Risk-adaptive electric power system operation order anti-error checking method and system

Abstract

The invention provides a risk self-adaptive electric power system operation order anti-error checking method and a risk self-adaptive electric power system operation order anti-error checking system. Comprising the following steps: collecting and solidifying a real-time state of a power grid to generate a space-time reference snapshot, carrying out macroscopic risk identification and correction on an operation order on the basis, and constructing a dynamic trajectory map reflecting operation sequence evolution; potential linkage risks are identified through a sliding scanning map, a risk thermodynamic distribution map is generated, a risk diffusion coefficient matrix is extracted accordingly, and a plurality of risk sensitive sub-scene clusters are divided in a self-adaptive mode. And performing transient simulation verification on each sub-scene cluster in parallel, generating a local adjustment scheme for a risk exceeding scene, and coordinating and integrating the scheme and reconstructing an operation order through iterative optimization circulation until the overall risk is converged. And outputting a multi-dimensional decision report and archiving whole-process data to a knowledge base, so that the operation ticket checking has the capabilities of deep risk prediction, local accurate suppression, global iterative optimization and experience self-evolution.

Description

Technical Field

[0001] This invention relates to the field of power system dispatching and safety control technology, specifically to a risk-adaptive power system operation ticket error prevention verification method and system. Background Technology

[0002] Verification of power system operation tickets to prevent errors is a crucial step in ensuring the safety of power grid switching operations. Traditional verification methods primarily rely on topology-based anti-error interlocking systems with fixed rules. The core of these systems is to perform static logical compliance checks on the operation steps, such as verifying the five-prevention rules. While this method can effectively intercept obvious erroneous operations, its verification granularity remains at the level of matching a single step with the current static topology. It cannot predict or assess the risks of dynamic power flow transfers, transient instability, and cascading failures that may arise during the execution of the operation sequence. Therefore, it represents a passive and superficial safety verification.

[0003] With the increasing proportion of renewable energy connected to the grid and the growing complexity of grid operation, the dynamism and uncertainty in the operation process have significantly increased. A single operation may trigger the propagation of hidden risks far beyond the local topology. Traditional rule-based and static analysis-based verification systems have revealed significant shortcomings: First, they lack the ability to model the spatiotemporal evolution characteristics of risks, making it difficult to locate potential risk nodes hidden in the gaps between operations. Second, the verification process is open-loop, making it impossible to automatically and iteratively optimize operation tickets based on simulation verification results. Third, the system's intelligence level is limited, and historical verification experience cannot be effectively accumulated and used to improve the accuracy and efficiency of future verifications.

[0004] Therefore, existing technologies suffer from prominent problems such as coarse verification granularity, lack of dynamic risk forecasting capabilities, and lack of adaptive learning and continuous optimization mechanisms. Summary of the Invention

[0005] To overcome the problems of coarse verification granularity and lack of dynamic risk prediction and adaptive learning capabilities in the background technology, this invention provides a risk-adaptive power system operation ticket error prevention verification method and system. By constructing a multi-dimensional spatiotemporal benchmark snapshot, performing macro-risk coarse screening and potential risk map positioning, dynamically dividing risk-sensitive sub-scenario clusters and implementing local fine simulation compensation, and driving closed-loop iterative optimization until risk convergence, it realizes the leap from single static rule verification to continuous dynamic risk adaptive control, enabling operation ticket verification to have deep risk prediction, local precise suppression, global iterative optimization and experience self-evolution capabilities, and ultimately upgrading from an error prevention tool to an intelligent safety navigation system.

[0006] The specific technical solution of this application is as follows:

[0007] According to one aspect of this application, a method for verifying error prevention of operation tickets in a risk-adaptive power system is provided, comprising:

[0008] In response to the input of the original operation ticket, the current power grid operating status is collected and fixed to generate a spatiotemporal reference snapshot; based on the spatiotemporal reference snapshot, the operation ticket is subjected to macro-risk identification and preliminary correction to generate a first optimized operation ticket, and a dynamic trajectory map reflecting the evolution of the operation sequence of the first optimized operation ticket is generated; potential cascading risks are identified by sliding scan of the map, and a risk heat map is generated.

[0009] Based on the risk heat map, the risk diffusion coefficient matrix is ​​extracted, and the overall operation trajectory is dynamically divided into multiple risk-sensitive sub-scene clusters according to the gradient clustering characteristics of the risk diffusion coefficient matrix.

[0010] For each risk-sensitive sub-scenario cluster, the transient simulation engine is independently invoked for safety verification. If the risk exceeds the standard, a local risk suppression and adjustment scheme is generated. The adjustment schemes of all sub-scenarios are integrated, and a new round of optimization operation tickets are obtained by reconstruction. Based on the spatiotemporal reference snapshot, the verification loop from macroscopic identification to local simulation is re-executed. Through iterative optimization, the overall risk meets the convergence condition.

[0011] After the verification loop converges, a multi-dimensional decision report is output, which includes risk summary, optimization trajectory, recommended operation sequence and risk inference information. The entire process data is archived to the knowledge base to support the verification of subsequent similar operation tickets.

[0012] As a further option of the method of the present invention, the step of generating a spatiotemporal reference snapshot includes:

[0013] In response to the verification command, the system synchronously issues data acquisition commands with time tags to the dispatch automation system, wide-area measurement system and equipment online monitoring system to collect power grid topology, steady-state telemetry, synchronization phasor and environmental planning information.

[0014] The collected multi-source heterogeneous data is processed by timestamp alignment, consistency fusion and bad data identification to form a structured power grid status dataset;

[0015] The power grid state dataset is encapsulated into a spatiotemporal reference snapshot object containing a snapshot identifier, a reference time, a topology snapshot, state estimation results, and equipment operating parameters.

[0016] As a further option of the method of the present invention, the step of generating dynamic trajectory maps includes:

[0017] Using the spatiotemporal reference snapshot as the initial state node, the power grid topology and power flow state after each operation step in the first optimized operation ticket are simulated sequentially to generate a series of state estimation nodes.

[0018] A directed edge is established between each adjacent state node. The directed edge carries the state change amount caused by the corresponding operation, the operation instruction text, and the step risk increment.

[0019] Each state node is assigned a comprehensive risk index value calculated from voltage and load rate, and all nodes and edges are constructed into a graph structure for storage and visualization.

[0020] As a further option of the method of the present invention, the step of generating the risk heat map includes:

[0021] A sliding time window is set along the dynamic trajectory map, and the power flow transfer and overload risk caused by the superposition of operation and simulated fault within the window is analyzed based on the risk transmission model.

[0022] Record the occurrence window, triggering path, risk path set, and severity of each potential cascading risk event;

[0023] All events are mapped to a two-dimensional grid with the operation step number and the power grid component number as coordinates. The thermal value of each grid point is calculated by the kernel density estimation method to generate a thermal distribution map that reflects the spatiotemporal distribution density of risk.

[0024] As a further option of the method of the present invention, the step of extracting the risk diffusion coefficient matrix and dividing it into sub-scene clusters includes:

[0025] Calculate the correlation coefficient of the time series of thermal values ​​of any two elements in the risk thermal distribution map, and extract the spatial gradient of thermal values ​​along the power grid connection relationship under high-risk operation steps;

[0026] A risk diffusion coefficient matrix is ​​constructed based on the correlation coefficient and spatial gradient. The elements of the risk diffusion coefficient matrix comprehensively characterize the temporal synchronicity and spatial diffusion intensity of risks among components.

[0027] The spectral clustering algorithm is used to cluster the row vectors of the matrix, and elements with similar risk transmission characteristics are divided into the same risk-sensitive sub-scene cluster.

[0028] As a further option of the method of the present invention, the step of performing security verification for each risk-sensitive sub-scenario cluster includes:

[0029] Based on spatiotemporal reference snapshots, local detailed equivalent models are constructed for each sub-scene cluster, and corresponding operation sub-sequences are converted into simulation event sequences and monitoring indicators and safety thresholds are set.

[0030] The transient simulation engine is invoked to execute time-domain simulations of each sub-scenario in parallel, and the transient voltage, power angle and frequency indicators are automatically evaluated to determine whether they exceed the limits.

[0031] For sub-scenarios where risks exceed limits, a local risk suppression and adjustment scheme is generated based on the knowledge base of electromagnetic transient safety adjustment strategies.

[0032] As a further option of the method of the present invention, the step of reconstructing the operation ticket by integrating the adjustment schemes of all sub-scenes includes:

[0033] Conflict detection and coordination are performed on the adjustment schemes generated independently in each sub-scenario. Conflict types include mutual exclusion of operations, resource competition and sequence contradiction.

[0034] Conflicts are resolved based on risk level, impact on global stability, and time adjustability to obtain a set of coordinated adjustment schemes. By inserting, replacing, deleting, or modifying operation steps, the coordinated schemes are applied to the current optimization operation ticket to generate a new round of optimization operation tickets.

[0035] As a further option of the method of the present invention, the convergence condition of the iterative optimization includes at least one of the following:

[0036] The overall risk indicators are below the preset target threshold;

[0037] The change in the overall risk index between two adjacent iterations is less than the convergence tolerance.

[0038] The number of iterations has reached the preset maximum limit;

[0039] If the convergence condition is not met, the verification loop from macroscopic identification to local simulation will be re-executed using a new round of optimization operation tickets and the original spatiotemporal reference snapshot as input.

[0040] Another aspect of this application provides a risk-adaptive power system operation ticket error prevention and verification system, the system comprising:

[0041] The spatiotemporal reference snapshot construction module is used to respond to the input of the original operation ticket, synchronously collect and solidify the current power grid operating status, and generate a spatiotemporal reference snapshot as a unified reference for subsequent verification.

[0042] The macro risk identification and dynamic trajectory construction module is connected to the spatiotemporal benchmark snapshot construction module. It is used to identify and preliminarily correct the operation ticket based on the spatiotemporal benchmark snapshot to generate the first optimized operation ticket, and simulate the evolution process of the operation sequence to map and generate a dynamic trajectory map containing state nodes, directed edges and comprehensive risk index values.

[0043] The cascading risk analysis and heat map generation module is connected to the macro-risk identification and dynamic trajectory construction module. It is used to identify potential cascading risk events based on the risk transmission model by sliding and scanning the dynamic trajectory map, and generate a risk heat map that reflects the spatiotemporal distribution density of the risk.

[0044] The risk scenario adaptive segmentation module is connected to the chain risk analysis and heat map generation module. It is used to extract the risk diffusion coefficient matrix that represents the spatiotemporal correlation and diffusion intensity of risks based on the risk heat map, and dynamically divide the overall operation trajectory into multiple risk-sensitive sub-scenario clusters according to its gradient clustering characteristics.

[0045] The parallel simulation verification and local adjustment module is connected to the risk scenario adaptive partitioning module. It is used to independently call the transient simulation engine for safety verification for each risk-sensitive sub-scenario cluster, and generate a local risk suppression and adjustment scheme when the risk exceeds the standard.

[0046] The iterative optimization control module is connected to the macro-risk identification and dynamic trajectory construction module and the parallel simulation verification and local adjustment module, respectively. It is used to integrate the adjustment schemes of all sub-scenarios to reconstruct a new round of optimization operation tickets, and restart the verification loop based on the spatiotemporal reference snapshot. Through iterative optimization, the overall risk meets the preset convergence conditions.

[0047] The multidimensional decision report generation module is connected to the iterative optimization control module. After the verification loop converges, it outputs a multidimensional decision support report including a risk summary, optimization trajectory comparison, recommended operation sequence, and risk simulation details.

[0048] The knowledge base management module is used to archive the input, process, and output data of the entire process into a structured knowledge base, supporting intelligent verification of subsequent similar operation tickets and system self-learning optimization.

[0049] The beneficial effects of this application are as follows:

[0050] This method fundamentally solves the problem of verification benchmark drift caused by dynamic changes in power grid conditions by constructing a spatiotemporal benchmark snapshot, providing a unified, accurate, and unchangeable initial anchor point for the entire process. This ensures that all subsequent analyses, simulations, and optimizations are performed within a consistent physical time context, significantly improving the reliability and credibility of the verification results and laying a solid foundation for data consistency in the safety assessment of complex operation sequences.

[0051] Compared to traditional linear or isolated verification methods, the core innovation of this invention lies in achieving intelligent risk management through macroscopic identification, graph analysis, scene segmentation, and parallel simulation. By constructing dynamic trajectory graphs and risk heat maps, the evolution of risks is visualized and quantified. Adaptive scene segmentation based on the risk diffusion coefficient matrix decomposes globally complex problems into parallel-processable homoprotic problems, greatly improving the efficiency of computationally intensive tasks such as transient simulations. This method can not only identify static risks but also proactively capture potential chain risks across steps and spaces, achieving a leap from passive verification to proactive early warning in risk prevention and control.

[0052] Through an iterative optimization mechanism, the system can automatically coordinate local adjustment plans, reconstruct operation tickets, and iteratively verify them until the overall risk meets the convergence conditions, thus forming a self-improving decision. The generated multi-dimensional decision reports provide dispatchers with panoramic and traceable operational guidance, while the archiving of the entire process data to the knowledge base accumulates valuable experience, providing a data-driven foundation for the system's future self-learning and continuous optimization. This achieves a paradigm shift in operation ticket error prevention and verification from single-task processing to continuous evolution of intelligent capabilities. Attached Figure Description

[0053] Figure 1 A schematic diagram of the overall process for the risk-adaptive power system operation ticket error prevention verification method;

[0054] Figure 2 Detailed flowchart of steps for the risk-adaptive power system operation ticket error prevention verification method S100;

[0055] Figure 3 Detailed flowchart of steps for the risk-adaptive power system operation ticket error prevention verification method S200;

[0056] Figure 4 Detailed flowchart of steps for the S300 method for verifying the error prevention of operation tickets in risk-adaptive power systems;

[0057] Figure 5 Detailed flowchart of steps for the S400 method for verifying the error prevention of operation tickets in risk-adaptive power systems;

[0058] Figure 6 Detailed flowchart of steps for the S500 method for verifying the error prevention of operation tickets in risk-adaptive power systems;

[0059] Figure 7 Detailed flowchart of steps for the S600 method for verifying the error prevention of operation tickets in risk-adaptive power systems;

[0060] Figure 8 Detailed flowchart of steps for the S700 method for verifying the error prevention of operation tickets in risk-adaptive power systems. Detailed Implementation

[0061] The technical solutions of the embodiments of this application will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of this application, and not all embodiments. Based on the embodiments of this application, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of this application.

[0062] The purpose of this invention is to provide a risk-adaptive closed-loop verification method. Through a comprehensive design encompassing benchmark construction, macroscopic identification, graph analysis, scenario segmentation, simulation verification, and iterative optimization, it achieves full-chain, intelligent risk control for operational tickets. The core innovation of this invention lies in introducing a spatiotemporal benchmark snapshot as the anchor point for the entire verification process, constructing a dynamic trajectory graph to visualize risk evolution, and implementing adaptive scenario segmentation and parallel simulation verification of complex operational sequences based on a risk diffusion coefficient matrix. Finally, through closed-loop iteration, the overall risk of the operational ticket satisfies the convergence condition.

[0063] Example 1

[0064] Please see Figure 1 This illustration shows a schematic flowchart of a risk-adaptive power system operation ticket error prevention verification method according to an embodiment of the present invention. The method mainly includes the following steps:

[0065] S100: In response to the original operation ticket input, collect and solidify the current power grid operating status, and generate a spatiotemporal reference snapshot.

[0066] S200: Based on spatiotemporal benchmark snapshots, macro-risk identification and preliminary correction are performed on the operation tickets, generating the first optimized operation ticket and its corresponding dynamic trajectory map.

[0067] S300: Identifies potential cascading risks by using a sliding scan dynamic trajectory map and generates a risk heat map.

[0068] S400: Extract the risk diffusion coefficient matrix based on the risk heat map, and divide the risk-sensitive sub-scene clusters according to its gradient clustering characteristics.

[0069] S500: For each risk-sensitive sub-scenario cluster, the transient simulation engine is independently invoked for security verification. If the risk exceeds the standard, a local risk suppression and adjustment scheme is generated.

[0070] S600: Integrate the adjustment schemes of all sub-scenarios to reconstruct the operation ticket, and re-execute the verification loop from S200 to S500 based on the spatiotemporal benchmark snapshot, and optimize the overall risk to meet the convergence conditions through iterative optimization.

[0071] S700: After the verification loop converges, it outputs a multi-dimensional decision report and archives all process data to the knowledge base.

[0072] The following provides a detailed explanation of each step.

[0073] like Figure 2 S100, as shown, involves synchronously acquiring and fixing the panoramic data of the power grid at the start of the verification process, forming an unalterable and precise spatiotemporal reference point to eliminate the verification benchmark drift problem caused by dynamic changes in the power grid state. Please refer to... Figure 2 It shows a detailed flowchart of S100.

[0074] S110: Synchronous triggering and collection of multi-source heterogeneous operation data across the entire network.

[0075] In this invention, the verification system master station uniformly publishes data with time information to data sources such as the scheduling automation system, the wide-area measurement system master station, and the equipment online monitoring system. Data acquisition commands from tags. A high-speed data bus ensures command synchronization and concurrent data acquisition. Real-time power grid data at time t_0 or as close as possible to time t_0.

[0076] In one possible implementation, the range of data collected includes: power grid topology information, steady-state telemetry information, synchronization phasor information, and environmental and planning information.

[0077] The system timestamps and command requirements for all collected data. A comparison is performed. For non-strictly synchronous systems such as SCADA, a certain time deviation is allowed. If the timestamp deviation of the key data source exceeds the time deviation, the current acquisition is deemed invalid, and the system waits for the next acquisition cycle to be triggered again, to ensure that all analysis data are aligned as closely as possible to the same physical moment in the time dimension.

[0078] S120: Data consistency fusion and bad data identification and processing.

[0079] In this invention, due to the diverse sources of data, there may be measurement errors, communication anomalies, or even mistakes. Therefore, the collected raw data needs to be cleaned, verified, and fused to form a complete, consistent, and reliable power grid status dataset.

[0080] S130: Build and store spatiotemporal benchmark snapshots.

[0081] In this invention, the aforementioned complete power grid state information is structurally encapsulated to form a spatiotemporal reference snapshot object. .

[0082] In one possible implementation, a spatiotemporal reference snapshot It is a structured data object containing the following key fields: snapshot identifier, baseline time. Topology snapshot, state estimation results, device operating parameters, and relevant context information.

[0083] Starting with a spatiotemporal baseline snapshot, S200 performs a rapid and comprehensive security scan of the operation ticket, identifying and correcting obvious static security issues, while simultaneously constructing a dynamic graph reflecting the evolution of the operation sequence. Please refer to [link / reference]. Figure 3 It shows a detailed flowchart of S200.

[0084] S210: Structured parsing of operation tickets and basic error prevention verification.

[0085] In this invention, the system receives the original operation ticket in natural language or semi-structured text format, first converts it into a machine-understandable structured instruction sequence, and then applies the classic five-prevention logic of the power system for preliminary verification.

[0086] In one possible implementation, the preliminary verification includes the following sub-steps:

[0087] Using an electrical engineering lexicon and named entity recognition technology, the operation ticket text is parsed to identify the equipment object, operation type, and operation parameters.

[0088] The parsed results are converted into a unified internal instruction representation, typically in triplet format: <Device ID, Operation Type, Parameter>.

[0089] The standardized instruction sequence is matched and verified against a pre-set anti-misoperation rule base. The rule base includes, but is not limited to: preventing the opening and closing of disconnect switches under load, preventing the closing of grounding switches while energized, preventing the closing and energizing of grounding switches, and preventing the accidental opening and closing of circuit breakers. The verification process simulates changes in equipment status. If a combination of operations that violates the rules is found, it is immediately marked as a hard error, and a clear description of the violation and a correction prompt are given.

[0090] S220: Static security risk screening and preliminary correction based on rapid trend.

[0091] In this invention, with Starting from the initial state, the network topology after each operation is simulated according to the sequence of operation tickets, and fast power flow calculation is performed to assess whether there are static safety risks such as power flow exceeding limits or voltage exceeding limits, and attempts are made to automatically correct them.

[0092] In one possible implementation, the static security risk screening and preliminary correction based on rapid current flow executes the following process:

[0093] Establish a variable power grid topology model, with initial state and The topology is consistent. Operation commands are processed sequentially, and the topology model is updated in real time according to the command type.

[0094] After each simulation step, the DC power flow method or the fast decomposition power flow method is used to calculate the network power flow distribution and approximate voltage under the new topology.

[0095] The power flow calculation results are compared with the equipment safety limits. The main checks are for equipment overload and voltage exceeding limits.

[0096] For operational steps where risks are detected, the system attaches a risk flag to the instruction triple, including the risk type and severity level. Simultaneously, the system accesses a pre-built static risk correction strategy knowledge base to attempt to generate corrective recommendations.

[0097] After automatic or manual confirmation by the system, the original operation ticket is modified to generate the first optimized operation ticket. .

[0098] S230: Simulation execution and dynamic trajectory map construction.

[0099] In this invention, based on the first optimized operation ticket and spatiotemporal reference snapshot It fully simulates the execution process of the entire operation sequence and constructs a visualized graph structure, namely a dynamic trajectory map, by taking a series of key system state snapshots and their transition relationships during the process.

[0100] In one possible implementation, the dynamic trajectory map construction process is as follows:

[0101] by Initial state node .for The first in Each operation updates the topology and performs power flow calculations to obtain a post-operation system state estimate. and use it as the new state node. The final set of nodes is obtained. ,in This represents the number of operation steps.

[0102] From node arrive Create a directed edge This indicates that an operation has been performed. .

[0103] side The core weight vector is the amount of state change caused by the operation it represents. To facilitate analysis and visualization, extract... The key components are used as edge labels. Simultaneously, the operation instruction text and step risk increments are included. As an additional attribute of the edge.

[0104] For each state node Calculate and attach the comprehensive risk index value. The comprehensive risk index value is calculated from factors such as node voltage and branch load rate.

[0105] The completed map Stored in the form of a graph database or a dedicated data structure.

[0106] The S300 extracts cascading risk paths from dynamic trajectory maps, identifying potential cascading risks arising from multiple operational steps or the combination of operations and potential failures. It then presents the risk distribution density across a two-dimensional time-space plane in an intuitive graphical format. Please refer to [link / reference]. Figure 4 It shows a detailed flowchart of S300.

[0107] S310: Define the sliding analysis window and risk transmission model.

[0108] In this invention, a fixed-length sliding time window is introduced to capture the risk coupling effect across steps in the operation sequence. Simultaneously, a risk propagation model is established to describe how local disturbances propagate in the power grid.

[0109] In one possible implementation, the risk propagation model is based on the fundamental principle that a fault leads to a power flow shift, which may trigger a new overload. For the grid state within the window, the model focuses on: a) changes in power flow distribution caused by the operation itself; and b) how the power flow will be redistributed if any critical component fails and goes out of service under the grid state.

[0110] S320: Sliding scan and detection of potential cascading risk events.

[0111] In this invention, for each sliding window, the system behavior under the combined effect of its internal operation and simulated faults is analyzed to detect whether there is a high-risk mode that could cause overload of subsequent components.

[0112] In one possible implementation, for each window Perform the following tests:

[0113] Calculate the initial state of the window and the end state Changes in the distribution of currents Identify the set of lines where power flow increases significantly due to operations within the window. .

[0114] for Each heavy-load line in Simulate its state at the end of the window. A power outage occurs. Using the power transfer distribution factor matrix, the power flow transfer caused by this fault is quickly calculated. Power transfer distribution factor matrix elements Indicates the line When the line loses 1 unit power, The increase in power.

[0115] The transferred power flow is superimposed on the initial power flow at the end of the window to obtain the predicted power flow after the fault: .examine Are there other lines in the middle? The trend exceeded its safety limits. A certain proportion. If it exists, record a potential cascading risk event. .

[0116] For each detected event Record its characteristics, including: the occurrence window number. Triggering circuit A set of lines that may be at risk of overload The severity of the overload and the specific operational steps involved.

[0117] S330: Generate a risk heat map.

[0118] In this invention, the spatial and temporal location information of all detected chain risk events is mapped onto a two-dimensional grid, and a continuous and intuitive risk intensity distribution map is generated through kernel density estimation.

[0119] In one possible implementation, a risk heat map

[0120] A discrete two-dimensional grid is formed by using the operation step number as the horizontal axis and the numbers of all monitored lines or busbars of the power grid as the vertical axis.

[0121] For each risk event Its occurrence window An interval corresponding to the x-axis The risk routes involved This corresponds to a specific location on the vertical axis. For each grid cell affected by the event, the severity score of that event is accumulated.

[0122] Kernel density estimation is performed on discrete event mapping results. Grid points. The heat value at the point It is obtained by weighted summation of the contributions of all events to this point: ;in, It is an event The severity It is a two-dimensional kernel function. Grid points To the event The spatial and temporal distance at the center.

[0123] The calculated The matrix is ​​normalized and mapped onto a chromatogram to generate the final risk thermal distribution map.

[0124] Based on the risk clustering characteristics revealed by the risk heatmap, S400 intelligently decomposes complex verification problems covering the entire operation process and the entire network into multiple sub-problems with strong internal risk homogeneity and weak inter-problem coupling. This creates conditions for parallel processing of subsequent time-consuming fine simulations. Please refer to [reference needed]. Figure 5 It shows a detailed flowchart of the S400.

[0125] S410: Extract the risk diffusion coefficient matrix based on the risk heat map.

[0126] In this invention, the risk heat map contains information about the correlation of risks in both the spatial and temporal dimensions. By analyzing the spatial distribution pattern and temporal evolution sequence of heat values, a matrix can be constructed to characterize the strength of risk correlation between different power grid locations.

[0127] In one possible implementation, from the risk heatmap extract Risk diffusion coefficient matrix The steps are as follows:

[0128] Calculate the risk heatmap matrix In, any two elements and The corresponding thermodynamic value time series and correlation coefficient .

[0129] In the highest-risk operation steps, the spatial gradient of the thermal value along the grid connection is calculated.

[0130] Risk diffusion coefficient matrix elements This combines the above-mentioned temporal correlation and spatial transitivity:

[0131] ;

[0132] in, and These are weighting coefficients; It is a component and Electrical distance between them; It is the attenuation constant; It is a component The highest thermal value throughout the entire process; It is the threshold of risk significance; It is an indicator function. The first term of the formula captures temporal synchronicity, and the second term captures the spatial diffusion effect from high-risk sources.

[0133] S420: Automatic sub-scene segmentation based on gradient spectrum clustering.

[0134] In this invention, the matrix Each row is considered as a point in the risk diffusion feature space. Using a clustering algorithm, points with similar features are grouped together to form sub-scene clusters.

[0135] In one possible implementation, a spectral clustering algorithm is used to group points with similar features into one class, thereby forming sub-scene clusters.

[0136] S500 is an inspection performed on each sub-scene cluster divided by S400. It uses high-precision time-domain simulation to verify whether the system can withstand various disturbances during the execution of operation sub-sequences, and generates specific rectification plans for situations that do not meet safety requirements. Please refer to [link / reference]. Figure 6 It shows a detailed flowchart of the S500.

[0137] S510: Sub-scene simulation task preparation and parallel distribution.

[0138] In this invention, each sub-scene cluster Prepare complete, independent transient simulation task packages and utilize the parallel computing capabilities of the computing cluster to execute multiple simulations simultaneously.

[0139] In one possible implementation, for a sub-scenario Simulation task preparation includes: taking snapshots based on a spatiotemporal reference. Based on the complete model in [the original text], construct a cluster-focused [system / mechanism]. A detailed local model. The corresponding state variables are precisely set to each element of the local equivalent model, and the operation subsequence is... Convert the data into an event sequence executable by the simulation software, and set the occurrence time of each event. Define the risk indicators that need to be monitored and evaluated in the simulation, and set their safety thresholds. Package the above model, data, events, and monitoring settings into a single simulation task file.

[0140] S520: Transient simulation execution and automatic evaluation of safety criteria.

[0141] In this invention, at each computing node, a commercial or self-developed transient simulation program reads the task file, performs numerical integration to solve a system of differential-algebraic equations, and simulates the dynamic process of the power grid. After the simulation is completed, it automatically extracts the time-domain waveform data of predefined monitoring quantities from the output file. Based on the time-domain waveform data, key safety indicators are calculated, such as transient voltage drop area, maximum power angle difference, and minimum frequency point.

[0142] The calculated indicators are compared with the set safety thresholds. If any indicator exceeds its threshold, the sub-scene is deemed invalid. The risk exceeds the limit under the current operating procedure. Otherwise, it is judged as passing the verification. The judgment results and detailed indicator values ​​of each sub-scenario are summarized and sent back to the main control program.

[0143] S530: Generate a local risk suppression and adjustment scheme.

[0144] In this invention, for sub-scenarios that are determined to exceed the risk limit, the simulation results are analyzed to locate the root cause of the problem, and targeted local adjustment schemes are generated by matching from the strategy knowledge base.

[0145] In one possible implementation, the system is connected to an electromagnetic transient safety adjustment strategy knowledge base. The knowledge base is organized as a mapping from failure / risk modes to response strategies. Based on the diagnosed root cause mode, all possible candidate strategies are retrieved. For each retrieved candidate strategy, [the system is then combined with...]. The specific network parameters and operating context are adapted.

[0146] S600 integrates the local adjustments to all sub-scenarios, reconstructs a globally superior new operation ticket, and then restarts the verification process for validation. Through multiple cycles of evaluation-adjustment-re-evaluation, the overall risk level of the operation ticket is continuously reduced until it meets the preset convergence criteria. Please refer to [reference needed]. Figure 7 It shows a detailed flowchart of the S600.

[0147] S610: Conflict detection and coordination fusion of multi-source local adjustment schemes.

[0148] In this invention, adjustment schemes are generated independently for different sub-scenes. and There may be conflicts between them, and coordination is necessary to merge them into a single operation ticket.

[0149] In one possible implementation, the conflict types and coordination rules include:

[0150] The two schemes require mutually exclusive operations to be performed on the same device at the same time. Coordination rules: prioritize the scheme proposed in the sub-scenario with the higher risk level; or adopt the scheme that is more beneficial to global stability after evaluation; or avoid conflicts by fine-tuning the execution time of the schemes.

[0151] Both solutions require the use of the same scarce resource. The coordination rule is to arrange the order of resource usage or choose a comprehensive solution that can cover the needs of multiple sub-scenarios.

[0152] Option A requires operation X to precede operation Y, while option B implicitly assumes Y precedes X. The coordination rule is to analyze the security reasons behind the two sequence constraints and find a new sequence that satisfies all critical security constraints.

[0153] After resolution, a set of coordinated adjustment schemes is obtained. .

[0154] S620: Operation Ticket Reconstruction Algorithm.

[0155] In this invention, the coordinated adjustment scheme set Optimization operation tickets applied to the current round Generate the next round of optimization operation tickets. .

[0156] In one possible implementation, the operation ticket reconstruction is a process of traversal and editing, specifically including:

[0157] If the solution requires a certain existing operation If a new preparatory or auxiliary operation sequence is inserted before or after it, these steps are inserted at the corresponding positions.

[0158] If the plan suggests using a safer alternative... Replace the original If so, then replace it.

[0159] If the scheme evaluation determines that a certain operation If the risk is extremely high and cannot be mitigated by adjustments, it is recommended to cancel; in this case, delete this step.

[0160] If the solution suggests modifying a parameter for an operation, then update that parameter.

[0161] S630: Iterative convergence judgment and loop control.

[0162] In this invention, a newly generated operation ticket is used. And the original spatiotemporal reference snapshot As input, restart a new round of verification process from S200 to S500. In the new round of S520, the risk indicator values ​​for each sub-scenario under the new invoice will be obtained. .

[0163] Define the overall risk metric $R^{(l+1)}$ for this iteration. It can be defined as the maximum value of the risk metrics for all sub-scenarios, or it can be the weighted average of the sub-scenarios based on their importance.

[0164] The convergence condition for an iterative process is typically set to one of the following three:

[0165] Condition 1: Risk meets the standard. ,in It is a preset target risk threshold, indicating that the risk has been reduced to an acceptable level.

[0166] Condition 2: The change converges: ,in It is the preset convergence tolerance.

[0167] Condition 3: Reaching the maximum number of iterations: Number of iterations The preset maximum limit has been reached. .

[0168] The system performs a judgment after each iteration. If condition 1 or 2 is met, the iteration is considered converged, and the process returns to S700 to output the final result. If only condition 3 is met, the iteration is forcibly terminated, the system outputs the current optimal result, but a warning of incomplete convergence is given. If no convergence condition is met, then... Return to S610 and continue the next round of conflict resolution-reconstruction-verification optimization cycle.

[0169] The S700 is the endpoint of the verification process. It extracts, integrates, and consolidates valuable information generated throughout the iterative optimization process. On one hand, it provides dispatchers with decision support reports that can be directly used to guide operations; on the other hand, it transforms the experience gained from this verification into knowledge, empowering future intelligent verifications. Please refer to [link / reference]. Figure 8 It shows a detailed flowchart of the S700.

[0170] S710: Generate multi-dimensional decision support reports.

[0171] In this invention, the report is presented in a clear, structured, and visual manner, covering comprehensive information from global risk assessment to specific operational instructions.

[0172] In one possible implementation, the system automatically integrates and formats the entire process verification data to generate a structured, multi-dimensional decision support report. The report's first page is an execution summary, clearly stating the verification task objectives, number of iterations, and final conclusions, and prominently marking the overall risk level and the elimination of key risks. The main body of the report visually displays the dynamic trajectory graphs and risk heatmaps of the original and final operation tickets through side-by-side comparison charts, and plots the curves showing the decrease in overall risk indicators with the number of iterations. The core content of the report is the final recommended operation ticket text with detailed change markers. All steps that have been inserted, deleted, replaced, or modified are highlighted with brief explanations of the reasons for the changes, and operation monitoring prompts are added for steps with residual risks. The report also includes details of key risk simulations, summarizing in tabular form the sub-scenarios where high risks occurred and their adjustment strategies and effects, accompanied by the most representative transient simulation waveform comparison charts for verification. Finally, the report provides a list of recommended operation executions and contingency plans, listing the plants and lines requiring close monitoring, and providing concise backup operation guidelines for potential anomalies.

[0173] S720: Full-process data archiving and case knowledge base updates.

[0174] In this invention, all input, process, and output data generated in this verification task are packaged, labeled, and stored in a standardized format to form reusable case knowledge.

[0175] In one possible implementation, the system standardizes, encapsulates, and archives the entire lifecycle data of this verification task, and uses this data to update the case knowledge base. The specific process is as follows: First, a structured data package is created, containing case metadata, an input set, a process dataset, and an output set. Then, the system assigns multi-dimensional feature labels to the case and stores it in a dedicated case database. Based on the archived case data, the system can further perform knowledge extraction; simultaneously, the accumulated case data can also be used to train machine learning models to improve the intelligence and efficiency of initial risk prediction and scenario classification for new operational tickets in the future, thereby achieving self-learning and continuous optimization of the system.

[0176] Example 2

[0177] A risk-adaptive power system operation ticket error prevention and verification system, the system includes:

[0178] The spatiotemporal reference snapshot construction module is used to respond to the input of the original operation ticket, synchronously collect and solidify the current power grid operating status, and generate a spatiotemporal reference snapshot as a unified reference for subsequent verification.

[0179] The macro risk identification and dynamic trajectory construction module is connected to the spatiotemporal benchmark snapshot construction module. It is used to identify and preliminarily correct the operation ticket based on the spatiotemporal benchmark snapshot to generate the first optimized operation ticket, and simulate the evolution process of the operation sequence to map and generate a dynamic trajectory map containing state nodes, directed edges and comprehensive risk index values.

[0180] The cascading risk analysis and heat map generation module is connected to the macro-risk identification and dynamic trajectory construction module. It is used to identify potential cascading risk events based on the risk transmission model by sliding and scanning the dynamic trajectory map, and generate a risk heat map that reflects the spatiotemporal distribution density of the risk.

[0181] The risk scenario adaptive segmentation module is connected to the chain risk analysis and heat map generation module. It is used to extract the risk diffusion coefficient matrix that represents the spatiotemporal correlation and diffusion intensity of risks based on the risk heat map, and dynamically divide the overall operation trajectory into multiple risk-sensitive sub-scenario clusters according to its gradient clustering characteristics.

[0182] The parallel simulation verification and local adjustment module is connected to the risk scenario adaptive partitioning module. It is used to independently call the transient simulation engine for safety verification for each risk-sensitive sub-scenario cluster, and generate a local risk suppression and adjustment scheme when the risk exceeds the standard.

[0183] The iterative optimization control module is connected to the macro-risk identification and dynamic trajectory construction module and the parallel simulation verification and local adjustment module, respectively. It is used to integrate the adjustment schemes of all sub-scenarios to reconstruct a new round of optimization operation tickets, and restart the verification loop based on the spatiotemporal reference snapshot. Through iterative optimization, the overall risk meets the preset convergence conditions.

[0184] The multidimensional decision report generation module is connected to the iterative optimization control module. After the verification loop converges, it outputs a multidimensional decision support report including a risk summary, optimization trajectory comparison, recommended operation sequence, and risk simulation details.

[0185] The knowledge base management module is used to archive the input, process, and output data of the entire process into a structured knowledge base, supporting intelligent verification of subsequent similar operation tickets and system self-learning optimization.

[0186] Those skilled in the art will understand that the embodiments of this application are provided as methods, systems, or computer program products. Therefore, this application takes the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, this application takes the form of a computer program product implemented on one or more computer storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer program code. The solutions in the embodiments of this application are implemented using various computer languages, exemplified by the object-oriented programming language Java and the interpreted scripting language JavaScript.

[0187] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, are implemented by computer program instructions. These computer program instructions are provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, create means for implementing the functions specified in the flowchart illustrations and / or block diagrams.

[0188] These computer program instructions are also stored in a computer read-memory that can direct a computer or other programmed data processing device to operate in a particular manner, such that the instructions stored in the computer read-memory produce an article of manufacture including instruction means that implement the functions specified in the flowchart or multiple flowcharts and / or block diagram blocks or multiple block diagrams.

[0189] These computer program instructions are also loaded onto a computer or other programming data processing device to cause a series of operational steps to be performed on the computer or other programming device to produce a computer-implemented process, such that the instructions, which execute on the computer or other programming device, provide steps for implementing the functions specified in the flowchart flow or multiple flows and / or the block diagram blocks or multiple blocks.

[0190] Although preferred embodiments of this application have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments as well as all changes and modifications falling within the scope of this application.

[0191] Obviously, those skilled in the art can make various modifications and variations to this application without departing from the spirit and scope of this application. Therefore, if such modifications and variations fall within the scope of the claims of this application and their equivalents, this application also intends to include such modifications and variations.

Claims

1. A method for verifying error prevention in operation tickets of a risk-adaptive power system, characterized in that, The method includes: In response to the input of the original operation ticket, the current power grid operating status is collected and fixed to generate a spatiotemporal reference snapshot; based on the spatiotemporal reference snapshot, the operation ticket is subjected to macro-risk identification and preliminary correction to generate a first optimized operation ticket, and a dynamic trajectory map reflecting the evolution of the operation sequence of the first optimized operation ticket is generated; potential cascading risks are identified by sliding scan of the map, and a risk heat map is generated. Based on the risk heat map, the risk diffusion coefficient matrix is ​​extracted, and the overall operation trajectory is dynamically divided into multiple risk-sensitive sub-scene clusters according to the gradient clustering characteristics of the risk diffusion coefficient matrix. For each risk-sensitive sub-scenario cluster, the transient simulation engine is independently invoked for safety verification. If the risk exceeds the standard, a local risk suppression and adjustment scheme is generated. The adjustment schemes of all sub-scenarios are integrated, and a new round of optimization operation tickets are obtained by reconstruction. Based on the spatiotemporal reference snapshot, the verification loop from macroscopic identification to local simulation is re-executed. Through iterative optimization, the overall risk meets the convergence condition. After the verification loop converges, a multi-dimensional decision report is output, which includes risk summary, optimization trajectory, recommended operation sequence and risk inference information. The entire process data is archived to the knowledge base to support the verification of subsequent similar operation tickets.

2. The method according to claim 1, characterized in that, The step of generating a spatiotemporal reference snapshot includes: In response to the verification command, the system synchronously issues data acquisition commands with time tags to the dispatch automation system, wide-area measurement system and equipment online monitoring system to collect power grid topology, steady-state telemetry, synchronization phasor and environmental planning information. The collected multi-source heterogeneous data is processed by timestamp alignment, consistency fusion and bad data identification to form a structured power grid status dataset; The power grid state dataset is encapsulated into a spatiotemporal reference snapshot object containing a snapshot identifier, reference time, topology snapshot, state estimation results, and equipment operating parameters.

3. The method according to claim 2, characterized in that, The steps for generating the dynamic trajectory map include: Using the spatiotemporal reference snapshot as the initial state node, the power grid topology and power flow state after each operation step in the first optimized operation ticket are simulated sequentially to generate a series of state estimation nodes. A directed edge is established between each adjacent state node. The directed edge carries the state change amount caused by the corresponding operation, the operation instruction text, and the step risk increment. Each state node is assigned a comprehensive risk index value calculated from voltage and load rate, and all nodes and edges are constructed into a graph structure for storage and visualization.

4. The method according to claim 1, characterized in that, The steps for generating the risk heat map include: A sliding time window is set along the dynamic trajectory map, and the power flow transfer and overload risk caused by the superposition of operation and simulated fault within the window is analyzed based on the risk transmission model. Record the occurrence window, triggering path, risk path set, and severity of each potential cascading risk event; All events are mapped to a two-dimensional grid with the operation step number and the power grid component number as coordinates. The thermal value of each grid point is calculated by the kernel density estimation method to generate a thermal distribution map that reflects the spatiotemporal distribution density of risk.

5. The method according to claim 1, characterized in that, The steps of extracting the risk diffusion coefficient matrix and dividing it into sub-scene clusters include: Calculate the correlation coefficient of the time series of thermal values ​​of any two elements in the risk thermal distribution map, and extract the spatial gradient of thermal values ​​along the power grid connection relationship under high-risk operation steps; A risk diffusion coefficient matrix is ​​constructed based on the correlation coefficient and spatial gradient. The elements of the risk diffusion coefficient matrix comprehensively characterize the temporal synchronicity and spatial diffusion intensity of risks among components. The spectral clustering algorithm is used to cluster the row vectors of the matrix, and elements with similar risk transmission characteristics are divided into the same risk-sensitive sub-scene cluster.

6. The method according to claim 1, characterized in that, The steps for performing security verification for each risk-sensitive sub-scenario cluster include: Based on spatiotemporal reference snapshots, local detailed equivalent models are constructed for each sub-scene cluster, and corresponding operation sub-sequences are converted into simulation event sequences and monitoring indicators and safety thresholds are set. The transient simulation engine is invoked to execute time-domain simulations of each sub-scenario in parallel, and the transient voltage, power angle and frequency indicators are automatically evaluated to determine whether they exceed the limits. For sub-scenarios where risks exceed limits, a local risk suppression and adjustment scheme is generated based on the knowledge base of electromagnetic transient safety adjustment strategies.

7. The method according to claim 1, characterized in that, The steps for reconstructing the operation ticket by integrating the adjustment scheme of all sub-scenes include: Conflict detection and coordination are performed on the adjustment schemes generated independently in each sub-scenario. Conflict types include mutual exclusion of operations, resource competition and sequence contradiction. Conflicts are resolved based on risk level, impact on global stability, and time adjustability to obtain a set of coordinated adjustment schemes. By inserting, replacing, deleting, or modifying operation steps, the coordinated schemes are applied to the current optimization operation ticket to generate a new round of optimization operation tickets.

8. The method according to claim 1, characterized in that, The convergence condition for the iterative optimization includes at least one of the following: The overall risk indicators are below the preset target threshold; The change in the overall risk index between two adjacent iterations is less than the convergence tolerance. The number of iterations has reached the preset maximum limit; If the convergence condition is not met, the verification loop from macroscopic identification to local simulation will be re-executed using a new round of optimization operation tickets and the original spatiotemporal reference snapshot as input.

9. A risk-adaptive power system operation ticket error prevention and verification system for performing the method as described in any one of claims 1 to 8, characterized in that, The system includes: The spatiotemporal reference snapshot construction module is used to respond to the input of the original operation ticket, synchronously collect and solidify the current power grid operating status, and generate a spatiotemporal reference snapshot as a unified reference for subsequent verification. The macro risk identification and dynamic trajectory construction module is connected to the spatiotemporal benchmark snapshot construction module. It is used to identify and preliminarily correct the operation ticket based on the spatiotemporal benchmark snapshot to generate the first optimized operation ticket, and simulate the evolution process of the operation sequence to map and generate a dynamic trajectory map containing state nodes, directed edges and comprehensive risk index values. The cascading risk analysis and heat map generation module is connected to the macro-risk identification and dynamic trajectory construction module. It is used to identify potential cascading risk events based on the risk transmission model by sliding and scanning the dynamic trajectory map, and generate a risk heat map that reflects the spatiotemporal distribution density of the risk. The risk scenario adaptive segmentation module is connected to the chain risk analysis and heat map generation module. It is used to extract the risk diffusion coefficient matrix that represents the spatiotemporal correlation and diffusion intensity of risks based on the risk heat map, and dynamically divide the overall operation trajectory into multiple risk-sensitive sub-scenario clusters according to its gradient clustering characteristics. The parallel simulation verification and local adjustment module is connected to the risk scenario adaptive partitioning module. It is used to independently call the transient simulation engine for safety verification for each risk-sensitive sub-scenario cluster, and generate a local risk suppression and adjustment scheme when the risk exceeds the standard. The iterative optimization control module is connected to the macro-risk identification and dynamic trajectory construction module and the parallel simulation verification and local adjustment module, respectively. It is used to integrate the adjustment schemes of all sub-scenarios to reconstruct a new round of optimization operation tickets, and restart the verification loop based on the spatiotemporal reference snapshot. Through iterative optimization, the overall risk meets the preset convergence conditions. The multidimensional decision report generation module is connected to the iterative optimization control module. After the verification loop converges, it outputs a multidimensional decision support report including a risk summary, optimization trajectory comparison, recommended operation sequence, and risk simulation details. The knowledge base management module is used to archive the input, process, and output data of the entire process into a structured knowledge base, supporting intelligent verification of subsequent similar operation tickets and system self-learning optimization.

Images