.

[0006] In a first aspect, the present invention provides a method for verifying and correcting the consistency of power grid emergency reports, comprising:

[0007] Acquire power grid operation data and emergency reports generated by a preset report generation model;

[0008] Extract the power grid operation events or operation status elements described in the emergency report, and construct a structured representation of the power grid operation events or operation status elements;

[0009] Consistency verification is performed on power grid operation data and structured representation based on preset verification dimensions to obtain verification results;

[0010] The verification result indicates that when there is a discrepancy between the emergency report and the power grid operation data, the emergency report is corrected using the power grid operation data.

[0011] Furthermore, extracting power grid operation events or operational status elements described in emergency reports and constructing structured representations characterizing these events or status elements includes:

[0012] Using a pre-trained entity recognition model, the system identifies power grid operation events or operation status elements described in emergency reports based on a predefined semantic slot structure, thereby obtaining a set of semantic entities.

[0013] Construct a structured representation of power grid operation events or operation status elements using a set of semantic entities.

[0014] Furthermore, consistency checks are performed on the power grid operation data and structured representation based on preset verification dimensions, and the verification results include:

[0015] Based on the preset entity mapping relationship, the verification data source is determined in the power grid operation data according to the structured representation;

[0016] The structured representation and the data source are validated for consistency based on the preset validation dimensions to obtain the validation results.

[0017] The preset verification dimensions include at least one of the following: numerical consistency, physical structure relationship consistency, and event timing consistency.

[0018] Furthermore, consistency verification of structured representations and data sources based on numerical consistency includes:

[0019] The operational attributes associated with power grid operation events are determined based on the structured representation;

[0020] The structured numerical verification function is determined from the preset verification function library based on the running attributes.

[0021] Numerical consistency verification is performed on power grid operation data and structured representation using numerical verification functions.

[0022] Furthermore, consistency verification of structured representations and data sources based on the consistency of physical structural relationships includes:

[0023] The grid topology relationships associated with grid operation events are determined based on the structured representation;

[0024] Power grid operation data includes power grid topology information representing the actual power grid topology, and the power grid topology information is used to verify the consistency between the power grid topology relationship and the actual power grid topology.

[0025] Furthermore, consistency verification of structured representations and data sources based on event-sequence consistency includes:

[0026] Get the preset event timing rules;

[0027] The event time parameters associated with each power grid operation event are determined based on the structured representation, and the time difference of each event time parameter is calculated.

[0028] Determine the event sequence records corresponding to the data source for verification;

[0029] The timing of each power grid operation event is verified to be consistent with the preset timing rules based on the timing difference. The timing of each power grid operation event described in the emergency report is also verified to be consistent with the event sequence record based on the event time parameters.

[0030] Furthermore, consistency checks are performed on the power grid operation data and structured representation based on preset verification dimensions, and the verification results include:

[0031] The number of times the power grid operation data and structured representation are inconsistent under each preset verification dimension is recorded as the number of conflict items;

[0032] The consistency confidence score is calculated by weighting the number of conflicting items corresponding to each preset verification dimension, and recorded as the verification result.

[0033] Secondly, the present invention provides a power grid emergency report consistency verification and correction device, comprising:

[0034] The verification data acquisition module is used to acquire power grid operation data and emergency reports generated by a preset report generation model;

[0035] The structured representation building module extracts the power grid operation events or operation status elements described in the emergency report and constructs a structured representation that characterizes the power grid operation events or operation status elements.

[0036] The consistency verification module is used to perform consistency verification on power grid operation data and structured representation based on preset verification dimensions, and obtain the verification results.

[0037] The report correction module is used to correct the emergency report using the grid operation data when the verification result characterization is inconsistent with the emergency report.

[0038] Thirdly, the present invention provides an electronic device including a memory storing computer-executable instructions and a processor, wherein when the computer-executable instructions are executed by the processor, the device performs the steps of the power grid emergency report consistency verification and correction method provided in the first aspect.

[0039] Fourthly, the present invention provides a readable storage medium storing a computer-executable program, which, when executed, can implement the various steps of the power grid emergency report consistency verification and correction method provided in the first aspect.

[0040] The present invention has the following beneficial effects:

[0041] This invention proposes a method and related apparatus for consistency verification and correction of power grid emergency reports. By simultaneously acquiring power grid operation data and emergency reports generated by a report generation model, and converting the natural language descriptions in the reports into structured representations, unstructured text can be directly compared with structured power grid operation data. This avoids the problem of relying on manual verification of multi-source data line by line in the traditional emergency report compilation process, improving the efficiency of emergency information processing. Furthermore, by utilizing a pre-trained entity recognition model and a predefined semantic slot structure, power grid operation events or operation status elements are extracted from the emergency reports, and corresponding structured representations are constructed. This allows the equipment, operation attributes, and event times claimed in the reports to be mapped to actual power grid operation data, introducing physical world constraints into the content generated by the large model from a technical perspective, effectively reducing equipment status... This invention addresses factual errors in state and operational values. Furthermore, it introduces a multi-dimensional verification dimension, including at least numerical consistency, physical structure consistency, and event timing consistency. This verifies the emergency report content from the perspectives of electrical quantity matching, power grid topology, and event sequence, avoiding misjudgments caused by relying on a single data dimension and making the verification results more comprehensive and reliable. When inconsistencies are detected, power grid operational data is used to correct the emergency report, thus forming a closed-loop processing mechanism of "generation-verification-correction." This invention can automatically verify and correct emergency reports generated by large models without deep human intervention, reducing the risk of scheduling misjudgments caused by inaccurate generated content. It is particularly suitable for applications such as intelligent dispatching, unmanned substations, and intelligent emergency command. Attached Figure Description

[0042] To more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on the provided drawings without creative effort.

[0043] Figure 1A flowchart illustrating the implementation of the power grid emergency report consistency verification and correction method provided in this embodiment of the invention;

[0044] Figure 2 A schematic diagram of the structure of the power grid emergency report consistency verification and correction device provided in an embodiment of the present invention;

[0045] Figure 3 This is an electronic device architecture diagram provided for an embodiment of the present invention. Detailed Implementation

[0046] 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.

[0047] See Figure 1 An embodiment of the present invention provides a method for consistency verification and correction of power grid emergency reports, comprising the following steps:

[0048] Step S110. Obtain power grid operation data and emergency reports generated by a preset report generation model.

[0049] This step simultaneously acquires various data reflecting the actual operating status of the power grid, as well as emergency reports generated by a preset report generation model, providing a foundation for subsequent structured feature extraction and consistency verification.

[0050] Among them, power grid operation data, serving as the factual benchmark, refers to objective operation records from power grid dispatch automation systems, protection information systems, wide-area measurement systems, or other operation monitoring platforms. It possesses the characteristics of objectivity, real-time performance, and traceability, and can truthfully reflect the power grid's operational status and event evolution within a specific time period. The factual benchmark can include continuously sampled time-series data, event trigger records, equipment action logs, operating mode change information, and steady-state cross-sectional data, etc. Its data structure is typically numerical, timestamped, or structured state-based. Power grid operation data serves as a "source of truthfulness reference," and all subsequent consistency judgments are based on this data as the objective basis for determining whether there are any deviations, omissions, or logical inconsistencies in the report content.

[0051] In some implementations, the power grid operation data may undergo further preprocessing operations, including time alignment, outlier removal, key feature extraction, and event segmentation, to form a standardized set of operational fact data suitable for semantic comparison.

[0052] The emergency report generated by the preset report generation model serves as the text to be inspected. This text is automatically generated by a natural language generation model based on operational data, dispatch logs, or manually input information. It is a semantic expression, typically presented in paragraph, sentence, or clause structure, and includes descriptive information such as the event's occurrence time, fault type, scope of impact, handling process, and recovery results. Unlike the power grid operational data, which serves as the factual benchmark, the text to be inspected is not raw objective data, but rather a linguistic abstraction and organized expression of operational facts. It may be affected by factors such as the generation model's understanding ability, training data distribution, or template rules, resulting in deviations in time representation, numerical citations, logical order, or causal relationship descriptions. Therefore, the text to be inspected is considered an object requiring verification, and its accuracy needs to be verified through semantic alignment and numerical comparison with the factual benchmark.

[0053] By clearly defining power grid operation data as the factual benchmark and the generated emergency report as the text to be inspected, the verification object and the reference object can be clearly distinguished at the system architecture level. This avoids the logical ambiguity caused by the mixing of the two types of information sources. At the same time, it provides a clear data backtracking basis for subsequent correction steps. That is, when the verification results show that the text stream to be inspected is inconsistent with the factual benchmark stream, the text content can be corrected in a targeted manner based on the factual benchmark stream, thereby ensuring that the final output emergency report is consistent with the actual operation data in terms of semantic expression.

[0054] Furthermore, power grid operation data refers to multi-source heterogeneous data that can describe the status of power grid equipment, electrical measurements, and network topology. It can come from power dispatch automation systems, energy management systems, or other power grid operation monitoring systems. The data collected by these systems may include remote signaling data, telemetry data, power grid topology information, and operating status data of various electrical equipment.

[0055] Furthermore, remote signaling data includes the open / close position status of circuit breakers and disconnectors, as well as the action signals of relay protection devices, used to reflect the discrete operating status of equipment and protection actions; telemetry data includes real-time measurements of active power, reactive power, bus voltage amplitude, and line current at key sections, used to reflect continuous changes in electrical quantities; and static topology information of the power grid is the basic network connection model, which can be represented as a graph structure, where the node set includes substations, buses, and other key equipment, and the branch set includes lines, transformers, and other connections, used for subsequent physical structure consistency verification.

[0056] In a more preferred embodiment, the power grid operation data includes both time-series data segments of the event process and steady-state snapshots reflecting the final state of the event, thereby ensuring that the verification process can both compare dynamic evolution and verify the accuracy of steady-state results.

[0057] Specifically, time-series data segments are used to depict the dynamic changes of power grid operation events from triggering, development, to completion of handling. Time-series data can include time-series information of electrical quantities, such as bus voltage amplitude and phase angle, line current, active and reactive power, and system frequency variation curves, etc., as they change over time; it can also include time-stamped data reflecting equipment behavior, such as protection device operation records, circuit breaker opening and closing times, automatic reclosing execution status, and stability control device operation information; and it can also include dispatch control operation logs, such as load shedding commands, unit output adjustment records, and reactive power compensation device switching processes, etc. The above data can be extracted based on the event trigger time by setting a time window to trace back and extend forward, thus forming time-series data segments covering the entire event process. In some implementations, unified time alignment processing can be performed on multi-source data, and data from different system sources can be standardized and calibrated through a time synchronization mechanism to ensure time consistency and comparability between different data sources.

[0058] Steady-state snapshot data is used to reflect the stable operating state of the power grid after the event is handled. It can correspond to the event end time, the system recovery judgment time, or the preset stability judgment time point. Its content includes, but is not limited to, bus voltage level, line load rate, system frequency, power balance status, power flow distribution results, and power grid topology information. It can also include the final status of circuit breakers and disconnectors, generator set operating status, load recovery status, and safety constraint verification results, etc.

[0059] In some embodiments, steady-state snapshot data can be obtained based on operational cross-sectional data generated by a scheduling automation system at a specific point in time, or it can be generated based on the steady-state estimation results output by a state estimation algorithm, or a state mirror data can be generated by automatically triggering a data freeze mechanism when preset steady-state determination conditions are met. The steady-state determination conditions may include a frequency change rate lower than a preset threshold, voltage fluctuation amplitude within an allowable range, no new protection actions generated, and completion of control command execution, thereby ensuring the objectivity and consistency of the steady-state data.

[0060] By simultaneously introducing time-series data fragments of the event process and steady-state snapshot data after the event's completion into the same data acquisition step, the obtained power grid operation data can support the verification of both dynamic descriptions in the report regarding the event's occurrence time, development process, and sequence of control actions, and outcome descriptions in the report regarding recovery status, voltage levels, and load recovery ratios. This avoids the potential bias issues that may arise from verification based solely on single outcome data or single process data, thereby improving the reliability and completeness of the entire emergency report consistency verification and correction method. The preset report generation model can be a rule-driven model, a text generation model based on statistical learning, or a large language model based on deep learning. Its input is power grid operation data, and its output is a natural language emergency report describing power grid operation events, fault processes, or equipment status changes. The text content includes a description of the fault process, the name of the faulty equipment, the scope of the power outage, the amount of lost load, and suggested remedial measures, ensuring that the subsequently extracted structured features can fully cover power grid operation events and status information.

[0061] In a further embodiment, to ensure the accurate correspondence between the report text and the power grid operation data in the time dimension, spatiotemporal data alignment can be performed on the two before verification.

[0062] Specifically, since most operational data is continuous time series, while emergency reports are typically generated for specific discrete events, and power grid operational data, as a factual benchmark, usually originates from different business systems or monitoring equipment, its timestamps may be generated based on different time synchronization mechanisms, resulting in sampling differences ranging from milliseconds to seconds. Furthermore, different systems may have different encoding rules for equipment identifiers, line names, or substation names. Emergency report text, as the text to be inspected, often uses natural language to express time, such as "the fault occurred at approximately 10:15" or "the situation was resolved within a few minutes." Such expressions are ambiguous and abstract, making direct numerical matching with precise timestamps impossible. If consistency verification is conducted directly without spatiotemporal alignment, differences in expression may be misjudged as factual deviations, or real inconsistencies may be overlooked, thus affecting the accuracy of the verification results. Therefore, establishing a unified time benchmark system and spatial object mapping system, placing text descriptions and operational data in the same reference coordinate system, is a prerequisite for ensuring the objectivity and accuracy of the verification.

[0063] Regarding time alignment, a unified standard time axis can be established. Specifically, the time synchronization time of the power grid dispatch master station system can be selected as the reference time source, and time standardization processing can be performed on data from different systems. For time-series data in power grid operation data, data with different sampling frequencies can be mapped to a unified time granularity, such as to the second level or a preset time interval, through timestamp resampling, interpolation processing, or time window normalization.

[0064] In some embodiments, different levels of time granularity can be set according to the accuracy requirements of event analysis. High-precision scenarios use a millisecond-level time axis, while regular analysis scenarios use a second-level time axis. For time expressions in the text stream to be examined, relative time expressions such as "at a certain moment," "subsequently," and "minutes later" can be converted into absolute time or time interval expressions using natural language time parsing algorithms and mapped to a unified time axis. If the text uses relative time descriptions, such as "power was restored within 5 minutes after the incident," the time can be extrapolated by combining the baseline event time in the report to generate the corresponding time interval. Through the above processing, time descriptions in the text can be converted into quantifiable time points or time segments, establishing a one-to-one correspondence with the timestamps in the factual baseline stream.

[0065] In terms of spatial alignment, the main issue addressed is the inconsistency in object identification. Power grid operation data typically uses equipment codes, topology node numbers, or asset management system identifiers as spatial location methods, while emergency report texts may use equipment names, abbreviations, or geographical descriptions. For example, an operation data entry might identify a line as a certain number, while the report text might describe it as "a 110kV line from a certain substation to a certain station."

[0066] To achieve spatial alignment, a device identifier mapping table can be constructed to establish a unified mapping relationship between device codes, full names, abbreviations, and historical naming information. In some implementations, a power grid topology model can be used to map all device objects to unified topology nodes or topology edge identifiers, using topology location as a unified spatial reference, thereby avoiding ambiguity issues caused by simple name matching. Furthermore, for cross-regional events or complex events involving multiple devices, a spatial association graph can be established, resolving multiple objects mentioned in the text into a topology association set, which is then matched with the device set in the factual baseline stream.

[0067] In a further embodiment, spatiotemporal alignment can also employ a hierarchical processing mechanism. First, coarse-grained alignment is performed, for example, determining the overall time segment of the event and the range of the main device. Then, fine-grained alignment is performed based on the coarse alignment, for example, precisely matching the specific action time with the specific device action record. The hierarchical alignment mechanism can improve processing efficiency and reduce the probability of mismatch.

[0068] In a further embodiment, the event timestamps recorded in the report text can be parsed. If the text does not explicitly record the event, the time of the report generation request can be used as the event time.

[0069] Set verification time window as follows:

[0070]

[0071] in, This is a lead time margin used to capture fault protection activation signals. This is a post-time tolerance used to capture the steady state after the switch trips.

[0072] Based on verification time window By extracting corresponding time-series data segments and final steady-state snapshots from power grid operation data, the "time of text description" is kept consistent with the "time of physical data sampling," thereby eliminating potential misjudgments in consistency verification due to time deviations.

[0073] Step S120. Extract the power grid operation events or operation status elements described in the emergency report, and construct a structured representation characterizing the power grid operation events or operation status elements.

[0074] This step transforms the unstructured natural language content in the emergency report into a structured representation that can be understood and logically processed by computers. This establishes a bridge between the text and power grid operation data, resolving the inherent differences in expression, semantic level, and data granularity between the text and the factual benchmark based on power grid operation data. This allows the text of the report to be inspected to participate in the verification in a quantifiable and comparable form, thereby ensuring the accuracy and reliability of the consistency judgment.

[0075] Structured representation refers to mapping power grid operation events or operation status elements in a report into formal objects with clear semantic fields and logical relationships, according to a predefined data structure.

[0076] Specifically, emergency reports are often presented in natural language, typically including information such as the time of the incident, the type of fault, the scope of impact, the handling process, and the recovery results. This type of information is semantically abstract and vaguely expressed. For example, the time may be expressed as "approximately 10:15" or "within the following minutes," equipment or line names may have abbreviations, aliases, or geographical descriptions, and the sequence of actions and causal relationships may only be expressed logically in natural language. If this textual information is not structured and directly compared with power grid operation data, it will face problems such as semantic misalignment, mismatched expression granularity, and misjudgments caused by ambiguity.

[0077] Structured representation can transform the event and state elements implicit in text into a unified data model. This can be achieved through explicit definitions such as time fields, device identifier fields, action type fields, and state value fields, giving previously scattered and ambiguous linguistic information a quantifiable expression. This transformation enables cross-modal comparisons in subsequent consistency checks, mapping the "action occurrence time—device—action type" in the text to the corresponding timestamps, device action records, and state changes in the factual benchmark, thus achieving a precise correspondence between textual semantics and power grid operation data.

[0078] Furthermore, structured representation facilitates the systematic analysis of causal relationships, evolution processes, and scope of impact of events. In some embodiments, event elements can be further organized into hierarchical or graph structures. For example, complex events involving multiple devices can be decomposed into unit action nodes, and an event causal network can be constructed through the sequential or topological relationships between nodes. This not only improves the granularity of verification but also provides a scalable data foundation for automatic correction, anomaly detection, and event statistical analysis. Specifically, a pre-trained entity recognition model can be used to parse emergency reports and, combined with a predefined semantic slot structure, identify power grid operation events or operating status elements involved in the report, forming a semantic entity set.

[0079] Furthermore, the semantic slot structure includes slots in multiple dimensions, primarily comprising three core dimensions: subject slots, state slots, and quantity slots. The subject slot identifies the physical equipment entities involved in the report, such as substation names, line names, main transformer numbers, and circuit breaker bay numbers, aiming to clarify the object affected by the event or state and establish a correspondence with equipment identifiers in the power grid topology or asset management system. The state slot extracts predicates from the text describing the operating status of equipment, such as "trip," "close," "loss of voltage," "overload," and "total shutdown." By identifying the semantic information of these actions or states, it can reflect the type of event, the nature of the equipment action, and the process of state change. The quantity slot captures numerical information related to the event or state, including voltage level, load loss, and number of households experiencing power outages, providing a quantitative description of the event's impact or operating indicators, and providing a basis for subsequent numerical comparison with the factual benchmark stream.

[0080] In the semantic slot extraction process, a pre-trained entity recognition model can first be used to perform semantic parsing on the report text, segmenting the text into identifiable entities and predicates. This model can extract information such as device names, action descriptions, and numerical indicators from the text through contextual semantic understanding and pattern recognition. Based on this, combined with predefined semantic slot templates, the extracted entities are mapped to the corresponding slot dimensions. For example, the text "At 10:15, the circuit breaker of XX substation tripped, causing a power outage for 50 households" can be parsed as follows: the main slot is filled with "XX substation circuit breaker", the status slot is filled with "tripped", and the value slot is filled with "power outage for 50 households". Furthermore, timestamps can be associated to form complete event entity units.

[0081] By utilizing a semantic slot structure, the originally scattered and ambiguous natural language information in the report text is uniformly mapped into a set of structured semantic entities. Each semantic entity contains key fields such as subject, state, quantity, and time, which can be directly compared with power grid operation data in the factual benchmark stream. The semantic slot structure not only provides a quantitative representation of the text content but also provides an operational data foundation for subsequent consistency verification and automatic correction, realizing cross-modal mapping from natural language information to structured facts.

[0082] The semantic entity set can include power grid object entities, operation attribute entities, and event time entities. Power grid object entities represent equipment such as substations, lines, main transformers, and circuit breakers; operation attribute entities represent specific operation parameters such as switch positions, active power, and voltage amplitude; and event time entities are used to identify the time of event occurrence or duration.

[0083] In a further embodiment, to enhance the ability to parse complex text, dependency parsing, contextual semantic embedding, or self-attention mechanisms can be used to assist in the identification of semantic entities. When constructing a structured representation, synonyms, numerical units, and state codes are unified to ensure that different expressions of event or device states are mapped to a unified logical object, thereby improving the accuracy of verification.

[0084] In a further embodiment, the semantic entity set is organized into a structured representation according to preset rules. The structured representation can adopt a vector, table, or multi-field data structure. Each record may include information such as device identifier, operating attribute type, status value, and time information, thereby providing a unified data interface for subsequent multi-dimensional consistency verification.

[0085] For example, the vectorized representation of the structured representation is defined as follows:

[0086]

[0087] in, A standardized, unique identifier for the device; For attribute types (e.g., switch position, active power, voltage amplitude); The status value recorded in the emergency report (such as a 0 / 1 logic state or a specific floating-point value); The claimed time when this state occurred.

[0088] For example, for the text "Zhongshan Station 220kV A and B lines tripped due to a fault at 14:00", its structured representation can be vectorized as follows: .

[0089] In a more preferred embodiment, the structured representation not only characterizes a single event or device state, but also logically associates causal relationships, such as fault propagation paths or topological relationships between dependent devices, laying the foundation for topology and timing verification.

[0090] In a further embodiment, to ensure the uniformity of the structured representation and the accuracy of verification, the system can perform synonym normalization processing on the semantic entity set. Specifically, for the descriptions of equipment status, operating attributes, or events in the emergency report, different synonyms or near-synonyms can be mapped to a unified standardized code. For example, descriptions such as "trip," "disconnect," and "open" can be unified as the "disconnect" state of the switch position, and "voltage drop," "low voltage," and "abnormal voltage drop" can be unified as the "low voltage" state.

[0091] In a more preferred embodiment, the unit system, numerical representation method and status mark can be standardized, for example, "kV" and "kilovolt" can be unified, and "MW" and "megawatt" can be unified, so that the structured representation can maintain consistency in subsequent numerical verification and logical judgment.

[0092] Synonym normalization can effectively eliminate the interference of diverse report text descriptions on subsequent numerical consistency, topological relationship and temporal consistency verification, thereby improving the accuracy and stability of the entire closed-loop verification and correction method.

[0093] Step S130. Perform consistency verification on the power grid operation data and structured representation based on the preset verification dimensions to obtain the verification results.

[0094] Specifically, this step uses multi-dimensional verification to determine whether the structured representation in the emergency report is consistent with the power grid operation data, thereby quantifying the report's credibility. Consistency verification involves comparing the structured representation with the power grid operation data in dimensions such as numerical values, physical structural relationships, and event timing to identify deviations or potential errors in the report.

[0095] Specifically, the preset verification dimension refers to the criteria or evaluation index system used to measure the consistency between the text description and the power grid operation data during the verification process. The design of the preset verification dimension should cover multiple aspects such as the time, space, state, quantity and logical relationship between events to ensure the comprehensiveness and accuracy of the consistency judgment.

[0096] In some embodiments, the time dimension can be verified for elements such as the event trigger time, the order of action execution, the event duration, and the recovery time. The time information described in the report text, such as "tripped at 10:15" or "power restored after a few minutes," needs to be compared with the corresponding timestamps and event evolution process in the factual baseline to determine whether the order of events and time intervals are consistent, thus ensuring the accuracy of the dynamic process description.

[0097] The spatial dimension can be used to locate and verify the main entities involved in the event, such as power grid equipment, lines, and substations. By mapping the equipment identifiers in the text body slots to power grid topology nodes and measurement points, the verification dimension can cover single equipment status verification, upstream and downstream topology influence relationships, and the chain reaction of events within a region, thereby ensuring that the spatial objects involved in the text description correspond one-to-one with the actual power grid entities.

[0098] The status dimension can verify the operating status of equipment or nodes, including actions or states such as tripping, closing, undervoltage, overload, and complete shutdown. By comparing the actions described in the status slots in the text with the remote signaling records or action logs in the power grid operation data, the correctness or deviation of the status descriptions in the text can be identified, thereby reflecting the authenticity of the event handling process.

[0099] In terms of numerical values, electrical quantities or statistical indicators can be verified, such as bus voltage, current magnitude, power load, number of households experiencing power outages, and load loss. By comparing the text-based quantity values ​​with the telemetry data of the corresponding measurement points in the factual baseline stream, the degree of impact of events or the accuracy of operational indicators can be verified.

[0100] Logical or topological relationship dimensions can be used to verify causal relationships, topological dependencies, and cascading effects between events. For example, when the text describes "A station's power outage caused B station's power loss," the verification dimension needs to determine whether the time of the power outage at station A and the time of the power loss at station B conform to the causal order, and verify whether the connection relationships between devices on the topological path support this causal link. This logical dimension verification can be used to identify inconsistencies or reasoning errors in complex event descriptions.

[0101] In some embodiments, before verification, the verification data source corresponding to the structured representation in the power grid operation data is determined according to the preset entity mapping relationship, so as to ensure that the data source of the verification process is accurate and reliable.

[0102] Specifically, to ensure a one-to-one correspondence between the structured representation and the actual power grid operation data, a mapping relationship from the text vector space to the physical data space is constructed. This mapping relationship records the correspondence between equipment entities and measurement point or data communication point addresses in the actual power grid topology. For each generated structured representation, including the equipment entity and its associated operating attributes, the measurement point or data communication point address is determined in the mapping relationship based on the equipment entity and its operating attributes, thereby accurately locking the verification data source in the power grid operation data.

[0103] Establishing this mapping relationship is a crucial step in creating a one-to-one correspondence between events and states described in natural language and actual power grid operation data. This ensures the accuracy and traceability of the comparison data sources selected during consistency verification. By maintaining the "entity-measuring point mapping relationship," this relationship records the correspondence between various equipment entities in the power grid topology model and the addresses of remote signaling / telemetry point tables in the power grid dispatch automation system.

[0104] After generating structured semantic entities through semantic slot parsing, each semantic entity carries information such as device identifiers and attribute types. For example, the substation name, line name, or circuit breaker number identified by the main slot, as well as the corresponding action type or indicator category in the status slot or value slot, all serve as core features for vectorized representation. By retrieving mapping relationships, these text vector features are converted into specific remote signaling or telemetry point addresses in the power grid operation data, thereby directly identifying the original data source used for verification and achieving accurate mapping from text semantic information to the physical data space.

[0105] In some embodiments, when the report text involves topological dependencies or causal relationships, such as "a power outage at a substation caused a loss of voltage in a downstream line," not only are the telemetry points of a single device mapped, but the events involved are also associated with the corresponding nodes and branches in the power grid topology model. During this process, the spatial relationship information carried in the semantic entity vectorization representation, such as upstream and downstream relationships, connected branches, and regional affiliation, is located to the set of nodes and edges in the power grid model through topology mapping logic, thereby providing an accurate data index for subsequent multidimensional consistency verification.

[0106] Through the aforementioned mapping mechanism, the abstract semantic information in the report text can be directly mapped to the specific source of the power grid operation data. Whether it's the status verification of a single device or the causal relationship verification involving multiple devices and cross-nodes, accurate and traceable data support can be obtained. This mapping mechanism not only improves the accuracy and reliability of verification but also provides a clear operational objective for subsequent automated correction: when a deviation is found between the text description and the factual baseline, the system can directly locate the corresponding measurement point data and make targeted corrections to the report content based on the actual operation data, ensuring that the generated emergency report is highly consistent with the actual operation of the power grid. Simultaneously, this method also supports the fusion and comparison of cross-system, multi-source data, enabling consistency verification to remain efficient and accurate in complex power grid environments.

[0107] In a further embodiment, the numerical consistency verification aims to verify the degree of numerical matching between the operating attributes in the structured representation and the actual measured values ​​of the power grid. Based on the operating attributes associated in the structured representation, a corresponding function can be selected from a preset verification function library to compare the reported numerical values ​​with the actual measured values ​​in the telemetry data and detect whether there is a deviation.

[0108] Specifically, the operational attributes associated with each power grid operation event are determined based on the structured representation, such as switch position, relay protection signal, line power, voltage amplitude, and bus current, to distinguish between discrete states (such as switch opening / closing, protection action signal) and continuous quantities (such as active power, voltage amplitude, and load). The two are determined by logical matching or relative error threshold.

[0109] For discrete state attributes, such as the opening / closing of circuit breakers and relay protection action signals, the system uses a logic matching function for verification, comparing the state values ​​recorded in the structured representation with the corresponding measurement point values ​​in the verification data source one by one; if the two are inconsistent, it is determined to be a "data conflict". For example, the emergency report describes the actual "opening" state as "closing".

[0110] For example, the following logical matching function can be defined to validate discrete-state data:

[0111]

[0112] in, This indicates the status value recorded in the report text. This represents the actual state value of the power grid operation data. Matching value. A value of 0 indicates a data fact conflict, such as a report text describing the circuit breaker's status as "closed" while the actual status is "open".

[0113] For continuous quantitative attributes, such as active power, reactive power, voltage amplitude, and line current, numerical verification is performed based on a preset relative error threshold.

[0114] Specifically, the state values ​​in the structured representation are compared with the measured values ​​from the corresponding verification data source, and the relative error is calculated:

[0115]

[0116] in, The status value recorded in the emergency report. To verify the actual measurement values ​​of the data source, This is the baseline value or rated capacity; if the relative error exceeds the allowable threshold... If the value is not met, it is determined to be "numerical precision deviation".

[0117] In a further embodiment, the physical structure relationship consistency check is used to verify whether the power grid topology relationship represented in the structured representation conforms to the physical connection of the actual power grid. The check output may include whether the causal topology relationship between devices is reasonable. For example, whether the equipment failure described in the report may lead to the loss of voltage or shutdown of other devices; or whether the power outage range described in the report conforms to the power grid topology constraints and backup power supply connection; or the check result of each topology relationship and the overall topology consistency score, providing a reference for subsequent comprehensive consistency scoring and closed-loop correction.

[0118] Specifically, the grid topology relationships associated with grid operation events are determined based on the structured representation, such as electrical connections between devices, fault propagation paths, and the scope of power supply impact. These grid topology relationships can include connections between nodes, connection status between branches, and power path dependencies of each device in the network. The structured representation is then validated using grid topology information representing the actual topology from the grid operation data.

[0119] For verifying electrical connectivity, the structured representation is mapped to the corresponding nodes and branches in the power grid topology information. Path search is performed in the power grid topology information to verify whether the electrical connections or causal relationships described in the emergency report are reasonable.

[0120] Specifically, connectivity logic verification involves analyzing the electrical connections between devices in the power grid topology diagram and conducting a physical feasibility test on the causal event links involved in the report, thereby determining whether the text description violates the basic physical constraints of power grid operation.

[0121] Taking the report claiming that "the power outage at substation A caused the loss of voltage at substation B" as an example, the main slot vectorized representation extracted from the report text is first mapped to the corresponding nodes in the power grid topology graph G, corresponding to nodes A and B respectively. Then, a path search algorithm, such as breadth-first search (BFS), depth-first search (DFS), or shortest path algorithm, is executed in the power grid topology graph to determine whether there is an electrical connection path between node A and node B.

[0122] The power grid topology graph G mentioned here consists of a set of nodes and a set of edges. Nodes represent equipment such as substations, lines, switches, and main transformers in the power grid, while edges represent electrical connections or branches. This information can be used to completely depict the topology of the power grid.

[0123] During the path search process, starting from node A, the search proceeds layer by layer along the topological edges to check if node B is reachable. If no reachable path is found during the entire search process, the connectivity relationship is determined to be non-existent. For example, the following connectivity criterion can be used:

[0124]

[0125] like If the causal relationship described in the report is not valid in terms of electrical physical structure, it is considered a "physical logic error." For example, if the report describes that downstream substation B lost voltage due to a fault in upstream line A, but in the actual power grid topology, A and B are not directly or indirectly connected through any conductive branch, then the event reasoning in the report has a deviation or error and needs to be marked or corrected.

[0126] In a further embodiment, connectivity logic verification can be further refined by combining branch attributes, device status, and operational constraints. For example, during path search, branches with open switches or lines under maintenance can be excluded, ensuring that connectivity judgments rely not only on topological connectivity but also on actual electrical feasibility. Furthermore, for multi-level causal links or composite events, the system can perform pairwise connectivity checks on each causal node pair in the event sequence, constructing a topological consistency matrix for overall event link consistency evaluation.

[0127] By using connectivity logic verification, potential logical errors in text can be identified from the perspective of physical electrical constraints. This not only uncovers irrationalities in causal descriptions between individual devices but also supports multi-level cascading verification of complex events, thereby ensuring the physical feasibility of emergency reports in terms of topological relationships and providing a reliable basis for subsequent automatic correction and high-level event reasoning. This method achieves cross-modal verification between natural language event descriptions and power grid physical models, enhancing the accuracy and reliability of the verification.

[0128] In a more preferred embodiment, the physical structure relationship consistency verification can further consider the redundant power supply path and network backup branch situation, which is used to determine whether the conclusions in the report regarding the power outage range or power supply status are consistent with the redundancy design principle and actual operation mode of the power grid. This step is mainly used to identify whether the conclusive descriptions in the report such as "complete shutdown", "complete loss of voltage" and "overall shutdown" are consistent with the multi-path power supply structure of the power grid, thereby avoiding the logical error of inferring the overall shutdown based on a single fault.

[0129] Specifically, when emergency report text contains expressions such as "Substation B is completely shut down" or "A certain area is completely without power," the system first uses semantic slot parsing to determine the main equipment B and its operating status description, mapping this main equipment to the corresponding node in the power grid topology model. Then, using this node as the center in the topology diagram, all connected upstream power supply paths are identified, including main power supply lines, backup lines, tie lines, and other power supply branches in the ring network structure. After identifying these paths, real-time operational data is used to retrieve the operating status, circuit breaker status, and load conditions of the corresponding lines for each path to determine if there is a viable path that can maintain power supply.

[0130] In determining the validity of a power supply path, it is necessary to verify not only whether the path is topologically connected, but also whether it is operational. For example, if a spare tie line L is topologically connected to substation B, but its circuit breaker is in the open state, then this path does not constitute a valid power supply path. If tie line L is in the closed state, but its load rate has reached or exceeded the safety limit, then this path may not be able to handle the additional load in actual operation and can also be determined as lacking power supply capacity. Only when the path is in a closed state and the operating parameters meet the safe operating conditions is it considered a valid redundant power supply path.

[0131] If at least one backup power supply path that meets the operating conditions is detected, it indicates that, under the actual operating conditions of the power grid, substation B should theoretically not be in a "complete shutdown" state. If, in this case, the report still concludes with a "complete shutdown," it can be determined that this is inconsistent with the power grid's redundant power supply structure and actual operating data, constituting a logical contradiction.

[0132] In some embodiments, the logical contradiction can be further subdivided into categories such as "redundant path not considered error" or "power supply capacity assessment error" for subsequent correction.

[0133] In a further implementation, power supply path redundancy verification can also be combined with the N-1 safety check principle of the power grid. Specifically, the power flow distribution under a single equipment or line failure condition can be simulated to assess whether the remaining power supply paths can bear the load. If the simulation results show that power supply can still be maintained under a single failure condition, but the report describes it as a complete outage, it indicates that the report's conclusions have not fully considered the redundancy structure or load transfer capability, which is a logical inconsistency.

[0134] By employing a power supply path redundancy verification mechanism, not only can simple device connectivity be verified, but the rationality of the report's conclusions can also be examined from the perspectives of power supply capacity and safe operation principles. This verification method combines the power grid's physical topology, real-time operating status, and safe operation constraints, extending consistency verification from a static judgment of "whether connected" to a dynamic judgment of "whether power supply capacity is available," thereby significantly improving the logical accuracy and verification depth of descriptions of complex power grid operation events.

[0135] For example, if an emergency report states "Substation B is completely down," it is necessary to check whether there is a backup power supply or a connecting line. ,like If the system is running and not overloaded, the conclusion of "complete shutdown" does not conform to the N-1 principle and is judged as "logical contradiction".

[0136] In a further embodiment, the event timing consistency check is used to verify whether the order of occurrence of each power grid operation event in the structured representation conforms to the physical laws of the power system and the preset processing timing rules.

[0137] Specifically, the event time parameters associated with each operational event are extracted based on the structured representation, including the event occurrence time, duration, and triggering conditions. The event time parameters can be derived from the time explicitly described in the emergency report text, or the report generation request time can be used as the default value to ensure that all events have comparable time identifiers.

[0138] In a further embodiment, a preset event timing rule base is obtained, including the sequence of routine fault handling, the delay of protection device operation, and the logic sequence of tripping and closing. By comparing the event time parameters of each structured representation with the reference sequence in the timing rule base, the timing difference between each power grid operation event is calculated to determine whether the event occurrence sequence is reasonable. For example, if the text description of the emergency report is "line voltage loss occurred before the switch tripping", it can be determined as a "timing logic error".

[0139] In a more preferred embodiment, the sequence of events (SOE) recorded in the power grid operation data is further compared to confirm whether the order of occurrence of each power grid operation event in the emergency report is consistent with the actual monitoring data. This comparison can not only discover logical errors in the text description, but also quantify the time deviation, which can be used to calculate the correction gradient in subsequent closed-loop correction.

[0140] In a further embodiment, after completing the consistency verification of multiple verification dimensions, the inconsistent results under each verification dimension can be statistically and weighted.

[0141] Specifically, the number of conflicting items that fail verification in each verification dimension is counted, and different weights are assigned to different verification dimensions. The confidence score of the overall consistency between the structured representation and the power grid operation data is obtained through weighted calculation.

[0142] For example, confidence score The calculation is as follows:

[0143]

[0144] in, The statistic representing conflict type i; Indicates the penalty weight for the type of conflict.

[0145] For example, data factual conflicts and physical logic errors can be assigned high weight (e.g., 20 points / item), time series deviations medium weight (e.g., 10 points / item), and numerical precision deviations low weight (e.g., 5 points / item). If confidence scoring... If the preset threshold is reached, the emergency report is considered qualified and proceeds directly to the output stage; otherwise, the report is corrected.

[0146] In a further embodiment, the statistical results can be recorded in the form of a log, including the specific type of each conflict item, the device entity and operating attributes involved, the corresponding verification data source and deviation value, etc., to provide accurate basis for subsequent correction and to support manual review or auditing.

[0147] Step S140. Verification result characterization: When the emergency report and the power grid operation data are inconsistent, the emergency report is corrected using the power grid operation data.

[0148] This step, based on consistency verification results, corrects inconsistencies or logical deviations in the emergency report, achieving closed-loop correction. Correction refers to adjusting the event states, operational attributes, or timing sequences described in the text according to actual power grid operating data, ensuring that the final report is consistent with the actual power grid operating status in terms of numerical values, topology, and timing, thereby improving the reliability and security of the emergency report.

[0149] Specifically, when the verification results show a deviation between the structured representation and the power grid operation data, the system can first generate a correction prompt or a partial modification suggestion. This prompt includes the conflict type, conflict location, actual data evidence, and corresponding correction instructions. For example, for numerical deviations, the prompt may include the actual measured value and the allowable error range; for topological or causal logic errors, the prompt may clearly identify the incorrect device relationships and the correct topological path; for timing deviations, the prompt may indicate the difference between the actual sequence of events and the textual description.

[0150] In a further embodiment, an iterative closed-loop correction mechanism is adopted to feed back correction prompts to the preset report generation model, so that the model can partially rewrite the text based on the provided measured data, topology information and time sequence rules. After each round of correction, the structured representation construction in step S120 and the consistency verification in step S130 are executed again until the verification result indicates that the emergency report and the power grid operation data are consistent.

[0151] In a more preferred embodiment, to improve efficiency, a partial rewriting strategy is adopted, adjusting the text only for the conflicting parts found during verification, without modifying the content that has already passed verification, thereby reducing unnecessary text changes and maintaining the overall stability of the report. Simultaneously, different conflict types can be assigned different correction priorities; for example, serious numerical and topological errors are treated as high priority, while minor timing deviations are treated as medium priority.

[0152] In a further embodiment, the correction process can combine physical data to generate a correction gradient, guiding the model to generate text that better reflects the actual operating state based on the quantified degree of conflict. For example, for continuous value deviations, the percentage deviation between the actual measured value and the reported claimed value can be calculated, and the prompt can explicitly require the model to adjust the description based on this deviation; for topological logic conflicts, topological paths and node relationships can be injected as factual evidence into the prompt to guide the model to deduce the correct causal relationship.

[0153] Through the above implementation methods, automated error correction of emergency reports can be achieved. Furthermore, through multiple iterations and gradient guidance, the reports can be made highly consistent in three dimensions: numerical values, topology, and time series, forming a closed-loop verification-correction mechanism. This ensures that the accuracy of emergency reports and the safety of the power system are guaranteed even without human intervention or with the assistance of large models.

[0154] The disclosed method can be implemented using various types of devices. Therefore, the present invention also discloses an apparatus corresponding to the above method, and specific embodiments are given below for detailed description.

[0155] like Figure 2 As shown, one embodiment of the present invention provides a power grid emergency report consistency verification and correction device, comprising:

[0156] The verification data acquisition module 202 is used to acquire power grid operation data and emergency reports generated by a preset report generation model.

[0157] The structured representation construction module 204 extracts the power grid operation events or operation status elements described in the emergency report and constructs a structured representation that characterizes the power grid operation events or operation status elements.

[0158] The consistency verification module 206 is used to perform consistency verification on the power grid operation data and structured representation based on a preset verification dimension, and obtain the verification result;

[0159] The report correction module 208 is used to correct the emergency report using the power grid operation data when the verification result characterization is inconsistent with the emergency report and the power grid operation data.

[0160] The device provided in this application embodiment has the same implementation principle and technical effect as the aforementioned method embodiment. For the sake of brevity, any parts not mentioned in the device embodiment can be referred to the corresponding content in the aforementioned method embodiment.

[0161] The methods and related apparatuses mentioned in the above embodiments are described with reference to the method flowcharts and / or structural diagrams provided in the embodiments of this application. Specifically, each block of the method flowchart and / or structural diagram, as well as combinations of blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing device to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing device, generate instructions for implementing the process. Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to operate in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 A schematic diagram of one or more processes and / or structures. Figure 1 The functions specified in one or more boxes. These computer program instructions may also be loaded onto a computer or other programmable data processing apparatus to cause a series of operational steps to be performed on the computer or other programmable apparatus to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable apparatus for implementing the process. Figure 1 A process or multiple processes and / or structures illustrate the steps of the functions specified in one or more boxes.

[0162] The following embodiments illustrate the application of this method to a computer device. It is understood that the computer device can be any device with computing and processing capabilities, including but not limited to servers or personal laptops. In one embodiment, the computer device can be an application server, which can be a server used to run the application under test.

[0163] See Figure 3 This document illustrates a hardware block diagram of an electronic device intended to represent various forms of digital computers, such as laptops, desktop computers, workstations, personal digital assistants, servers, blade servers, mainframes, and other suitable computers. The electronic device may also represent various forms of mobile devices, such as personal digital processors, cellular phones, smartphones, wearable devices, and other similar computing devices. The components shown herein, their connections and relationships, and their functions are merely illustrative and are not intended to limit the implementation of the present application described and / or claimed herein.

[0164] like Figure 3As shown, the electronic device includes: at least one processor 1, at least one communication interface 2, at least one memory 3, and at least one communication bus 4;

[0165] In this embodiment of the application, the number of processor 1, communication interface 2, memory 3, and communication bus 4 is at least one, and processor 1, communication interface 2, and memory 3 communicate with each other through communication bus 4;

[0166] Processor 1 may be a central processing unit (CPU), an application-specific integrated circuit (ASIC), or one or more integrated circuits configured to implement embodiments of the present invention.

[0167] Memory 3 may include high-speed RAM, and may also include non-volatile memory, such as at least one disk storage device;

[0168] The memory stores a program, which the processor can call. The program is used to implement the various processing steps of the aforementioned power grid emergency report consistency verification and correction scheme.

[0169] This invention also provides a readable storage medium storing a computer program thereon. When the computer program is executed by a processor, it implements the various processing flows of the power grid emergency report consistency verification and correction scheme provided in any possible implementation of the above embodiments and / or in combination with the embodiments.

[0170] The invention has been described in particular detail above with respect to possible scenarios, and those skilled in the art will recognize that the invention can be practiced through other embodiments. Specific naming of components, capitalization of terms, attributes, data structures, or any other programming or structural aspects are not mandatory or important, and the mechanisms or features of implementing the invention may have different names, forms, or procedures. The system can be implemented through a combination of hardware and software (as described), entirely through hardware elements, or entirely through software elements. The specific division of functions among the various system components described herein is merely exemplary and not mandatory; rather, the functions performed by a single system component can be performed by multiple components, or the functions performed by multiple components can be performed by a single component.

[0171] Those skilled in the art should understand that the various steps of the disclosed methods can be implemented using general-purpose computing devices. They can be centralized on a single computing device or distributed across a network of multiple computing devices. Optionally, they can be implemented using device-executable program code, which can then be stored in a storage device for execution by the computing device. Alternatively, they can be fabricated as separate integrated circuit modules, or multiple modules or steps can be fabricated as a single integrated circuit module. Therefore, the embodiments disclosed in this invention are not limited to any specific hardware and software combination.

[0172] The programs (also referred to as programs, software, software applications, or code) executable by these computing devices include machine instructions of a programmable processor and can be implemented using high-level procedural and / or object-oriented programming languages, and / or assembly / machine languages. As used herein, the terms “machine-readable medium” and “computer-readable medium” refer to any computer program product, device, and / or apparatus (e.g., disk, optical disk, memory, programmable logic device (PLD)) used to provide machine instructions and / or data to a programmable processor, including machine-readable media that receive machine instructions as machine-readable signals. The term “machine-readable signal” refers to any signal used to provide machine instructions and / or data to a programmable processor.

[0173] Certain aspects of this invention include the process steps and instructions described herein in algorithmic form. It should be noted that the process steps and instructions of this invention can be implemented in software, firmware, and / or hardware, and when implemented in software, they can be downloaded, stored on various operating systems and operated from said platforms.

[0174] Those skilled in the art will understand that the structures shown in the figures are merely block diagrams of some structures related to the present application and do not constitute a limitation on the terminal device to which the present application is applied. Specific terminal devices may include more or fewer components than those shown in the figures, or combine certain components, or have different component arrangements.

[0175] In the description of this specification, the use of terms such as "one embodiment," "some embodiments," "example," "specific example," or "possible design," etc., refers to a specific feature, structure, material, or characteristic described in connection with that embodiment or example, which is included in at least one embodiment or example of this application. In this specification, the illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in a suitable manner in any one or more embodiments or examples. Moreover, without contradiction, those skilled in the art can combine and integrate the different embodiments or examples described in this specification, as well as the features of different embodiments or examples.

[0176] The above embodiments are only used to illustrate the technical solutions of the present invention, and are not intended to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for grid emergency report consistency check and correction, characterized in that, include: Acquire power grid operation data and emergency reports generated by a preset report generation model; Extract the power grid operation events or operation status elements described in the emergency report, and construct a structured representation of the power grid operation events or operation status elements; Consistency verification is performed on power grid operation data and structured representation based on preset verification dimensions to obtain verification results; The verification result indicates that when there is a discrepancy between the emergency report and the power grid operation data, the emergency report is corrected using the power grid operation data.

2. The method of claim 1, wherein, The extraction of power grid operation events or operation status elements described in emergency reports, and the construction of a structured representation characterizing power grid operation events or operation status elements, includes: Using a pre-trained entity recognition model, the system identifies power grid operation events or operation status elements described in emergency reports based on a predefined semantic slot structure, thereby obtaining a set of semantic entities. Construct a structured representation of power grid operation events or operation status elements using a set of semantic entities.

3. The method of claim 1, wherein, The consistency verification of power grid operation data and structured representation based on a preset verification dimension yields the following verification results: Based on the preset entity mapping relationship, the verification data source is determined in the power grid operation data according to the structured representation; The structured representation and the data source are validated for consistency based on the preset validation dimensions to obtain the validation results. The preset verification dimensions include at least one of the following: numerical consistency, physical structure relationship consistency, and event timing consistency.

4. The method of claim 3, wherein, The consistency verification of the structured representation and the data source based on the numerical consistency includes: The operational attributes associated with power grid operation events are determined based on the structured representation; The structured numerical verification function is determined from the preset verification function library based on the running attributes. Numerical consistency verification is performed on power grid operation data and structured representation using numerical verification functions.

5. The method of claim 3, wherein, Consistency verification of structured representation and data source based on the consistency of the physical structural relationships includes: The grid topology relationships associated with grid operation events are determined based on the structured representation; Power grid operation data includes power grid topology information representing the actual power grid topology, and the power grid topology information is used to verify the consistency between the power grid topology relationship and the actual power grid topology.

6. The method of claim 1, wherein, Consistency verification of structured representations and data sources based on the event timing consistency includes: Get the preset event timing rules; The event time parameters associated with each power grid operation event are determined based on the structured representation, and the time difference of each event time parameter is calculated. Determine the event sequence records corresponding to the data source for verification; The timing of each power grid operation event is verified to be consistent with the preset timing rules based on the timing difference. The timing of each power grid operation event described in the emergency report is also verified to be consistent with the event sequence record based on the event time parameters.

7. The method of claim 1, wherein, The consistency verification of power grid operation data and structured representation based on a preset verification dimension yields the following verification results: The number of times the power grid operation data and structured representation are inconsistent under each preset verification dimension is recorded as the number of conflict items; The consistency confidence score is calculated by weighting the number of conflicting items corresponding to each preset verification dimension, and recorded as the verification result.

8. An emergency reporting consistency check and correction device for an electric power grid, characterized by include: The verification data acquisition module is used to acquire power grid operation data and emergency reports generated by a preset report generation model; The structured representation building module extracts the power grid operation events or operation status elements described in the emergency report and constructs a structured representation that characterizes the power grid operation events or operation status elements. The consistency verification module is used to perform consistency verification on power grid operation data and structured representation based on preset verification dimensions, and obtain the verification results. The report correction module is used to correct the emergency report using the grid operation data when the verification result characterization is inconsistent with the emergency report.

9. An electronic device, comprising: It includes a memory storing computer-executable instructions and a processor, which, when executed by the processor, causes the device to perform the power grid emergency report consistency verification and correction method as described in any one of claims 1 to 7.

10. A readable storage medium, characterized by, It stores a computer-executable program that, when executed, can implement the power grid emergency report consistency verification and correction method as described in any one of claims 1 to 7.

Images