A power grid operation risk prediction and collaborative governance system based on digital twinning

Abstract

The application discloses a power grid operation risk prediction and collaborative management system based on digital twinning, which comprises a basic support layer, a business management and control layer, an intelligent analysis layer and a collaborative management layer. The basic support layer constructs a power grid equipment operation state real-time mapping model through digital twinning technology; the business management and control layer realizes the cooperation of operation monitoring and maintenance management; the intelligent analysis layer analyzes data based on an improved Bayesian network, realizes power grid situation assessment and risk prediction; and the collaborative management layer executes hierarchical response according to the analysis result. The application solves the technical problems of information island, inaccurate risk prediction and untimely response in traditional power grid operation management, realizes the accurate perception of power grid operation state, the intelligent prediction of risk and the collaborative linkage of disposal, significantly improves the intelligentization and automation level of power grid operation management, and provides a powerful guarantee for the safe and stable operation of the power grid.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power system intelligent operation and maintenance, and particularly relates to a power grid operation risk prediction and collaborative governance system based on digital twinning. BACKGROUND

[0002] With the deepening of the construction of smart grid and the large-scale grid connection of new energy, the structure of power grid is becoming more and more complex, and the operation mode is more and more diverse, which puts forward higher requirements for the operation and control of power grid dispatching. Although modern power grid has a perfect automation and informatization foundation, there are still the following problems in intelligent analysis and collaborative governance:

[0003] 1. The value of data is not fully excavated. Although the existing SCADA system has high-frequency sampling capability and extensive coverage, it still needs to be strengthened in multi-source data fusion analysis, equipment operation mechanism modeling, etc. In particular, there is a lack of deep mining and knowledge extraction of massive operation data, which is difficult to support more accurate state evaluation and decision analysis.

[0004] 2. The depth of business collaboration is not enough. Although the power grid dispatching technical support system has realized data integration, the depth of business collaboration is still insufficient. The linkage mechanism of operation monitoring, maintenance management, fault disposal and other links needs to be improved, and it is difficult to fully realize the value brought by data sharing.

[0005] 3. The risk prediction capability is limited. The existing early warning system mainly focuses on the abnormal detection of single equipment or local system, and lacks consideration of the correlation between equipment and fault propagation characteristics. There is a lack of intelligent analysis model based on network topology and historical experience, which makes it difficult to realize accurate prediction and active prevention of risks.

[0006] 4. The emergency response mechanism needs to be optimized. Although a relatively perfect emergency plan system has been established, there is still room for improvement in rapid judgment and intelligent decision support.

[0007] 5. The level of global collaboration is not high. Although the existing control system covers all business links, it lacks a systematic collaborative governance mechanism. The linkage effect between links is not fully realized, and it is difficult to realize active management of the whole life cycle of the power grid.

[0008] Although various intelligent power grid operation and maintenance management systems have been developed in the industry, there are still the following technical challenges in deep application:

[0009] 1. The existing digital twinning application lacks unified modeling of equipment topology relationship, maintenance operation and dispatching operation, which affects the deep mining of data value;

[0010] 2. The existing risk prediction model lacks consideration of the correlation between equipment and the fault propagation mechanism, especially in the use of Bayesian networks for multi-dimensional analysis.

[0011] 3. The operation monitoring and maintenance management coordination mechanism is imperfect, and it is difficult to realize seamless connection and closed-loop management of risk early warning and emergency response.

[0012] Therefore, it is necessary to develop a power grid operation management and control system integrating digital twin modeling, Bayesian network analysis and collaborative governance mechanism to improve the intelligent level and safety guarantee capability of power grid operation. SUMMARY

[0013] The problem to be solved by the present application is how to provide a power grid operation management and control system based on digital twin technology, with intelligent risk prediction capability and supporting global collaborative governance. In order to solve the above technical problems, the technical solutions adopted by the present application are as follows:

[0014] A power grid operation risk prediction and collaborative governance system based on digital twin, characterized in that it comprises a basic support layer, a business management layer, an intelligent analysis layer and a collaborative governance layer;

[0015] The basic support layer comprises a data acquisition submodule, a data processing submodule and a data modeling submodule, acquires and stores power grid operation data, and constructs a power grid digital twin model including device topology relationship, maintenance operation procedure and dispatching operation process after processing the acquired data through data cleaning and standardization. The digital twin model reflects the operation state of the power grid physical device in real time.

[0016] The business management layer comprises an operation monitoring submodule and a maintenance management submodule, and constructs a business monitoring system based on the digital twin model, realizes operation monitoring and maintenance management, and transmits structured monitoring data to the intelligent analysis layer.

[0017] The intelligent analysis layer comprises a situation awareness submodule and a risk analysis submodule, and analyzes the data based on the improved Bayesian network, outputs the power grid situation assessment and risk prediction results, and transmits them to the collaborative governance layer.

[0018] The collaborative governance layer comprises an emergency response submodule, executes emergency response according to the analysis results, and issues disposal instructions through a mobile application terminal.

[0019] Further, the basic support layer comprises a data acquisition submodule for acquiring real-time operation data of the power grid, specifically including: device operation attributes including voltage level, operating current, active power, reactive power, device opening and closing state, device operation mode; transmission parameters including data sampling period, data quality identifier, data timestamp, data source identifier; device basic information including device basic attributes and connection relationship; device maintenance information; spatial location information.

[0020] The data processing submodule processes the collected data to remove noise, remove duplicates, and standardize the data, and establishes a unified data format standard.

[0021] The data modeling submodule constructs a digital twin model of the power grid based on the processed data, wherein the digital twin model includes a device topology relationship model, a maintenance operation procedure model, and a dispatch operation process model.

[0022] Further, the method for constructing the digital twin model of the power grid by the data modeling submodule includes:

[0023] The device topology relationship modeling abstracts the power devices in the power grid as nodes and the electrical connection relationship between the nodes as edges to construct a directed graph model of the power grid structure; the nodes of the directed graph model include basic attributes and operating attributes of the devices; and the edges of the directed graph model include connection attributes and transmission attributes.

[0024] The maintenance operation procedure modeling decomposes the maintenance operation process into three levels of operation items, operation steps, and operation points; each operation item includes specific operation guidance, safety measures, and quality standards; each operation step sets execution timing and completion standards; and each operation point records key control points and acceptance criteria.

[0025] The dispatch operation process modeling divides the dispatch operation process into three stages of pre-arrangement, execution, and supervision, establishes operation step sequences for each stage, defines trigger conditions, execution actions, and completion flags for each step, and automatically records execution time, operation personnel, and operation results for each step.

[0026] The digital twin model is synchronized with the actual power grid state by the following methods: determining an optimal collection period based on device importance and operating state and dynamically adjusting the collection period; using data caching and version control mechanisms to ensure data consistency; setting up model update exception handling and rollback mechanisms; updating power grid operating data according to a preset collection period; updating device state data according to a preset collection period; updating operating state in response to dispatch operation instructions; and updating maintenance state in response to maintenance operation information.

[0027] Further, the specific implementation method of the operation monitoring submodule in the business control layer includes:

[0028] Based on the device topology relationship model, monitoring objects are established, device nodes are mapped to monitoring measurement points, topology relationships are mapped to monitoring associations, and parameter threshold ranges for each monitoring measurement point are established.

[0029] Based on the dispatch operation process model, an operation condition library is established, parameter variation characteristics of standard operation steps are extracted, a correspondence between operation types and parameter variation patterns is established, and a standard condition characteristic library is formed.

[0030] Performing operation monitoring, collecting monitoring point data in real time, identifying parameter out-of-limit conditions, extracting associated point data, matching operating condition characteristics, and generating structured abnormality descriptions;

[0031] Further, the specific implementation method of the maintenance management submodule in the business management layer includes:

[0032] Based on the maintenance operation procedure model, establish operation standards, convert operation procedures into standard operation items, define operation item execution acceptance rules, and establish the association between operation items and equipment states.

[0033] Performing maintenance management, tracking maintenance operation execution, recording operation process data, associating equipment operating states, and generating standardized maintenance records.

[0034] Data correlation analysis, establishing time sequence correlation between maintenance operations and equipment states, analyzing parameter changes before and after maintenance, and extracting maintenance effect characteristics.

[0035] Further, the intelligent analysis layer specifically includes: constructing observation variables and state variables of the improved Bayesian network, taking the equipment abnormality descriptions and operating condition characteristics output by the operation monitoring submodule as network observation variables X, taking the operation records and equipment states output by the maintenance management submodule as network state variables Y, and determining the connection relationship E between network nodes based on the equipment topology relationship matrix T in the digital twin model.

[0036] Further, the specific implementation of the improved Bayesian network includes:

[0037] Based on the network observation variables, state variables, and connection relationships, a dynamic Bayesian network structure B(V, E) containing T time segments is constructed, where V = {X U Y} represents the network node set, each node contains state variables at times t and t-1, and E represents the time sequence transition relationship between nodes.

[0038] Based on the equipment associated fault records, a node conditional probability table (CPT) is constructed, and the EM algorithm is used to iteratively optimize the CPT parameters. The EM algorithm calculates the conditional probability of unobserved states through the E step, maximizes the expected likelihood estimation probability parameters of complete data through the M step, and iteratively optimizes the optimal parameters. The optimization process of the EM algorithm is:

[0039] E step, estimate the conditional probability distribution of unobserved data based on the current parameter θ:

[0040] P(Y|D,CPT)

[0041] Where D is the observed data, i.e., the equipment abnormality descriptions and operating condition characteristics output by the operation monitoring submodule, and Y is the state variable, i.e., the operation records and equipment states output by the maintenance management submodule.

[0042] M step, maximize the expected likelihood function of the complete data to get new parameter estimates:

[0043] CPT * = argmax P(D|CPT)

[0044] Where CPT * is the optimal conditional probability estimate, P(D|CPT) is the likelihood probability of observing data D under the current conditional probability;

[0045] The probability transition matrix is corrected by the device importance weight, and the formula is:

[0046] P ′ (Y|X) = W x P(Y|X)

[0047] Where P ′ (Y|X) is the corrected conditional probability, P(Y|X) is the original conditional probability, X is the observation variable, i.e. the device abnormality description and working condition characteristics output by the operation monitoring submodule, and W is the device importance weight matrix, which is determined based on the voltage level in the device basic attribute, the connection relationship in the topology relationship matrix, and the operation attribute; For the variable X, the probability distribution of the network state variable is calculated using the MAP criterion, and the formula is:

[0048] P(Y|X) = argmax P(X|Y) P(Y)

[0049] Where P(X|Y) is the likelihood probability (dependence of observation variable on state variable), and P(Y) is the prior probability of state variable;

[0050] The forward algorithm is used to calculate the risk probability at time t+k, and the formula is:

[0051] P(Y t+k |X 1:t ) = ∑ P(Y t+k |Y t ) P(Y t |X 1:t )

[0052] Where Y t+k is the state at time t+k, X 1:t is the observation sequence from time 1 to t, P(Y t+k |Y t ) is the state transition probability, and P(Y t |X 1:t ) is the state posterior probability at time t;

[0053] The risk propagation path is calculated based on the conditional risk propagation probability matrix, and the formula is:

[0054] P = P(Y|X) x R

[0055] Wherein, R is the risk correlation matrix, the correlation between the devices is described based on the device topology relationship matrix T;

[0056] Further, the specific implementation of the collaborative governance layer includes: based on the risk probability distribution and the propagation path output by the intelligent analysis layer, establishing a hierarchical response mechanism, mapping the risk probability P(Y|X) to four response levels, determining the key devices within the influence range according to the risk propagation probability matrix P, and pushing the risk level and key device information to the mobile application terminal of the relevant personnel;

[0057] Further, the system also includes a mobile application terminal:

[0058] The mobile application terminal interacts with the system in real time through the 4G / 5G network, receives the risk level and key device information output by the intelligent analysis layer, displays the disposal instructions issued by the collaborative governance layer, and real-time feedbacks the on-site disposal situation to the system, including the device state change and disposal progress, and cooperates with the collaborative governance layer to form a closed-loop management.

[0059] The beneficial effects of the present application are as follows:

[0060] 1. By constructing a digital twin model integrating device topology relationship, maintenance operation procedures and dispatching operation processes, unified modeling and real-time mapping of power grid operation state are realized, providing a comprehensive data basis for risk prediction;

[0061] 2. Based on the improved Bayesian network and the device importance weight correction mechanism, combined with the conditional risk propagation probability matrix, the accuracy and reliability of risk prediction are improved, realizing the breakthrough from single device fault early warning to global risk prediction;

[0062] 3. A hierarchical response mechanism based on risk level and influence range is adopted, and real-time feedback of on-site disposal situation is realized through the mobile terminal, forming a complete closed-loop management from risk early warning to on-site disposal; BRIEF DESCRIPTION OF DRAWINGS

[0063] Figure 1 It is a system architecture diagram for power grid operation risk prediction and collaborative governance based on digital twin.

[0064] Figure 2 It is a schematic diagram of power grid digital twin model. DETAILED DESCRIPTION

[0065] The application provides a power grid operation risk prediction and collaborative management system based on digital twinning, which comprises a basic support layer, a business management and control layer, an intelligent analysis layer and a collaborative management layer, and forms a closed-loop management mechanism of data circulation and business collaboration between the layers. The technical solutions of the application will be described in detail below in combination with specific embodiments.

[0066] Embodiment 1: Power grid operation risk prediction and collaborative management system based on digital twinning

[0067] With reference to Figure 1 The system of the embodiment adopts a hierarchical architecture design, and realizes the prediction and management of power grid operation risks through the collaborative cooperation of the basic support layer, the business management and control layer, the intelligent analysis layer and the collaborative management layer. The specific implementation process of the system is as follows:

[0068] The basic support layer collects real-time operation data of the power grid through a data acquisition submodule. The submodule adopts a distributed data acquisition architecture, and data acquisition units are deployed at each substation and key node. The data acquisition units adopt redundant design, each unit contains two sets of acquisition systems, namely main and standby systems, and when the main system fails, it automatically switches to the standby system to ensure the reliability of data acquisition. The collected data includes the following categories:

[0069] Device operation attribute data, specifically including: voltage level (35kV, 110kV, 220kV, 500kV, etc.), operating current (accurate to 0.1A), active power (accurate to 0.1kW), reactive power (accurate to 0.1kVar), device opening and closing state (open, closed, fault), device operation mode (normal, maintenance, standby) and other parameters. The sampling period of these parameters is differentiated according to the importance of the device, and the sampling period of important devices such as main transformers is 100ms, and the sampling period of general devices is 1s.

[0070] Transmission parameter data, including: data sampling period (configurable, range 100ms-60s), data quality identifier (normal, suspicious, error), data timestamp (accurate to milliseconds), data source identifier (device code, acquisition unit number) and other information. The system monitors the data quality in real time, and when it finds that the data quality identifier is "suspicious" or "error", it automatically triggers the data review mechanism.

[0071] Device basic information data, including: device model, manufacturer, commissioning time, rated parameter, maintenance period and other basic attributes, as well as the primary system connection relationship and secondary system connection relationship between devices. These data are synchronously obtained from the device asset management system, and full-quantity synchronization is performed every 24 hours, and incremental synchronization is used to update the changed data during the period.

[0072] The equipment maintenance information data includes historical maintenance records, fault records, technical transformation records, etc. The system establishes an equipment maintenance archive library to record detailed information of each maintenance activity, including maintenance time, maintenance content, maintenance personnel, maintenance effect evaluation, etc.

[0073] The spatial position information data is obtained by using the Beidou positioning system to obtain the accurate geographic coordinates of the equipment. The position accuracy of outdoor equipment is better than 10 meters, and the position coordinates of indoor equipment are determined by pre-calibration.

[0074] The collected data is processed by a data processing submodule. The submodule adopts a three-level data processing mechanism: the first level is real-time processing, which filters noise of the collected data, uses a median filter algorithm to eliminate sudden interference, and uses a mean filter algorithm to smooth the data curve for analog quantities; the second level is quasi-real-time processing, including data deduplication, data completion and outlier processing, the system establishes a data change characteristic model through historical data analysis for identification and processing of outliers; the third level is offline processing, including data cleaning, data standardization and data quality evaluation, which processes historical data in depth through batch processing.

[0075] The standardized processing adopts a unified data format standard, including: the time stamp is unified as UTC time, and the time zone information is recorded; the analog quantity data is converted into the International System of Units; the state quantity data uses standard state coding; the equipment identification uses a unified coding rule. The processed data is stored in a distributed time series database according to time sequence, facilitating subsequent rapid retrieval and analysis.

[0076] Referring to Figure 2 , the data modeling submodule constructs a power grid digital twin model based on the processed standard data. The model adopts a hierarchical modeling method, including three levels of device layer, system layer and business layer. In the device layer, the digital model of the independent device is constructed, including the geometric model, physical model and behavior model of the device; in the system layer, the correlation model between devices is constructed to describe the physical connection, logical relationship and information interaction between devices; in the business layer, the business process model is constructed to describe the rules and processes of maintenance, operation, dispatching and other business activities.

[0077] In the process of modeling the device topology relationship, a graph database is used to store the connection relationship between devices. The power equipment in the power grid is abstracted as a node, each node contains basic attributes such as device ID, device type, voltage level, and running attributes such as operating current and active power. The electrical connection relationship between nodes is abstracted as an edge, and the attributes of the edge include connection type (bus connection, T connection, etc.), conductor type, line length, and transmission attributes such as power flow direction and impedance parameters. The system supports dynamic updating of the topology relationship, and automatically updates the topology model when the device state changes (such as switch opening and closing, line outage, etc.).

[0078] In the modeling process of maintenance operation procedures, a hierarchical model structure is adopted. At the operation item level, each operation item contains detailed operation guidance (text description, operation video, precautions, etc.), safety measures (personal protection requirements, on-site safety measures, etc.), and quality standards (key parameter requirements, acceptance criteria, etc.). At the operation step level, the execution timing between steps (serial, parallel, conditional execution, etc.) is defined, and the completion criteria for each step (parameter requirements, state confirmation, etc.) are set. At the operation point level, the specific requirements (operation methods, precautions, etc.) and acceptance criteria (qualification standards, acceptance methods, etc.) for each key control point are recorded.

[0079] In the modeling process of dispatching operation procedures, a state machine model is used to describe the operation procedures. In the pre-arrangement phase, the system automatically generates operation tickets containing detailed operation steps and safety measures; in the execution phase, the system tracks the operation execution in real time, records the execution time, operation personnel, and operation results for each step; in the supervision phase, the system monitors the operation process in real time and alarms in case of abnormal conditions. Each operation step defines clear trigger conditions (time conditions, state conditions, etc.), execution actions (opening and closing operations, parameter adjustments, etc.), and completion criteria (state confirmation, parameter verification, etc.).

[0080] To ensure that the digital twin model is synchronized with the actual power grid state, the system adopts a multi-level synchronization mechanism:

[0081] 1. Collection cycle optimization: The system dynamically adjusts the collection cycle based on the importance of the device (evaluated by factors such as device type, voltage level, and access capacity) and the running state (normal, abnormal, alarm, etc.). The collection cycle for important devices in normal state is 1s, and it is automatically shortened to 100ms when an abnormality occurs; the collection cycle for general devices is 5s, and it is shortened to 1s when an abnormality occurs.

[0082] 2. Data consistency guarantee: A double-layer cache mechanism is adopted, the first layer is memory cache, used to store real-time data in the last 5 minutes; the second layer is persistent cache, used to store historical data in the last 24 hours. The system uses a version control mechanism to assign a unique version number to each data update, ensuring data consistency.

[0083] 3. Abnormal handling mechanism: Set up a model update abnormal handling process, including data anomaly detection (based on statistical characteristics and machine learning algorithms), abnormality grading (divided into warning, serious, and fatal three levels), abnormal handling (automatic handling and manual intervention combined), and abnormal record (detailed record of abnormal situation and handling process). When a fatal level abnormality occurs, the model rollback mechanism is automatically triggered to restore to the latest stable version.

[0084] The business management layer is based on the digital twin model to carry out operation monitoring and maintenance management. The operation monitoring submodule first establishes a monitoring object model, maps the device nodes to monitoring points, and each monitoring point includes a monitoring point ID, a monitoring point type, a collection period, an alarm threshold, and the like. The system supports multiple types of threshold settings: fixed threshold, dynamic threshold (automatically adjusted according to historical data statistics characteristics), and associated threshold (considering the state of related monitoring points).

[0085] When establishing the operating condition library, the system analyzes based on the above scheduling operation process model. Through the analysis of the operation data in the three stages of pre-arrangement, execution, and supervision, the system extracts the parameter change characteristics of the standard operation steps. For example, for transformer voltage regulation operation, the key parameter characteristics including voltage change rate, load current change, voltage value before and after voltage regulation are extracted; for line switching operation, the characteristic parameters such as switch action timing, bus voltage change, and power flow transfer are extracted. The system establishes a corresponding relationship between these characteristics and the corresponding operation type to form a standard condition characteristic library. Based on the established standard condition characteristic library, the system carries out real-time monitoring. First, real-time monitoring point data is collected, including the above operation-related voltage, current, power, and the like. When it is found that the parameters exceed the preset threshold range, the system identifies the parameter out-of-limit condition. Subsequently, the system automatically extracts the related monitoring point data that has a topological correlation with the out-of-limit parameter. By matching the real-time collected parameter change characteristics with the standard condition characteristic library, it is determined whether there is an abnormal condition at present. Finally, the system generates a structured abnormality description based on the matching result.

[0086] The maintenance management submodule establishes a standard job library based on the maintenance job procedure model. Each standard job item includes detailed execution specifications, such as job content, job tools, job environment requirements, and the like. The system establishes a mapping relationship between the job item and the device state, and automatically evaluates the necessity and priority of maintenance through state monitoring data. During the maintenance process, the system records the job execution in real time, including job time, job personnel, job content, detection data, and the like, and evaluates the job quality through image recognition and sensor data analysis.

[0087] Through data correlation analysis, the system establishes a time sequence correlation model between the maintenance job and the device state. The model includes three dimensions: time dimension (before maintenance, during maintenance, and after maintenance), state dimension (operation parameters, alarm information, and fault records), and job dimension (job content, job quality, and job effect). The system analyzes the maintenance effect through this model, evaluates the maintenance quality, and continuously optimizes the maintenance strategy.

[0088] The intelligent analysis layer performs analysis based on an improved Bayesian network. In constructing the network structure, the system first determines the set of observation variables X and the set of state variables Y. The observation variables include: device operating parameters (voltage, current, etc.), device state indicators (temperature, vibration, etc.), alarm information (type, level, etc.), operation records (type, time, etc.), and the like. The state variables include: device health status, fault type, risk level, maintenance recommendations, and the like.

[0089] The implementation process of the improved Bayesian network adopts a distributed computing framework to improve the computing efficiency of large-scale networks. The system first constructs a dynamic Bayesian network structure B(V, E) containing T time segments (generally T=24, representing 24 hours). In constructing the node conditional probability table (CPT), the system first constructs an initial node conditional probability table based on the device-associated fault records. Specifically, the system analyzes the fault propagation chain in the device-associated fault records to determine the conditional probability relationship between nodes. For example, when A device fails, the probability of failure of the adjacent B device is analyzed according to the associated fault records, thereby constructing the conditional probability table item from node A to node B. For each node pair in the network, the system constructs the corresponding conditional probability table item by analyzing its historical associated fault records. After completing the initial construction of the CPT, the system performs parameter iterative optimization using the EM algorithm, and the specific implementation of the EM algorithm includes:

[0090] 1. E step: Calculate the conditional probability P(Y|D, CPT) of unobserved state Y. The system uses the variational inference method to approximate the posterior distribution to improve the computing efficiency.

[0091] 2. M step: Maximize the expected likelihood function of the complete data. The system uses the gradient ascent method to optimize the parameters, and the learning rate uses an adaptive adjustment strategy.

[0092] 3. Convergence judgment: Stop iteration when the parameter change is less than a preset threshold (such as 0.001) or reaches the maximum number of iterations (such as 100 times).

[0093] The system corrects the probability transition matrix through the device importance weight. The calculation of the weight matrix W considers the following factors:

[0094] 1. Voltage level weight: 500kV is 1.0, 220kV is 0.8, 110kV is 0.6, and 35kV is 0.4.

[0095] 2. Topology importance weight: calculated according to the degree centrality of the device in the network.

[0096] 3. Operating state weight: calculated according to the operating time, load rate, and other factors of the device.

[0097] In risk prediction, the system uses a forward algorithm to calculate the risk probability at time t+k. The specific calculation formula is:

[0098] P(Y t+k |X1: t )=∑P(Y t+k |Y t )P(Y t |X1: t )

[0099] Where Y t+k is the state at time t+k, X 1:t is the observation sequence from 1 to t, P(Y t+k |Y t ) is the state transition probability, and P(Y t |X 1:t ) is the state posterior probability at time t. Through this calculation process, the system can predict the risk probability at future time based on the current observed device state sequence.

[0100] For the calculation of risk propagation path, the system analyzes based on the conditional risk propagation probability matrix, and the calculation formula is:

[0101] P=P(Y|X)×R

[0102] Where R is the risk correlation matrix, which describes the correlation between devices based on the device topology relationship matrix T. Through this calculation process, the system can determine the propagation path of risk between devices.

[0103] The collaborative governance layer establishes a four-level response mechanism based on the risk analysis result P(Y|X):

[0104] 1. First-level response (red): risk probability > 0.8, large impact range, need to be disposed immediately

[0105] 2. Second-level response (orange): 0.6 < risk probability ≤ 0.8, medium impact range, need to be disposed in time

[0106] 3. Third-level response (yellow): 0.4 < risk probability ≤ 0.6, impact range

[0107] 4. Fourth-level response: risk probability ≤ 0.4

[0108] The system determines the key devices within the impact range according to the risk propagation probability matrix P. For each response level, the system generates corresponding disposal instructions, including risk description, key device list, and disposal suggestions, etc.

[0109] The mobile application terminal of the embodiment interacts with the system in real time through a 4G / 5G network. In a preferred embodiment, the basic functions of the terminal include:

[0110] 1. Receiving risk level information output by the intelligent analysis layer

[0111] 2. Displaying the disposal instructions issued by the collaborative governance layer

[0112] 3. Returning the on-site disposal situation, including: equipment state change information, disposal progress.

[0113] The mobile application terminal will return the collected on-site information to the system in real time, forming a closed-loop management with the collaborative governance layer. The system records the information of the entire disposal process, realizing the whole-process tracking and management of the risk disposal process.

[0114] Through the application of multi-level architecture design and digital twin technology, the embodiment realizes the accurate prediction and collaborative governance of power grid operation risk. The system uses improved Bayesian network for risk analysis, and realizes closed-loop management of on-site disposal through the mobile application terminal. Through actual application verification, the system can effectively improve the power grid operation risk control ability, and has significant practical value and promotion significance.

Claims

1. A power grid operation risk prediction and collaborative governance system based on digital twinning, characterized in that, The system comprises a basic support layer, a business management and control layer, an intelligent analysis layer, and a collaborative governance layer. The basic support layer comprises a data acquisition sub-module, a data processing sub-module, and a data modeling sub-module, collects and stores power grid operation data, and constructs a power grid digital twin model comprising device topology relationship, maintenance operation procedure, and dispatching operation process after the collected data is processed through data cleaning and standardization. The business management and control layer comprises an operation monitoring sub-module and a maintenance management sub-module, constructs a business monitoring system based on the digital twin model, realizes operation monitoring and maintenance management, and transmits structured monitoring data to the intelligent analysis layer. The intelligent analysis layer comprises a situation awareness sub-module and a risk analysis sub-module, analyzes data based on an improved Bayesian network, outputs power grid situation assessment and risk prediction results, and transmits them to the collaborative governance layer. The collaborative governance layer comprises an emergency response sub-module, executes emergency response according to the analysis results, and issues disposal instructions through a mobile application terminal. The specific implementation of the improved Bayesian network comprises: Based on network observation variables, state variables and connection relationships, a dynamic Bayesian network structure containing T time segments is constructed wherein denotes a set of network nodes, each node containing state variables at t and t-1 time, and E denotes the time sequence transition relationship between nodes A node conditional probability table (CPT) is constructed based on device associated fault records, and an EM algorithm is used to iteratively optimize CPT parameters, wherein the EM algorithm calculates the conditional probability of unobserved states through an E step, maximizes the expected likelihood estimation probability parameters of complete data through an M step, and iteratively optimizes the optimal parameters, wherein the optimization process of the EM algorithm is as follows: In the E step, the conditional probability distribution of unobserved data is estimated according to the current parameter θ: Wherein D is the observed data, i.e., the device abnormality description and working condition characteristics output by the operation monitoring sub-module, and Y is the state variable, i.e., the operation record and device state output by the maintenance management sub-module. In the M step, the new parameter estimation is obtained by maximizing the expected likelihood function of complete data: wherein, is the optimal conditional probability estimate, P(D|CPT) is the likelihood of observing the data D given the current conditional probability. The probability transition matrix is corrected by the device importance weight, and the formula is as follows: wherein, is the revised conditional probability, is the original conditional probability, X is the observation variable, i.e., the equipment abnormality description and operating condition characteristics output by the operating monitoring sub-module, is an equipment importance weight matrix, determined based on the voltage level in the equipment basic attribute, the connection relationship in the equipment topology relationship matrix T, and the operating attribute; for the variable X, the probability distribution of the network state variable is calculated using the MAP criterion, and the formula is: Wherein P(X|Y) is the likelihood probability, and P(Y) is the prior probability of the state variable. The forward algorithm is used for the calculation The risk probability at the moment is calculated by the formula: wherein, is the state at time t + k, is the observation sequence from time 1 to t, is the state transition probability, is the state posterior probability at time t; The risk propagation path is calculated based on the conditional risk propagation probability matrix, and the formula is as follows: Wherein R is the risk correlation matrix, and the correlation between devices is described based on the device topology relationship matrix T.

2. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The basic support layer comprises: The data acquisition sub-module is used to collect real-time operation data of the power grid, and specifically comprises: device operation attributes, including voltage level, operating current, active power, reactive power, device opening and closing state, and device operation mode; transmission parameters, including data sampling period, data quality identifier, data timestamp, and data source identifier; device basic information, including device basic attributes and connection relationship; device maintenance information; and spatial location information. The data processing sub-module is used to denoise, remove duplicates, and standardize the collected data, and establish a unified data format standard. The data modeling sub-module is used to construct a power grid digital twin model based on the processed data, and the digital twin model comprises a device topology relationship model, a maintenance operation procedure model, and a dispatching operation process model.

3. The data modeling sub-module of claim 2, wherein, The method for constructing the power grid digital twin model by the data modeling sub-module comprises: A device topology relationship model is built, in which power devices in the power grid are abstracted as nodes, and electrical connection relationships between the nodes are abstracted as edges, to construct a power grid structure directed graph model; the nodes of the directed graph model include device basic attributes and operating attributes; the edges of the directed graph model include connection attributes and transmission attributes; A maintenance operation procedure model is built, in which a maintenance operation flow is divided into three levels of operation items, operation steps and operation points, each operation item includes specific operation guidance, safety measures and quality standards, each operation step sets an execution time sequence and a completion standard, and each operation point records a key control point and an acceptance criterion; A dispatch operation flow model is built, in which a dispatch operation process is divided into three stages of pre-arrangement, execution and supervision, an operation step sequence is established for each stage, and a trigger condition, an execution action and a completion flag of each step are defined; the system automatically records an execution time, an operator and an operation result of each step; The digital twin model is synchronized with an actual power grid state in the following ways: an optimal collection period is determined based on device importance and operating state and is dynamically adjusted; a data cache and a version control mechanism are used to ensure data consistency; a model update exception handling and rollback mechanism is set; power grid operating data are updated according to a preset collection period; device state data are updated according to a preset collection period; an operating state is updated in response to a dispatch operation instruction; and a maintenance state is updated in response to maintenance operation information.

4. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The specific implementation method of the operation monitoring submodule in the business management and control layer includes: A monitoring object is established based on the device topology relationship model, device nodes are mapped to monitoring measurement points, and topology relationships are mapped to monitoring associations, to establish a parameter threshold range of each monitoring measurement point; An operation condition library is established based on the dispatch operation flow model, parameter variation characteristics of a standard operation step are extracted, a corresponding relationship between an operation type and a parameter variation mode is established, and a standard condition characteristic library is formed; Operation monitoring is performed, monitoring measurement point data are collected in real time, parameter out-of-limit conditions are identified, associated measurement point data are extracted, operation condition characteristics are matched, and a structured abnormality description is generated.

5. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The specific implementation method of the maintenance management submodule in the business management and control layer includes: A job standard is established based on the maintenance operation procedure model, a job procedure is converted into a standard operation item, an execution and acceptance rule of the operation item is defined, and an association between the operation item and a device state is established; Maintenance management is performed, a maintenance operation execution is tracked, job process data are recorded, a device operating state is associated, and a standardized maintenance record is generated; Data association analysis is performed, a time sequence association between a maintenance operation and a device state is established, parameter variations before and after maintenance are analyzed, and a maintenance effect characteristic is extracted.

6. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The intelligent analysis layer specifically includes: An improved Bayesian network is constructed, observation variables and state variables are determined, device abnormality descriptions and condition characteristics output by the operation monitoring submodule are taken as network observation variables X, job records and device states output by the maintenance management submodule are taken as network state variables Y, and a connection relationship E between network nodes is determined based on a device topology relationship matrix T in the digital twin model.

7. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The specific implementation of the emergency response submodule of the collaborative governance layer includes: Based on the risk probability distribution and propagation path output by the intelligent analysis layer, a hierarchical response mechanism is established to distribute the risk probability... Mapped to a Level 4 response level, key equipment within the affected area is determined based on the risk propagation probability matrix P, and the risk level and key equipment information are pushed to the mobile application terminals of relevant personnel.

8. The power grid operation risk prediction and collaborative governance system based on digital twinning of claim 1, wherein, The system further includes a mobile application terminal: The mobile application terminal performs real-time data interaction with the system through a 4G / 5G network, receives the risk level and key equipment information output by the intelligent analysis layer, displays the disposal instructions issued by the collaborative management layer, and returns the on-site disposal situation to the system in real time, including the equipment state change and disposal progress, and forms a closed-loop management with the collaborative management layer.

Images