Power distribution network risk auxiliary decision-making method, system and equipment and medium

Abstract

The invention discloses a power distribution network risk auxiliary decision-making method, system and device and a medium. The method comprises the steps of collecting power grid multi-source data; performing protocol unified conversion on the power grid multi-source data, converting various heterogeneous communication protocols into a unified data format, performing real-time analysis and anomaly detection on the converted power grid multi-source data through an edge calculation layer, and generating edge layer early warning information; the edge layer early warning information and the key detection data are uploaded to a cloud platform layer for deep data analysis, historical operation data are utilized to construct a power distribution network digital twin model, and equipment fault risk assessment data and spatial risk distribution data are generated; and performing multi-channel early warning publishing on the power distribution network risk assessment result and the emergency disposal scheme. According to the power distribution network risk auxiliary decision-making system with the five-layer architecture, the comprehensiveness and the real-time performance of power distribution network risk monitoring in extreme weather are improved through multi-source heterogeneous data fusion and edge-cloud cooperative processing.

Description

Technical Field

[0001] This invention relates to the field of distribution network risk decision-making technology, and in particular to a distribution network risk auxiliary decision-making method, system, device and medium. Background Technology

[0002] With global climate change, extreme weather events (such as typhoons, rainstorms, and ice storms) are becoming increasingly frequent, posing a serious threat to the safe and stable operation of power distribution networks. Under extreme weather conditions, power distribution networks face problems such as increased equipment failures and decreased power supply reliability. Traditional power distribution network risk monitoring and decision-making systems have many shortcomings. On the one hand, existing systems struggle to effectively integrate multi-source heterogeneous data; data from sensors and monitoring equipment are isolated, failing to comprehensively and accurately reflect the operational status of the power distribution network under extreme weather conditions. On the other hand, decision support lacks intelligence and precision, relying heavily on human experience, making it difficult to assess risks accurately and in real-time and generate scientific emergency plans. For example, faced with complex meteorological conditions and equipment operating parameters, traditional systems cannot quickly correlate and analyze them, leading to untimely warnings and inappropriate response plans. Furthermore, traditional systems have significant shortcomings in data processing efficiency, risk assessment accuracy, and the relevance of emergency plans, failing to meet the high requirements for the safe operation of power distribution networks under extreme weather conditions. Summary of the Invention

[0003] In view of the aforementioned existing problems, the present invention is proposed.

[0004] Therefore, this invention provides a method, system, device, and medium for assisting decision-making on distribution network risks to solve the problems of traditional distribution network risk assisting decision-making systems, such as difficulty in integrating multi-source heterogeneous data, lack of intelligent and precise decision support, low data processing efficiency and accuracy of risk assessment, and poor targeting of emergency plans.

[0005] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0006] In a first aspect, the present invention provides a risk-assisted decision-making method for power distribution networks, comprising the following steps:

[0007] Collect multi-source data from the power grid;

[0008] The multi-source data of the power grid is subjected to unified protocol conversion, which converts various heterogeneous communication protocols into a unified data format. The converted multi-source data of the power grid is then analyzed and anomaly detected in real time through the edge computing layer to generate edge layer early warning information.

[0009] The edge layer early warning information and key detection data are uploaded to the cloud platform layer for in-depth data analysis. Historical operation data is used to build a digital twin model of the power distribution network and generate equipment failure risk assessment data and spatial risk distribution data.

[0010] The equipment failure risk assessment data and spatial risk distribution data are correlated with extreme weather failure patterns. Distribution network risk assessment results and emergency response plans are generated through distribution network risk knowledge graphs and digital twin models.

[0011] The risk assessment results and emergency response plans for the power distribution network will be released to the public through multiple channels.

[0012] As a preferred embodiment of the distribution network risk auxiliary decision-making method described in this invention, the step of collecting multi-source data of the power grid includes:

[0013] The controller collects current, voltage, and temperature rise parameters of the power distribution network equipment to obtain the operating data of the power distribution network equipment;

[0014] Meteorological monitoring data are obtained by collecting wind speed, wind direction, rainfall, temperature, and air pressure through regional micro-meteorological stations;

[0015] The AI-powered recognition terminal identifies abnormal device states and obtains abnormal device detection data.

[0016] Mobile monitoring equipment is used to inspect and monitor power distribution network facilities and obtain inspection and testing data.

[0017] Location data is obtained by collecting the real-time location of power distribution equipment and mobile workers through positioning terminals;

[0018] The collected information is integrated and output as multi-source power grid data.

[0019] As a preferred embodiment of the distribution network risk auxiliary decision-making method described in this invention, the steps of real-time analysis and anomaly detection of multi-source data of the power grid include:

[0020] Protocol parsing is performed on multi-source power grid data in a unified data format to extract data fields from the multi-source power grid data;

[0021] Build a streaming computing engine to perform real-time processing and statistical analysis on the data fields through time windows;

[0022] By using a preset safety threshold rule base, rule matching and risk assessment are performed on the power distribution network equipment operation data, meteorological monitoring data, and equipment anomaly detection data;

[0023] When the detection parameters exceed the preset safety threshold, the edge layer warning information of the corresponding risk level is triggered, and the corresponding key detection data is marked.

[0024] The advantages of this preferred technical solution are: reducing data transmission latency through real-time edge processing, improving early warning response speed, and enabling timely detection of abnormal trends through streaming computing and time window analysis, thus preventing the escalation of faults.

[0025] As a preferred embodiment of the distribution network risk auxiliary decision-making method described in this invention, the step of performing deep data analysis includes:

[0026] The edge layer early warning information and key detection data are encrypted and then uploaded to the cloud platform layer;

[0027] The key detection data are analyzed spatially using a geographic information system to generate a geographical distribution risk map of power distribution network equipment.

[0028] The current key detection data and historical operation data are correlated and analyzed using a distributed computing platform to train the equipment fault prediction algorithm model;

[0029] The fault probability value of each power distribution network device is output through the equipment fault prediction algorithm model to form the equipment fault risk assessment data.

[0030] Regional spatial risk distribution data are generated by combining geographical risk maps.

[0031] The beneficial effects of this preferred technical solution are: encryption ensures data transmission security and prevents the leakage of sensitive information; geographic information system analysis enables spatial visualization of risks and facilitates rapid location of problem areas.

[0032] As a preferred embodiment of the distribution network risk auxiliary decision-making method described in this invention, the step of constructing the distribution network digital twin model includes:

[0033] Collect historical operation data, fault record data, and network topology data of the power distribution network;

[0034] A mathematical simulation model of the distribution network equipment is established based on the historical operating data to simulate the operating characteristics of the real distribution network. The model is then trained and a digital twin model of the distribution network is output.

[0035] Input the current meteorological monitoring data and equipment operating status into the digital twin model of the power distribution network;

[0036] The operation response of the distribution network is simulated using the digital twin model of the distribution network, and the simulation results of the distribution network are output.

[0037] The beneficial effects of this preferred technical solution are: the digital twin model can simulate the operating characteristics of a real power distribution network, providing a simulation environment close to reality, and the model trained based on historical operating data has learning capabilities, which can continuously optimize the prediction accuracy.

[0038] As a preferred embodiment of the distribution network risk auxiliary decision-making method of the present invention, the steps for generating the distribution network risk assessment results and emergency response plans include:

[0039] Construct a risk knowledge graph for the power distribution network and establish the correlation between equipment types, meteorological factors, failure modes, and emergency strategies;

[0040] The equipment failure risk assessment data and spatial risk distribution data are input into the power distribution network risk knowledge graph for knowledge reasoning and failure mode matching.

[0041] By combining the results of knowledge reasoning and fault mode matching with the simulation results of the digital twin model of the power distribution network, the key equipment and fault propagation paths affected by extreme weather are determined.

[0042] Based on the fault mode matching results and simulation analysis, a distribution network risk assessment result is generated.

[0043] Based on the risk assessment results of the power distribution network, corresponding emergency resource allocation and response steps are matched to generate an emergency response plan.

[0044] The beneficial effects of this preferred technical solution are: the knowledge graph establishes a systematic relationship, realizes intelligent fault mode identification, and, combined with simulation results, can accurately determine the scope of fault impact and propagation path, thereby improving the pertinence of emergency response.

[0045] As a preferred embodiment of the distribution network risk auxiliary decision-making method of the present invention, the step of disseminating the distribution network risk assessment results and emergency response plans through multiple channels includes:

[0046] Based on the risk level in the power distribution network risk assessment results, early warning information is generated and disseminated through communication channels.

[0047] The monitoring interface displays the real-time operating status of the power distribution network and the progress of emergency response plans.

[0048] Receive emergency response execution status and equipment status update information transmitted back by on-site maintenance personnel via mobile terminals;

[0049] The effectiveness of the emergency response is evaluated based on the feedback data from the implementation.

[0050] Secondly, the present invention provides a power distribution network risk auxiliary decision-making system, including a data acquisition module, an edge computing module, a cloud platform analysis module, a digital twin module, and an early warning release module:

[0051] The data acquisition module is responsible for collecting power distribution network equipment operation data, meteorological monitoring data, and location data, and outputting multi-source power grid data.

[0052] The edge computing module performs unified protocol conversion on the multi-source data of the power grid, and performs real-time analysis and anomaly detection on the converted multi-source data of the power grid to generate edge layer early warning information;

[0053] The cloud platform analysis module performs in-depth data analysis on the edge layer early warning information and key detection data to generate equipment failure risk assessment data and spatial risk distribution data.

[0054] The digital twin module associates equipment failure risk assessment data and spatial risk distribution data with extreme weather failure patterns, and generates distribution network risk assessment results and emergency response plans through the distribution network risk knowledge graph and digital twin model.

[0055] The early warning release module will release the risk assessment results and emergency response plans of the power distribution network through multiple channels.

[0056] Thirdly, the present invention provides an electronic device, comprising:

[0057] Memory and processor;

[0058] The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the power distribution network risk auxiliary decision-making method.

[0059] Fourthly, the present invention provides a computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the power distribution network risk auxiliary decision-making method.

[0060] Compared with the prior art, the beneficial effects of the present invention are as follows:

[0061] A five-layer architecture-based power distribution network risk support decision-making system enhances the comprehensiveness and real-time performance of power distribution network risk monitoring under extreme weather conditions through multi-source heterogeneous data fusion and edge-cloud collaborative processing. A protocol conversion gateway enables unified access for devices from different manufacturers and using different protocols, resolving the issues of data silos and poor compatibility in traditional systems. The real-time processing capabilities of the edge computing layer effectively reduce data transmission latency, while the deep analysis of the cloud platform layer provides robust computational support for risk assessment, forming a highly efficient collaborative architecture that combines real-time edge response with intelligent cloud analysis.

[0062] Driven by both a distribution network risk knowledge graph and a digital twin model, a fundamental shift from traditional experience-based decision-making to intelligent scientific decision-making has been achieved. The knowledge graph establishes systematic relationships between equipment, weather, and faults, enabling the system to automatically identify complex fault propagation patterns. The digital twin model provides a near-realistic simulation environment, offering a scientific basis for the formulation and verification of emergency plans. The resulting closed-loop mechanism of "monitoring-decision-response-feedback" not only improves the efficiency and accuracy of emergency response but also continuously enhances the system's intelligence level through continuous learning and optimization, providing reliable technical support for the safe and stable operation of the power grid under extreme weather conditions. Attached Figure Description

[0063] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0064] Figure 1 This is a schematic diagram of the overall process of the power distribution network risk auxiliary decision-making method according to an embodiment of the present invention.

[0065] Figure 2 This is a structural interaction diagram of a power distribution network risk auxiliary decision-making method according to an embodiment of the present invention.

[0066] Figure 3 This is a data acquisition flowchart of a power distribution network risk auxiliary decision-making method according to an embodiment of the present invention. Detailed Implementation

[0067] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0068] Example 1, referring to Figure 1 As an embodiment of the present invention, a distribution network risk auxiliary decision-making method is provided, comprising the following steps S1 to S5:

[0069] S1. Collect multi-source data from the power grid;

[0070] S2. Perform unified protocol conversion on multi-source data of the power grid, convert various heterogeneous communication protocols into a unified data format, and perform real-time parsing and anomaly detection on the converted multi-source data of the power grid through the edge computing layer to generate edge layer early warning information.

[0071] S3. Upload edge layer early warning information and key detection data to the cloud platform layer for in-depth data analysis, use historical operation data to build a digital twin model of the power distribution network, and generate equipment failure risk assessment data and spatial risk distribution data;

[0072] S4. The equipment failure risk assessment data and spatial risk distribution data are correlated with extreme weather failure patterns. The distribution network risk assessment results and emergency response plans are generated through the distribution network risk knowledge graph and digital twin model.

[0073] S5. Disseminate the risk assessment results and emergency response plans of the power distribution network through multiple channels.

[0074] It should be noted that with the intensification of global climate change, extreme weather events such as typhoons, rainstorms, and ice storms are becoming increasingly frequent, posing a serious threat to the safe and stable operation of power distribution networks. Under extreme weather conditions, power distribution networks face problems such as increased equipment failures and decreased power supply reliability. Traditional power distribution network risk monitoring and decision-making systems have many shortcomings. On the one hand, existing systems struggle to effectively integrate multi-source heterogeneous data; data generated by various sensors and monitoring equipment are isolated and cannot comprehensively and accurately reflect the operational status of the power distribution network under extreme weather conditions. On the other hand, decision support lacks intelligence and precision, relying heavily on manual experience and judgment, making it difficult to assess risks accurately and in real time and generate scientific emergency plans. For example, faced with complex meteorological conditions and changes in equipment operating parameters, traditional systems cannot quickly perform correlation analysis, leading to untimely warnings and unreasonable response plans, affecting the safe operation of the power grid and the reliability of power supply.

[0075] Therefore, to address the aforementioned difficulties in data fusion and insufficient intelligent decision-making, a five-layer architecture for distribution network risk auxiliary decision-making is constructed through steps S1-S5. This system enables unified access and real-time processing of multi-source heterogeneous data, improves data processing efficiency and early warning response speed through an edge-cloud collaborative architecture, and achieves intelligent identification and correlation analysis of fault modes under extreme weather conditions based on a distribution network risk knowledge graph and a digital twin model. This automatically generates scientific and reasonable emergency response plans. Simultaneously, a complete risk management system is formed through multi-channel early warning dissemination and a closed-loop feedback mechanism, significantly improving the safety and reliability of the distribution network under extreme weather conditions and providing intelligent technical support for the stable operation of the power grid.

[0076] Example 2, refer to Figures 1-3 As an embodiment of the present invention, a distribution network risk auxiliary decision-making method is provided based on the above embodiment.

[0077] In this embodiment of the application, step S1, the acquisition of multi-source power grid data, includes steps A1 to A6:

[0078] A1. Collect current, voltage, and temperature rise parameters of power distribution network equipment through the controller to obtain the operating data of the power distribution network equipment;

[0079] A2. Collect wind speed, wind direction, rainfall, temperature, and air pressure through regional micro-meteorological stations to obtain meteorological monitoring data;

[0080] A3. Identify abnormal device states through AI recognition terminals and obtain device anomaly detection data;

[0081] A4. Conduct inspections and monitoring of power distribution network facilities using mobile monitoring equipment to obtain inspection and testing data;

[0082] A5. The real-time location of power distribution equipment and mobile workers is collected through the positioning terminal to obtain location data;

[0083] A6. Integrate the collected information and output it as multi-source power grid data.

[0084] Specifically, in step A1, the controller includes a programmable logic controller (PLC) deployed on key equipment such as transformers and circuit breakers. It collects real-time operating parameters such as current, voltage, and temperature rise rate every 50ms and transmits the data via the Modbus protocol. The accuracy requirements for data acquisition are: current measurement error ≤ ±0.5% and voltage measurement error ≤ ±0.2%.

[0085] Specifically, in step A2, the regional micro-weather station integrates multiple sensors, including wind speed sensors, rain gauges, temperature and humidity sensors, and is deployed in high-risk areas such as mountainous areas and coastal areas. It collects data every 10 seconds and has a built-in GPS module for accurate positioning. The measurement range is: wind speed 0-70m / s, rainfall 0-500mm / h, and temperature -40~+80℃.

[0086] Specifically, in step A3, the AI ​​recognition terminal is equipped with a target detection algorithm and high-definition cameras are installed in key sections of substations and transmission lines to identify abnormal equipment conditions such as insulator damage, conductor galloping, and tower tilting in real time. The abnormality detection results are transmitted in real time via the RTSP protocol.

[0087] Specifically, in step A4, the mobile monitoring equipment includes drones and mobile inspection vehicles. The drones are equipped with infrared thermal imagers, visible light cameras, and lidar. They automatically inspect the power transmission lines according to preset routes. When an anomaly is detected, they automatically hover and transmit high-definition image data. They support switching between manual and automatic modes.

[0088] Specifically, in step A5, the positioning terminal integrates a Beidou / GPS dual-mode positioning module and is installed on power distribution equipment, emergency repair vehicles, and mobile work personnel equipment. It collects location information in real time at a frequency of 1Hz, with a positioning accuracy of ≤3 meters. Combined with electronic fence technology, it realizes real-time monitoring and trajectory recording of equipment and personnel.

[0089] It should be noted that in step A6, the data integration process achieves unified format conversion through the data acquisition gateway, supporting adaptive recognition and conversion of multiple communication protocols such as Modbus, TCP / IP, MQTT, and LoRa, ensuring that data from different manufacturers and models of equipment can be seamlessly accessed into the system.

[0090] In one optional implementation, the collection of multi-source power grid data in step S1 can also be preprocessed by an edge smart gateway, including data cleaning, format standardization, timestamp synchronization, and other functions, to reduce redundant data transmission and improve data quality and overall system processing efficiency.

[0091] In another optional implementation, the collection of multi-source power grid data in step S1 can also be combined with an Internet of Things (IoT) sensor network, deploying distributed sensor nodes, and using self-organizing network technology to achieve data aggregation and intelligent routing nearby, thereby enhancing the reliability and coverage of data collection in complex environments.

[0092] In this embodiment of the application, step S2, which involves real-time analysis and anomaly detection of multi-source data from the power grid, includes steps B1 to B4:

[0093] B1. Perform protocol parsing on multi-source power grid data in a unified data format and extract the data fields from the multi-source power grid data;

[0094] B2. Build a streaming computing engine to perform real-time processing and statistical analysis of data fields through time windows;

[0095] B3. Through a preset safety threshold rule base, perform rule matching and risk assessment on power distribution network equipment operation data, meteorological monitoring data, and equipment anomaly detection data;

[0096] B4. When the detection parameters exceed the preset safety threshold, trigger the edge layer warning information of the corresponding risk level and mark the corresponding key detection data.

[0097] Specifically, in step B1, the protocol parsing uses a parsing algorithm that combines regular expressions and finite state machines to perform in-depth parsing of multi-source data from the power grid, extracting key data fields such as device ID, timestamp, monitoring values, status identifiers, and data quality codes. It supports dynamically expanding parsing rules and generates standardized JSON data formats for easy subsequent processing.

[0098] Specifically, in step B2, a distributed streaming computing engine is built based on the Apache Flink framework. A sliding time window is set for real-time data processing. The window size is 1 minute and the sliding interval is 10 seconds. Data fields are filtered, aggregated, and statistically analyzed in real time to calculate statistical parameters such as the real-time average, maximum, minimum, and standard deviation of the power distribution network equipment operation data.

[0099] Specifically, in step B3, the preset safety threshold rule base is constructed based on expert experience and historical fault data. It includes a multi-dimensional rule set such as equipment operation threshold rules, meteorological warning rules, and equipment anomaly identification rules. An improved Rete algorithm is used for efficient rule matching, which supports compound condition judgment, such as "when the wind speed is ≥25m / s, the rainfall is ≥50mm / h, and the line current fluctuation is ≥15%, a line tripping risk warning is triggered".

[0100] Specifically, in step B4, the risk level is divided into three levels: normal, warning, and emergency. When the detection parameter exceeds the preset safety threshold, the system automatically calculates the risk score and determines the warning level, generates edge layer warning information containing the anomaly type, risk level, affected equipment, and timestamp, and marks the key detection data that triggers the warning for priority uploading. The warning information generation time is ≤5 seconds.

[0101] It should be noted that the above-mentioned real-time parsing and anomaly detection mechanisms effectively reduce data transmission pressure and improve the real-time performance of anomaly detection by using edge computing for local processing. The sliding window design of the streaming computing engine can capture the temporal change characteristics of data, and the dynamic update mechanism of the preset rule base ensures the accuracy and adaptability of anomaly detection.

[0102] In an optional implementation, in step S2, real-time parsing and anomaly detection can also be combined with machine learning algorithms to identify changes in data patterns by training anomaly detection models, including the Isolation Forest algorithm to detect data anomalies and the LSTM neural network to predict data trend anomalies, thereby improving the ability to identify unknown anomaly patterns.

[0103] In another optional implementation, in step S2, real-time parsing and anomaly detection can also adopt an adaptive threshold adjustment mechanism, which dynamically adjusts the safety threshold parameters according to the statistical characteristics of historical operating data and seasonal variation patterns, so as to avoid false alarms and missed alarms caused by fixed thresholds and improve the accuracy of anomaly detection.

[0104] In the embodiments of this application, step S3, the step of performing deep data analysis, includes C1 to C5:

[0105] C1. After encrypting the edge layer early warning information and key detection data, upload them to the cloud platform layer;

[0106] C2. Analyze the spatial location of key detection data using a geographic information system to generate a geographical distribution risk map of power distribution network equipment;

[0107] C3. Use a distributed computing platform to perform correlation analysis between current key detection data and historical operating data, and train the equipment fault prediction algorithm model.

[0108] C4. Output the failure probability value of each distribution network equipment through the equipment failure prediction algorithm model to form equipment failure risk assessment data;

[0109] C5. Generate regional spatial risk distribution data by combining geographical risk distribution maps.

[0110] Specifically, in step C1, the edge layer early warning information and key detection data are encrypted using the AES-256 symmetric encryption algorithm. The encryption key is securely distributed using the RSA-2048 asymmetric algorithm. A secure communication tunnel between the edge computing layer and the cloud platform layer is established through the TLS1.3 protocol. Digital signature verification and integrity verification are supported during data transmission.

[0111] Specifically, in step C2, a distribution network geographic information system is constructed based on the ArcGIS platform, integrating high-precision electronic maps, equipment spatial coordinates, and topographic data. Spatial interpolation analysis is performed on key detection data, and continuous risk contour maps are generated using the Kriging interpolation algorithm. Combined with attribute information such as equipment type, voltage level, and load importance, a multi-level geographical distribution risk map of distribution network equipment is generated.

[0112] Specifically, in step C3, a big data analytics environment is built based on the Hadoop / Spark distributed computing platform to perform deep correlation analysis on the current key detection data and historical operating data from the past three years. Multiple machine learning algorithms, including random forest, gradient boosting tree, and neural networks, are used for integrated modeling. The training data includes multi-dimensional features such as equipment operating parameters, meteorological data, fault records, and maintenance history. Model training employs cross-validation and grid search for hyperparameter optimization.

[0113] In C4, the specific calculation form of the equipment failure prediction algorithm model is as follows:

[0114]

[0115] In the formula, P fault (t) represents the failure probability of the equipment at time t; w i f is the weight coefficient of the i-th feature; i (X t ) represents the value of the i-th characteristic function at time t; n is the total number of characteristics; α is the maintenance time decay coefficient; T maintainβ is the time since the last maintenance; γ is the time decay constant; W is the extreme weather impact coefficient; W extreme The current extreme weather index; W normal This is the baseline value for normal weather.

[0116] It should be noted that in the formula This is a basic fault probability calculation term based on multidimensional features. To account for the time decay correction term that considers the impact of maintenance intervals on equipment health status, γ·max(0,W) extreme -W normal () represents a risk amplification correction term under extreme weather conditions.

[0117] Specifically, in step C5, the regional risks are classified by combining the geographical distribution risk map and equipment failure risk assessment data, and the distribution network coverage area is divided into three levels: low-risk area (green), medium-risk area (yellow), and high-risk area (red). The spatial risk distribution data includes information such as risk level, impact range, list of key equipment, and expected impact time. The risk area is dynamically updated and visualized through GIS spatial analysis function.

[0118] It should be noted that the aforementioned in-depth data analysis, through multi-dimensional data fusion and intelligent algorithm modeling, has achieved a leap from single-point monitoring to regional risk assessment. The spatial analysis capabilities of the geographic information system provide intuitive support for the visualization of risks, and the distributed computing platform ensures the real-time performance and accuracy of large-scale data processing.

[0119] In an optional implementation, in step S3, deep data analysis can also introduce digital twin technology to construct a virtual simulation model of the distribution network. By integrating the physical model with the data-driven model, the dynamic response process of the distribution network under extreme weather conditions can be simulated, thereby improving the interpretability of the physical mechanism of fault prediction and the accuracy of prediction.

[0120] In another optional implementation, in step S3, deep data analysis can also employ federated learning technology to integrate historical operating data of multiple regional power distribution networks for collaborative modeling while protecting data privacy, thereby improving the model's generalization ability and adaptability to new regions and new equipment types, and enhancing the system's scalability.

[0121] In this embodiment of the application, step S3, the step of constructing a digital twin model of the distribution network, includes D1 to D4:

[0122] D1. Collect historical operating data, fault record data, and network topology data of the power distribution network;

[0123] Specifically, historical operation data collection includes operation data of distribution network equipment for the past 5 years, covering operating parameters such as transformer load rate, line current, voltage level, and power factor. Fault record data includes detailed information such as fault type, occurrence time, fault cause, repair time, and scope of impact. Network topology data includes static information such as equipment connection relationships, geographical coordinates, equipment technical parameters, and protection configuration.

[0124] D2. Establish a mathematical simulation model of distribution network equipment based on historical operating data, simulate the operating characteristics of the real distribution network, train the model, and output a digital twin model of the distribution network.

[0125] The mathematical simulation model is constructed based on power system theories such as power flow calculation, short-circuit analysis, and stability analysis. It uses the node admittance matrix method to establish network equations, combines the Newton-Raphson iterative algorithm to solve for power flow distribution, and introduces the random forest algorithm to model equipment failure probabilities. The core equations of the digital twin model are expressed as follows:

[0126]

[0127] In the formula, ΔP and ΔQ are the corrections for active and reactive power, respectively; J is the block submatrix of the Jacobian matrix; Δθ and ΔV are the corrections for the node voltage phase angle and amplitude; υ is the meteorological influence coefficient; F weather (T,W,R) represents the meteorological influence function of temperature T, wind speed W, and rainfall R; ο represents the historical fault influence coefficient; G fault (H data ) is a fault impact function based on historical operation data.

[0128] D3. Input the current meteorological monitoring data and equipment operating status into the digital twin model of the power distribution network;

[0129] Specifically, the real-time data input interface adopts a standardized data format. Current meteorological monitoring data includes parameters such as real-time temperature, wind speed, wind direction, rainfall, and humidity. Equipment operating status includes real-time monitoring values ​​such as load current, bus voltage, transformer oil temperature, and switch status. The data input frequency is updated every 10 seconds. Anomaly detection and missing value filling are performed through the data preprocessing module to ensure the quality and continuity of the input data.

[0130] D4. Simulate the operation response of the distribution network using a digital twin model of the distribution network and output the simulation results of the distribution network.

[0131] Specifically, the simulation calculation process of the digital twin model of the distribution network includes multiple modules such as power flow calculation, fault propagation analysis, and load forecasting. The simulation results output includes key indicators such as voltage distribution at each node, power flow distribution on the line, equipment load rate, system loss, fault impact range, and load transfer scheme. The simulation calculation adopts parallel computing technology.

[0132] It should be noted that the aforementioned digital twin model achieves a high-fidelity digital mapping of complex power distribution network systems through a combination of physical modeling and data-driven modeling. The meteorological influence function F in the model... weather (T,W,R) can quantify the specific impact of extreme weather on power grid operation, and the fault impact function G fault (H data Based on historical failure mode learning, the accuracy and interpretability of failure prediction are improved.

[0133] In an optional implementation, in step S3, the construction of the digital twin model of the distribution network can also employ deep learning technology, using graph neural networks to model the topology of the distribution network, processing time-series data through convolutional neural networks, and combining attention mechanisms to capture key influencing factors, thereby improving the model's ability to model complex nonlinear relationships.

[0134] In another optional implementation, in step S3, the construction of the digital twin model of the distribution network can also introduce an incremental learning mechanism to continuously update the model parameters based on the newly added operating data and fault cases. Through online learning algorithms, the model can adapt to changes in the operating mode of the distribution network, maintain the timeliness and predictive accuracy of the model, and realize the self-optimization and evolution of the model.

[0135] In this embodiment of the application, step S4, the steps for generating the distribution network risk assessment results and emergency response plan, include E1 to E5:

[0136] E1. Construct a risk knowledge graph for the power distribution network and establish the relationships between equipment types, meteorological factors, failure modes, and emergency strategies.

[0137] E2. Input the equipment failure risk assessment data and spatial risk distribution data into the distribution network risk knowledge graph to perform knowledge reasoning and failure mode matching.

[0138] E3. Combine the results of knowledge reasoning and fault mode matching with the simulation results of the digital twin model of the distribution network to determine the key equipment and fault propagation paths affected by extreme weather.

[0139] E4. Based on the fault mode matching results and simulation analysis, generate distribution network risk assessment results;

[0140] E5. Based on the risk assessment results of the distribution network, match the corresponding emergency resource allocation and handling steps to generate an emergency response plan.

[0141] Specifically, in step E1, a knowledge graph of distribution network risks is constructed using the Neo4j graph database. The graph nodes include equipment entities (transformers, switches, lines, etc.), meteorological entities (typhoons, rainstorms, ice storms, etc.), fault entities (short circuits, grounding, overloads, etc.), and strategy entities (load transfer, equipment switching, emergency power generation, etc.). The relationship edges include "impact relationships", "causal relationships", "temporal relationships", and "spatial relationships", etc., and are trained through expert knowledge and historical fault cases.

[0142] Specifically, in step E2, knowledge reasoning adopts a hybrid reasoning method that combines a rule-based reasoning engine with a graph neural network. Equipment failure risk assessment data and spatial risk distribution data are used as reasoning inputs. The graph traversal algorithm searches for matching failure mode paths. An example of the reasoning rule is "IF typhoon AND wind speed > 25 m / s AND overhead line THEN conductor galloping risk = high". Fault mode matching uses semantic similarity calculation, and the matching threshold is set to 0.8.

[0143] Specifically, in step E3, the calculation model for critical equipment identification and fault propagation path analysis is as follows:

[0144] R critical (i)=w1·P fault (i)+w2·I importance (i)+w3·∑ j∈N(i) P cascade (i,j);

[0145]

[0146] In the formula, R critical (i) represents the criticality assessment value of device i; P fault (i) represents the failure probability of device i; I importance (i) represents the importance index of device i; P cascade (i,j) represents the fault propagation probability from device i to device j; N(i) is the set of adjacent devices of device i; T topology (i,j) represents the topological connectivity strength; d ij λ is the distance between devices; W is the distance attenuation coefficient. weather is the extreme weather amplification factor; w1, w2, and w3 are weighting coefficients.

[0147] It should be noted that in the formula w1·P fault (i) Reflects the inherent risk of equipment failure, w2·I importance (i) Reflecting the importance of the equipment in the power grid, w3·∑ j∈N(i) P cascade (i,j) quantifies the impact of fault propagation and simulates the spatial characteristics of fault propagation through an exponential distance decay function.

[0148] Specifically, in step E4, the distribution network risk assessment results are generated, including dimensions such as risk level classification, impact range determination, and time prediction. The risk levels are divided into four levels: Level I (low risk), Level II (medium risk), Level III (high risk), and Level IV (extremely high risk). The assessment results include information such as risk type, probability of occurrence, list of affected equipment, expected duration, and possible load loss.

[0149] Specifically, in step E5, the emergency response plan is generated using a multi-objective optimization algorithm. The optimization objectives include minimizing power outage time, minimizing load loss, and minimizing resource consumption. The constraints include personnel quantity, equipment inventory, and transportation time. The plan content includes emergency team dispatch, backup equipment configuration, load transfer strategy, repair operation process, and material transportation route. The optimal plan is solved using a genetic algorithm.

[0150] It should be noted that the above-mentioned risk assessment and emergency response plan generation process combines the semantic reasoning ability of knowledge graphs with the simulation and prediction capabilities of digital twin models, realizing the transformation from data-driven to knowledge-driven intelligent decision-making. The graph structure of knowledge graphs can naturally express complex causal relationships, and multi-objective optimization algorithms ensure the scientific nature and operability of emergency plans.

[0151] In an optional implementation, in step S4, the risk assessment and emergency response plan generation can also incorporate reinforcement learning technology. By interacting with the environment, the optimal decision-making strategy can be learned, and the decision-making model can be continuously optimized based on feedback from historical emergency response effects, thereby improving the practicality and effectiveness of the emergency response plan.

[0152] In another optional implementation, in step S4, the risk assessment and emergency response plan generation can also be combined with blockchain technology to establish a multi-departmental collaborative emergency decision-making mechanism. The emergency response process can be automatically triggered through smart contracts to ensure the transparency and traceability of the emergency response process and improve the efficiency of cross-departmental collaboration.

[0153] In this embodiment of the application, step S5, which involves disseminating the distribution network risk assessment results and emergency response plans through multiple channels, includes steps F1 to F4:

[0154] F1. Based on the risk level in the distribution network risk assessment results, generate early warning information and release it through communication channels;

[0155] F2. Display the real-time operating status of the power distribution network and the execution progress of emergency response plans on the monitoring interface;

[0156] F3. Receive emergency response execution status and equipment status update information transmitted back by on-site maintenance personnel via mobile terminal;

[0157] F4. Evaluate the effectiveness of emergency response based on feedback data from the implementation.

[0158] Specifically, in step F1, the early warning information is automatically generated, including core elements such as risk type, risk level, affected area, expected duration, and recommended measures. Differentiated release strategies are adopted according to the risk level: Level I risks are alerted through the internal monitoring system, Level II risks are alerted via SMS, Level III risks are alerted via telephone voice alarm, and Level IV risks are alerted via emergency broadcast across all channels, including SMS platform, telephone system, mobile APP push, web system pop-up window, WeChat work group, email, etc.

[0159] Specifically, in step F2, the monitoring interface is built on a large-screen display system and adopts a layered and partitioned visual layout to display information such as the primary wiring diagram of the power distribution network, equipment operating status, meteorological information distribution, risk heat map, and emergency team location in real time. The execution progress of the emergency response plan is presented intuitively through progress bars, timelines, status indicators, etc., including key indicators such as task allocation status, personnel arrival status, equipment preparation progress, and work completion rate. The interface refreshes every 5 seconds and supports simultaneous access by multiple users and hierarchical permission management.

[0160] Specifically, in step F3, on-site maintenance personnel transmit execution status via a dedicated mobile terminal APP. The mobile terminal integrates functions such as GPS positioning, photo and video recording, voice recognition, and barcode scanning. The transmitted information includes detailed data such as personnel arrival time, work start time, on-site photos, equipment inspection results, fault handling progress, and completion confirmation. It supports offline data storage and automatic synchronization after network recovery. Data transmission is protected by both 4G / 5G networks and Beidou satellite communication.

[0161] Specifically, in step F4, the emergency response effectiveness evaluation adopts a multi-dimensional indicator system, including response time evaluation, resource allocation efficiency evaluation, response quality evaluation, and user satisfaction evaluation. The response time evaluation includes statistics on time nodes such as early warning release time, personnel arrival time, and fault repair time. The resource allocation efficiency evaluation analyzes indicators such as personnel utilization rate, equipment utilization rate, and material consumption ratio. The response quality evaluation is measured by indicators such as fault recovery rate, secondary failure rate, and safety accident incidence rate. The evaluation results automatically generate an evaluation report containing problem analysis, improvement suggestions, and experience summary, providing data support for subsequent emergency plan optimization.

[0162] It should be noted that the aforementioned multi-channel early warning release mechanism ensures the timely delivery and effective coverage of early warning information through differentiated information push strategies, the visual monitoring interface provides comprehensive situational awareness capabilities for emergency command, the on-site data feedback from mobile terminals forms an information closed loop, and the multi-dimensional effect evaluation system provides a scientific basis for continuous system improvement.

[0163] In an optional implementation, the multi-channel early warning release in step S5 can also be combined with artificial intelligence voice technology, which automatically generates voice broadcast content through natural language generation technology, supports multilingual broadcast and voice synthesis, improves the comprehensibility and dissemination effect of early warning information, and is particularly suitable for information transmission to remote areas and special groups.

[0164] In another optional implementation, the multi-channel early warning release in step S5 can also incorporate social media platform integration, releasing power outage information and restoration progress to the public through new media channels such as official Weibo and WeChat public accounts, while establishing a public opinion monitoring mechanism to respond promptly to public concerns and media reports, thereby enhancing the social image and credibility of power companies.

[0165] In summary, the five-layer architecture of the distribution network risk auxiliary decision-making system enhances the comprehensiveness and real-time performance of distribution network risk monitoring under extreme weather conditions through multi-source heterogeneous data fusion and edge-cloud collaborative processing. The protocol conversion gateway enables unified access for devices from different manufacturers and using different protocols, resolving the issues of data silos and poor compatibility in traditional systems. The real-time processing capabilities of the edge computing layer effectively reduce data transmission latency, while the deep analysis of the cloud platform layer provides robust computational support for risk assessment, forming a highly efficient collaborative architecture that combines real-time edge response with intelligent cloud analysis.

[0166] Driven by both a distribution network risk knowledge graph and a digital twin model, a fundamental shift from traditional experience-based decision-making to intelligent scientific decision-making has been achieved. The knowledge graph establishes systematic relationships between equipment, weather, and faults, enabling the system to automatically identify complex fault propagation patterns. The digital twin model provides a near-realistic simulation environment, offering a scientific basis for the formulation and verification of emergency plans. The resulting closed-loop mechanism of "monitoring-decision-response-feedback" not only improves the efficiency and accuracy of emergency response but also continuously enhances the system's intelligence level through continuous learning and optimization, providing reliable technical support for the safe and stable operation of the power grid under extreme weather conditions.

[0167] Example 3 illustrates a schematic scheme for a distribution network risk auxiliary decision-making method. It should be noted that the technical solution of this distribution network risk auxiliary decision-making system belongs to the same concept as the technical solution of the aforementioned distribution network risk auxiliary decision-making method. Details not described in detail in this embodiment can be found in the description of the technical solution of the aforementioned distribution network risk auxiliary decision-making method.

[0168] This embodiment also provides a power distribution network risk auxiliary decision-making system, including a data acquisition module, an edge computing module, a cloud platform analysis module, a digital twin module, and an early warning release module:

[0169] The data acquisition module is responsible for collecting power distribution network equipment operation data, meteorological monitoring data, and location data, and outputting multi-source power grid data.

[0170] The edge computing module performs unified protocol conversion on multi-source data from the power grid, and performs real-time analysis and anomaly detection on the converted multi-source data to generate edge layer early warning information;

[0171] The cloud platform analysis module performs in-depth data analysis on edge layer early warning information and key detection data to generate equipment failure risk assessment data and spatial risk distribution data;

[0172] The digital twin module links equipment failure risk assessment data and spatial risk distribution data with extreme weather failure patterns, and generates distribution network risk assessment results and emergency response plans through distribution network risk knowledge graph and digital twin model;

[0173] The early warning release module will release the risk assessment results and emergency response plans of the power distribution network through multiple channels.

[0174] This embodiment also provides an electronic device suitable for power distribution network risk auxiliary decision-making, including: a memory and a processor; the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions to implement the power distribution network risk auxiliary decision-making method proposed in the above embodiment.

[0175] This embodiment also provides a storage medium storing a computer program that, when executed by a processor, implements the distribution network risk auxiliary decision-making method proposed in the above embodiments.

[0176] The storage medium proposed in this embodiment belongs to the same inventive concept as the method for implementing distribution network risk auxiliary decision-making proposed in the above embodiments. Technical details not described in detail in this embodiment can be found in the above embodiments, and this embodiment has the same beneficial effects as the above embodiments.

[0177] Based on the above description of the implementation methods, those skilled in the art will clearly understand that the present invention can be implemented using software and necessary general-purpose hardware, and of course, it can also be implemented using hardware. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, can be embodied in the form of a software product. This computer software product can be stored in a computer-readable storage medium, such as a computer floppy disk, read-only memory (ROM), random access memory (RAM), flash memory, hard disk, or optical disk, etc., including several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute the methods of the various embodiments of the present invention.

[0178] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

Claims

1. A risk-assisted decision-making method for power distribution networks, characterized in that, Includes the following steps: Collect multi-source data from the power grid; The multi-source data of the power grid is subjected to unified protocol conversion, which converts various heterogeneous communication protocols into a unified data format. The converted multi-source data of the power grid is then analyzed and anomaly detected in real time through the edge computing layer to generate edge layer early warning information. The edge layer early warning information and key detection data are uploaded to the cloud platform layer for in-depth data analysis. Historical operation data is used to build a digital twin model of the power distribution network and generate equipment failure risk assessment data and spatial risk distribution data. The equipment failure risk assessment data and spatial risk distribution data are correlated with extreme weather failure patterns. Distribution network risk assessment results and emergency response plans are generated through distribution network risk knowledge graphs and digital twin models. The risk assessment results and emergency response plans for the power distribution network will be released to the public through multiple channels.

2. The distribution network risk auxiliary decision-making method as described in claim 1, characterized in that, The steps for collecting multi-source data from the power grid include: The controller collects current, voltage, and temperature rise parameters of the power distribution network equipment to obtain the operating data of the power distribution network equipment; Meteorological monitoring data are obtained by collecting wind speed, wind direction, rainfall, temperature, and air pressure through regional micro-meteorological stations; The AI-powered recognition terminal identifies abnormal device states and obtains abnormal device detection data. Mobile monitoring equipment is used to inspect and monitor power distribution network facilities and obtain inspection and testing data. Location data is obtained by collecting the real-time location of power distribution equipment and mobile workers through positioning terminals; The collected information is integrated and output as multi-source power grid data.

3. The distribution network risk auxiliary decision-making method as described in claim 2, characterized in that, The steps for real-time analysis and anomaly detection of multi-source power grid data include: Protocol parsing is performed on multi-source power grid data in a unified data format to extract data fields from the multi-source power grid data; Build a streaming computing engine to perform real-time processing and statistical analysis on the data fields through time windows; By using a preset safety threshold rule base, rule matching and risk assessment are performed on the power distribution network equipment operation data, meteorological monitoring data, and equipment anomaly detection data; When the detection parameters exceed the preset safety threshold, the edge layer warning information of the corresponding risk level is triggered, and the corresponding key detection data is marked.

4. The distribution network risk auxiliary decision-making method as described in claim 3, characterized in that, The steps involved in conducting in-depth data analysis include: The edge layer early warning information and key detection data are encrypted and then uploaded to the cloud platform layer; The key detection data are analyzed spatially using a geographic information system to generate a geographical distribution risk map of power distribution network equipment. The current key detection data and historical operation data are correlated and analyzed using a distributed computing platform to train the equipment fault prediction algorithm model; The fault probability value of each power distribution network device is output through the equipment fault prediction algorithm model to form the equipment fault risk assessment data. Regional spatial risk distribution data are generated by combining geographical risk maps.

5. The distribution network risk auxiliary decision-making method as described in claim 4, characterized in that, The steps for constructing the digital twin model of the power distribution network include: Collect historical operation data, fault record data, and network topology data of the power distribution network; A mathematical simulation model of the distribution network equipment is established based on the historical operating data to simulate the operating characteristics of the real distribution network. The model is then trained and a digital twin model of the distribution network is output. Input the current meteorological monitoring data and equipment operating status into the digital twin model of the power distribution network; The operation response of the distribution network is simulated using the digital twin model of the distribution network, and the simulation results of the distribution network are output.

6. The distribution network risk auxiliary decision-making method as described in claim 5, characterized in that, The steps for generating the distribution network risk assessment results and emergency response plan include: Construct a risk knowledge graph for the power distribution network and establish the correlation between equipment types, meteorological factors, failure modes, and emergency strategies; The equipment failure risk assessment data and spatial risk distribution data are input into the power distribution network risk knowledge graph for knowledge reasoning and failure mode matching. By combining the results of knowledge reasoning and fault mode matching with the simulation results of the digital twin model of the power distribution network, the key equipment and fault propagation paths affected by extreme weather are determined. Based on the fault mode matching results and simulation analysis, a distribution network risk assessment result is generated. Based on the risk assessment results of the power distribution network, corresponding emergency resource allocation and response steps are matched to generate an emergency response plan.

7. The distribution network risk auxiliary decision-making method as described in claim 6, characterized in that, The steps for disseminating the risk assessment results and emergency response plans for the power distribution network through multiple channels include: Based on the risk level in the power distribution network risk assessment results, early warning information is generated and disseminated through communication channels. The monitoring interface displays the real-time operating status of the power distribution network and the progress of emergency response plans. Receive emergency response execution status and equipment status update information transmitted back by on-site maintenance personnel via mobile terminals; The effectiveness of the emergency response is evaluated based on the feedback data from the implementation.

8. A distribution network risk auxiliary decision-making system, employing the method described in any one of claims 1-7, characterized in that, It includes a data acquisition module, an edge computing module, a cloud platform analysis module, a digital twin module, and an early warning release module: The data acquisition module is responsible for collecting power distribution network equipment operation data, meteorological monitoring data, and location data, and outputting multi-source power grid data. The edge computing module performs unified protocol conversion on the multi-source data of the power grid, and performs real-time analysis and anomaly detection on the converted multi-source data of the power grid to generate edge layer early warning information; The cloud platform analysis module performs in-depth data analysis on the edge layer early warning information and key detection data to generate equipment failure risk assessment data and spatial risk distribution data. The digital twin module associates equipment failure risk assessment data and spatial risk distribution data with extreme weather failure patterns, and generates distribution network risk assessment results and emergency response plans through the distribution network risk knowledge graph and digital twin model. The early warning release module will release the risk assessment results and emergency response plans of the power distribution network through multiple channels.

9. An electronic device, comprising: Memory and processor; The memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, they implement the steps of the power distribution network risk auxiliary decision-making method according to any one of claims 1 to 7.

10. A computer-readable storage medium storing computer-executable instructions that, when executed by a processor, implement the steps of the power distribution network risk auxiliary decision-making method according to any one of claims 1 to 7.

Images