Power grid intelligent monitoring and dynamic scheduling system and method based on digital twinning

Abstract

The invention discloses a power grid intelligent monitoring and dynamic scheduling system and method based on digital twinning, and belongs to the technical field of power grid scheduling. The system comprises a power grid topology generation module for generating a power grid topology and a full-life-cycle file of power equipment; the modeling module and the knowledge graph establishing module are used for acquiring power grid operation situation data, and establishing a digital twin model and a knowledge graph according to the power grid operation situation data, power grid topology and power equipment full life cycle archives; the mapping module is used for establishing a mapping model to obtain a mapping data set; the risk analysis module performs multi-risk analysis through the mapping data set to generate a comprehensive risk analysis report; and the dispatching module is used for obtaining a power grid dispatching strategy by adopting a multi-target adaptive optimization algorithm and a load balance algorithm according to the comprehensive risk analysis report. The method effectively improves the anti-risk capability of the scheduling strategy, and improves the accuracy of the scheduling strategy through the comprehensive application of the digital twin model and the knowledge graph.

Description

Technical Field

[0001] The present invention relates to the technical field of power grid dispatching, and in particular to a power grid intelligent monitoring and dynamic dispatching system and method based on digital twins. Background Art

[0002] In recent years, the proportion of renewable energy has gradually increased, and the load on the power grid has increased year by year. However, the intermittent and volatile nature of renewable energy has brought unprecedented challenges to the dispatching of power systems. The power generation of renewable energy such as wind power and solar power is extremely dependent on weather conditions, such as wind speed and light intensity. These natural factors are highly unpredictable, which makes the power generation of renewable energy fluctuate greatly and difficult to accurately predict. This uncertainty makes it difficult for power system dispatchers to formulate stable and reliable power generation plans in advance. In actual operation, it is often necessary to frequently adjust the dispatching plan, which increases the complexity and difficulty of dispatching work, and may lead to problems such as imbalance in supply and demand of the power system and voltage fluctuations, which have an adverse impact on the stability and economy of the system.

[0003] Existing multi-objective optimization scheduling methods attempt to find a balance point between multiple objectives, but the complex and changing power grid environment significantly increases the difficulty and computational complexity of optimization. When dealing with multi-objective optimization problems, it is often difficult to find a reasonable balance point between multiple objectives such as quality and cost.

[0004] To sum up, traditional power grid dispatching optimization methods have at least one of the following technical problems: insufficient optimization accuracy, high labor costs, and weak risk resistance. Summary of the invention

[0005] In view of this, the purpose of the present invention is to provide a digital twin-based power grid intelligent monitoring and dynamic dispatching system and method to solve at least one technical problem that the existing power grid dispatching strategy is inaccurate in multi-objective optimization and susceptible to risks.

[0006] To achieve the above objectives, in a first aspect, the present invention provides a power grid intelligent monitoring and dynamic dispatching system based on digital twins, comprising: A power grid topology generation module, used to obtain power grid geographic data and power equipment basic data, and generate power grid topology and power equipment full life cycle archives based on the power grid geographic data and the power equipment basic data; A modeling module, used to obtain power grid operation status data, and establish a digital twin model based on the power grid operation status data, the power grid topology and the full life cycle archive of the power equipment; A knowledge graph establishment module, used to establish a knowledge graph based on the power grid operation status data, the power grid topology and the power equipment life cycle archive; A mapping module, used to establish a mapping model according to the knowledge graph and the digital twin model, input the power grid operation status data into the mapping model, and obtain a mapping data set; A risk analysis module, used for performing multi-risk analysis based on the mapping data set to generate a comprehensive risk analysis report; The dispatching module is used to obtain a power grid dispatching strategy by adopting a multi-objective adaptive optimization algorithm and a load balancing algorithm according to the comprehensive risk analysis report.

[0007] In a second aspect, the present invention provides a method for intelligent monitoring and dynamic dispatching of a power grid based on digital twins, comprising the following steps: Obtaining grid geographic data and power equipment basic data, and generating grid topology and power equipment full life cycle archives based on the grid geographic data and the power equipment basic data; Acquire power grid operation status data, and establish a digital twin model based on the power grid operation status data, the power grid topology and the full life cycle archive of the power equipment; Establishing a knowledge graph based on the power grid operation status data, the power grid topology and the full life cycle archive of the power equipment; Establishing a mapping model according to the knowledge graph and the digital twin model, inputting the power grid operation status data into the mapping model, and obtaining a mapping data set; Performing a multi-risk analysis based on the mapped data set to generate a comprehensive risk analysis report; According to the comprehensive risk analysis report, a multi-objective adaptive optimization algorithm and a load balancing algorithm are used to obtain a power grid dispatching strategy.

[0008] The above technical solution has the following beneficial technical effects: The present invention generates a full life cycle archive of power grid topology and power equipment, and then constructs a digital twin model to simulate power grid operation. Then, a knowledge graph is constructed by acquiring power grid operation status data in combination with the power grid topology and the full life cycle archive of power equipment. The knowledge graph and the twin digital model are mapped to obtain a mapping data set. A comprehensive risk analysis is performed on the mapping data set, and then a scheduling strategy is generated based on the comprehensive risk analysis report. Compared with the prior art, the risk resistance of the scheduling strategy is effectively improved by combining the comprehensive risk analysis report. At the same time, the accuracy of the scheduling strategy is improved through the comprehensive application of the digital twin model and the knowledge graph. BRIEF DESCRIPTION OF THE DRAWINGS

[0009] The accompanying drawings are used to better understand the present invention and do not constitute an improper limitation of the present invention. Figure 1 It is a structural block diagram of a power grid intelligent monitoring and dynamic dispatching system based on digital twin in an embodiment of the present invention; Figure 2is a structural block diagram of a power grid topology generation module in an embodiment of the present invention; Figure 3 is a structural block diagram of a modeling module in an embodiment of the present invention; Figure 4 It is a structural block diagram of a knowledge graph establishment module in an embodiment of the present invention; Figure 5 is a structural block diagram of a mapping module in an embodiment of the present invention; Figure 6 is a structural block diagram of a risk analysis module in an embodiment of the present invention; Figure 7 is a structural block diagram of a scheduling module in an embodiment of the present invention; Figure 8 It is a flow chart of a method for intelligent monitoring and dynamic dispatching of a power grid based on digital twins in an embodiment of the present invention; Fig. 9 It is a schematic diagram of the structure of a computer system according to an embodiment of the present invention. DETAILED DESCRIPTION

[0010] The following is a description of exemplary embodiments of the present invention in conjunction with the accompanying drawings, including various details of the embodiments of the present invention to facilitate understanding, which should be considered as merely exemplary. Therefore, it should be recognized by those of ordinary skill in the art that various changes and modifications may be made to the embodiments described herein without departing from the scope and spirit of the present invention. Similarly, for clarity and conciseness, the description of well-known functions and structures is omitted in the following description.

[0011] Embodiment 1 like Figure 1 As shown, this embodiment provides a power grid intelligent monitoring and dynamic scheduling system based on digital twins, which specifically includes: a power grid topology generation module, a modeling module, a knowledge graph establishment module, a mapping module, a risk analysis module and a scheduling module.

[0012] A power grid topology generation module, used to obtain power grid geographic data and power equipment basic data, and generate power grid topology and power equipment full life cycle archives based on the power grid geographic data and the power equipment basic data; A modeling module, used to obtain power grid operation status data, and establish a digital twin model based on the power grid operation status data, the power grid topology and the full life cycle archive of the power equipment; A knowledge graph establishment module, used to establish a knowledge graph based on the power grid operation status data, the power grid topology and the power equipment life cycle archive; A mapping module, used to establish a mapping model according to the knowledge graph and the digital twin model, input the power grid operation status data into the mapping model, and obtain a mapping data set; A risk analysis module, used for performing multi-risk analysis based on the mapping data set to generate a comprehensive risk analysis report; The dispatching module is used to obtain a power grid dispatching strategy by adopting a multi-objective adaptive optimization algorithm and a load balancing algorithm according to the comprehensive risk analysis report.

[0013] The embodiment of the present invention provides a digital twin-based power grid intelligent monitoring and dynamic dispatching system, which realizes comprehensive perception, accurate modeling and intelligent dispatching of power grid operation status through the collaborative work of multiple modules such as integrated power grid topology generation, digital twin modeling, knowledge graph construction and mapping analysis. The system can not only fully capture power grid geographic data and equipment life cycle information, but also perform multi-dimensional risk analysis based on knowledge graphs and digital twin models, and generate dynamic dispatching strategies using multi-objective adaptive optimization algorithms and load balancing algorithms, which improves the intelligence level of power grid operation, risk warning capabilities and dispatching accuracy, and is conducive to ensuring the safe and stable operation of the power grid and improving the operational efficiency of the power system.

[0014] Specifically, Figure 2 As shown, in the power grid topology generation module, it includes: a first data acquisition unit, a power grid topology construction unit and a file generation unit. The first data acquisition unit is used to acquire power grid geographic data, power equipment basic data and power equipment connection relationship data. The power grid geographic data includes the power grid geographic map and the location of each device in the power grid, and the power equipment basic data includes power equipment factory parameters, power equipment operation and maintenance records and power equipment historical fault data. The power equipment connection relationship data can come from the power system planning document, the line connection diagram or the geographic information system. The power equipment connection relationship data is a detailed description of the physical and logical connection mode between each device in the power system, specifically including the bus connection topology in the substation, the connection path between the lines, the incoming and outgoing line configuration of the equipment, etc. For example, in a 110kV substation, it can include the connection mode between the main transformer and the bus (single bus, double bus or ring network connection), the connection relationship between the feeder switch and the bus, the tie switch configuration between the lines, and the connection path and direction of each feeder. These data not only record the physical connection status of the power equipment, but also reflect the power supply path, protection configuration and operation logic of the power grid. They can be obtained through technical documents such as substation floor plans, line wiring diagrams, relay protection diagrams, etc., and are the basic data for building the power grid topology.

[0015] The power grid topology construction unit is used to extract a power grid geographic map from the power grid geographic data, extract device connection relationships from the power equipment basic data, and embed the device connection relationships into the power grid geographic map to obtain a power grid topology; the power grid geographic map includes the location of each device in the power grid, and the location of each device in the power grid is taken as a point, and the device connection relationship is taken as an edge. Points are set in the power grid geographic map according to the location of each device in the power grid, and edges are set between the points according to the device connection relationship, thereby generating a power grid topology structure.

[0016] The archive generation unit is used to perform a full life cycle analysis on the basic data of the power equipment to obtain a full life cycle archive of the power equipment. The full life cycle archive of the power equipment is a complete digital mapping of the power equipment from production to scrapping. The equipment factory parameters, equipment operation and maintenance records, and historical fault data are obtained based on the basic data of the power equipment. The equipment factory parameters, equipment operation and maintenance records, and historical fault data are then subjected to semantic association and full life cycle analysis to obtain the full life cycle archive of the power equipment. The full life cycle archive of the power equipment is the basic image of the digital twin model.

[0017] Specifically, in the data collection phase, basic data of power equipment is collected from each stage of equipment production, transportation, operation, maintenance and scrapping, mainly including equipment factory parameters, equipment operation and maintenance records and historical fault data. Factory parameters include equipment model, specifications, manufacturing date, technical indicators, etc.; operation and maintenance records cover the equipment's operating load, working environment, and regular maintenance; historical fault data records equipment fault type, occurrence time, maintenance process and processing results. These data together constitute the basic information set of the equipment's entire life cycle.

[0018] Semantic association is built through ontology modeling and knowledge graph construction to establish semantic links between different stages and types of equipment data. For example, the factory parameters of the equipment are associated with the actual operation data to identify equipment performance deviations; historical fault data is associated with operation and maintenance records to explore the potential laws and warning signals of fault occurrence.

[0019] The full life cycle analysis uses data-driven analysis methods to systematically evaluate and predict the basic data of power equipment. It mainly uses one or more of the following analysis methods: Time series trend analysis is the first level method. By statistically analyzing the historical data of equipment operating parameters (such as temperature, vibration, current, etc.), the equipment performance change curve is drawn to identify the fluctuation characteristics and long-term development trend of equipment performance. Using basic time series algorithms such as moving average and exponential smoothing can preliminarily reveal the evolution law of equipment performance.

[0020] The status assessment model is the second-level method. A multi-dimensional status assessment model is constructed based on the equipment operation data, and the equipment status is divided into three basic levels: normal, sub-healthy, and abnormal. By setting the performance index threshold, the current status of the equipment is quantitatively evaluated. For example, a status judgment model based on fuzzy evaluation can be established, which comprehensively considers multiple operation indicators and gives a comprehensive score for the health status of the equipment.

[0021] Fault prediction is the third level method. Classification and regression algorithms in machine learning are used to establish equipment fault prediction models. Available methods include logistic regression, decision tree, support vector machine, and random forest. These algorithms can learn the characteristics of historical fault data, identify pre-fault warning signals, and predict the type and probability of future faults.

[0022] Remaining life prediction is the fourth level method. Based on the equipment's operating data and performance degradation characteristics, a equipment life prediction model is constructed. The Weibull distribution model or the Cox proportional risk regression model in survival analysis can be used to estimate the remaining service life of the equipment. By analyzing the performance decay curves of key components of the equipment, a more accurate life prediction can be given.

[0023] Risk assessment is the fifth level method. The equipment risk assessment model is constructed by integrating the equipment status, failure prediction and life prediction results. Multi-dimensional risk quantification is performed based on the health status, failure probability and potential impact of the equipment to provide a basis for maintenance decisions.

[0024] The final generated power equipment life cycle archive is a complete digital mapping, covering the entire process information of power equipment from production to scrapping. This archive is not only an electronic archive of the equipment, but also the basic image of the digital twin model. It records the comprehensive life course of the equipment, including factory information, operating status, maintenance history, fault records, etc., providing comprehensive, dynamic and intelligent data support for the full life cycle management of the equipment.

[0025] Specifically, Figure 3 As shown, the modeling module includes: a second data acquisition unit, an initial modeling unit, a simulation unit and a model optimization unit.

[0026] The second data acquisition unit is used to acquire power grid operation status data. The power grid operation status is real-time collected data, including power grid operation data, data acquisition and monitoring control system data, smart meter data or substation monitoring system data.

[0027] The grid operation data covers multiple indicators. Voltage data includes the real-time voltage values ​​of each voltage level line (such as 220kV, 110kV, 35kV), and records the voltage amplitude and phase angle changes at both ends and nodes of the line. Current data records the current size and direction of the line, transformer, and bus, including the current value of each phase, current imbalance, and load rate. Power data is divided into active power and reactive power, recording power changes in power generation, transmission, and power consumption, reflecting the energy transmission and conversion of the power grid. Frequency data monitors the overall frequency of the power grid, recording the real-time value, fluctuation range, and change trend of the frequency.

[0028] The data of the Supervisory Control and Data Acquisition (SCADA) system mainly includes the equipment operation status and alarm information. The switch status data records the on and off positions of the circuit breaker, disconnector, and grounding switch, including the switch opening and closing time, operation times, and action characteristics. The equipment alarm information includes voltage over-limit alarm, current over-limit alarm, temperature abnormality alarm, equipment failure alarm, etc., recording the type, level, occurrence time, and location of the alarm. The protection device action record includes the protection type, action reason, action time, and protection range. The operation log records the detailed information of remote control operations and manual intervention operations, including the operator, operation time, operation instructions, and operation results.

[0029] Smart meter data provides detailed information on the electricity consumption side. Real-time electricity consumption data records the user's instantaneous power consumption, cumulative power consumption, and time-sharing power consumption. The power load curve shows the time-varying characteristics of the power load, including daily load curve, weekly load curve, and monthly load curve. Power quality indicators include voltage fluctuation, flicker, harmonic content, imbalance, etc., reflecting the stability and quality of power supply. Abnormal power consumption behavior identification includes sudden increase in power consumption, sudden decrease in power consumption, abnormal power consumption period, and abnormal power consumption mode.

[0030] Substation monitoring system data focuses on the operating status and health status of equipment. Transformer temperature data includes winding temperature, oil temperature and top oil temperature, and monitors equipment temperature changes. Partial discharge signals record the discharge characteristics of equipment insulation components, including discharge intensity, discharge frequency and discharge location. Insulation status monitoring includes insulation resistance, dielectric loss angle, partial discharge test results, etc., to evaluate equipment insulation performance. Equipment vibration and noise data records bearing vibration, mechanical vibration spectrum and noise intensity, reflecting the mechanical status and potential faults of the equipment.

[0031] The initial modeling unit is used to build an initial digital twin model based on the power grid topology and the full life cycle archive of the power equipment.

[0032] When constructing the initial digital twin model, first, the goal and scope of the initial digital twin model are clarified based on the grid topology and the life cycle archive of the power equipment, including the simulation target of the grid operation system, the operation status of each power equipment, the effect achieved by the simulation, and the simulation optimization target. For example, for the grid system, the goal can be to monitor the status of grid equipment in real time, predict equipment failures, optimize energy resource allocation, etc. Grid-related data, including electrical quantity data, equipment status data, environmental data, etc., are obtained through the grid topology and the life cycle archive of the power equipment. These data will be used to build and update the initial digital twin model.

[0033] Clean and format the collected power grid-related data to ensure data consistency and comparability. This step includes removing noise, filling missing values, and standardizing data.

[0034] Based on the preprocessed grid-related data, a mathematical model is established to describe the behavior of the grid and grid equipment. This can include physical modeling, statistical modeling or machine learning methods, and the appropriate modeling method is selected according to the complexity and needs of the actual situation. Specifically, after the data preprocessing is completed, the next step is to establish a mathematical model that can accurately describe the behavior of the grid and power equipment. According to the complexity and specific needs of the power grid system, a variety of modeling methods can be used: physical modeling accurately describes the electrical characteristics and dynamic behavior of power grid equipment by constructing mathematical equations based on physical laws and power system principles; statistical modeling uses historical data and statistical analysis methods to establish a probability model and / or trend prediction model for the operating status of the equipment; machine learning methods, such as neural networks, support vector machines, etc., can learn the nonlinear characteristics of the power grid system through a large amount of training data to achieve more intelligent and adaptive modeling. These modeling methods are not mutually exclusive, but can be flexibly combined and integrated according to specific application scenarios.

[0035] According to the established mathematical model, the initial digital twin model is constructed. This includes determining the topology of the power grid, setting equipment parameters, selecting simulation algorithms, etc., in order to simulate and monitor the behavior of power grid equipment. Specifically, based on the mathematical model established earlier, the final construction of the initial digital twin model requires refinement and parameterization. This includes restoring the topology of the power grid, such as node connections, line configurations, substation layouts, etc.; setting detailed parameters of each power equipment, such as the impedance of the transformer, the characteristic curve of the generator, and the transmission parameters of the line; selecting appropriate simulation algorithms, such as power flow calculation algorithms, transient stability analysis algorithms, equipment fault simulation algorithms, etc. Through these steps, a digital twin model that can dynamically and real-time simulate and monitor the behavior of power grid equipment can be constructed.

[0036] The simulation unit is used to obtain a number of simulation results by adopting a cross-scale multi-physics field coupling modeling algorithm based on the full life cycle archive of the power equipment and the power grid topology.

[0037] The model optimization unit is used to optimize the initial digital twin model using a deep reinforcement learning algorithm according to the simulation results and the power grid operation status data to obtain a digital twin model. Specifically, the model optimization unit uses a deep reinforcement learning algorithm to iteratively optimize the initial model. Specifically, the model optimization unit continuously optimizes the initial digital twin model based on the simulation results and the power grid operation status data through a deep reinforcement learning algorithm. Specifically, the unit first constructs a multi-dimensional state space to integrate the operation data, historical fault information and real-time monitoring data of the power grid equipment, such as the temperature curve, insulation resistance change, partial discharge characteristics, load change rate, etc. of the transformer. These data are standardized and normalized and converted into feature vectors that can be directly learned by the neural network. The action space is defined as a series of executable equipment maintenance and control strategies, such as adjusting cooling system parameters, local insulation repair, load reduction operation or early warning replacement of components. The deep reinforcement learning algorithm, represented by the deep Q network (DQN), constructs a complex state-action mapping model through a multi-layer neural network, and designs a reward function based on equipment health, failure risk and maintenance cost. In a simulated power grid environment, the algorithm will conduct multiple rounds of decision-making learning. Each round of decision-making will record state transitions, execution actions, and immediate rewards obtained. Through the experience replay mechanism, it will randomly sample from the historical decision sequence for deep learning, gradually adjust the neural network weights, and optimize the decision-making strategy. For example, for a key transformer, the model learns that when the temperature exceeds 75°C and the insulation resistance drops by 20%, the optimal strategy is to start the backup cooling system and reduce the operating load by 30% to minimize the risk of equipment failure and maintenance costs. By continuously evaluating the deviation between the model prediction results and the actual operating data, and dynamically adjusting the learning rate and exploration strategy, the model optimization unit can achieve adaptive optimization of the digital twin model, continuously improve the simulation and prediction capabilities of the dynamic characteristics of the complex power grid system, and ultimately form a digital twin model with highly intelligent and self-learning characteristics.

[0038] Specifically, the simulation unit includes: an electromagnetic field coupling subunit, a thermal field coupling subunit and a mechanical stress coupling subunit.

[0039] The electromagnetic field coupling subunit is used to construct an electromagnetic field distribution model of the power equipment according to the power grid topology and the full life cycle archive of the power equipment, and obtain electromagnetic field simulation results by using a finite element analysis method based on the electromagnetic field distribution model. The electromagnetic field distribution model is used to simulate the electromagnetic field distribution around the power equipment, and the electromagnetic field intensity distribution inside and outside the power equipment is calculated by using a finite element analysis method. The potential interference effect of the electromagnetic field on the insulation, communication and control system of the power equipment is analyzed based on the electromagnetic field intensity distribution to obtain electromagnetic field simulation results.

[0040] The thermal field coupling subunit is used to construct a device temperature field distribution model based on the full life cycle archive of the power equipment to obtain thermal field simulation results. The implementation steps of the thermal field coupling subunit are as follows: first, the equipment material parameters and equipment operating conditions are obtained based on the full life cycle archive of the power equipment. Then, the equipment material parameters and the equipment operating conditions are input into the transient heat conduction equation to obtain the equipment temperature field distribution model. Finally, the impact of temperature changes on the equipment structural integrity and equipment performance is calculated based on the equipment temperature field distribution model, and the thermal field simulation results are generated based on the impact of the temperature changes on the equipment structural integrity and equipment performance.

[0041] Specifically, equipment material parameters refer to indicators that describe the thermodynamic and physical properties of equipment materials, mainly including the thermal conductivity, specific heat capacity, thermal expansion coefficient, density, etc. of the material. These parameters reflect the thermodynamic behavior of the material under different temperature and stress conditions. For example, for transformer windings, it is necessary to obtain specific values ​​such as the thermal conductivity of copper wire, the thermal expansion coefficient of insulating materials, and the specific heat capacity of transformer oil. The acquisition of these parameters not only depends on the factory technical specifications provided by the manufacturer, but also needs to be combined with real-time monitoring data during equipment operation, such as insulation aging, material performance changes, etc. The equipment operating conditions are comprehensive indicators that describe the working environment and operating conditions of the equipment during actual operation, mainly including load current, ambient temperature, cooling method, operating time, load rate, etc. Specifically, for power equipment, the operating conditions can include daily average load change curves, seasonal load characteristics, geographical location environmental factors, etc.

[0042] The mechanical stress coupling subunit is used to obtain mechanical stress simulation results based on the stress-strain curve and cumulative damage theory according to the full life cycle archive of the power equipment. The mechanical stress simulation results are used to reflect the structural response of the power equipment under complex loads. The implementation steps of the mechanical stress coupling subunit are first to use the finite element dynamics analysis method to calculate the mechanical stress distribution of the equipment under power grid operation, external disturbances and extreme climate conditions. Then, based on the mechanical stress distribution of the equipment under power grid operation, external disturbances and extreme climate conditions, based on the stress-strain curve and cumulative damage theory, the fatigue damage degree of the equipment and the remaining service life of the equipment are calculated. Finally, the mechanical stress simulation results are generated according to the fatigue damage degree of the equipment and the remaining service life of the equipment.

[0043] Specifically, Figure 4 As shown, the knowledge graph establishment module includes: a preprocessing unit, a semantic association unit, an ontology unit, a semantic reasoning unit and a knowledge graph establishment unit.

[0044] The preprocessing unit is used to preprocess the power grid operation status data to obtain a monitoring data set. The power grid operation status data is multi-source data collected in real time. The preprocessing includes multi-protocol adaptation conversion and data standardization processing. The data standardization processing includes data cleaning, abnormal data identification, abnormal data repair and normalization processing. Standardized data in a unified format is obtained through the preprocessing.

[0045] The semantic association unit is used to extract the topological features and life cycle features of the monitoring data set according to the power grid topology and the full life cycle archive of the power equipment, and generate semantic constraint relationships according to the topological features and the life cycle features. In the semantic association unit, the topological features of the monitoring data set are extracted according to the power grid topology, the life cycle features in the monitoring data set are extracted according to the full life cycle archive of the power equipment, and the semantic association between the topological features and the life cycle features is set. Specifically, in the semantic association unit, the topological features of the monitoring data set are first extracted through the power grid topology data, and the specific method includes analyzing the connection relationship between device nodes, the spatial layout of power lines and the hierarchical structure of the power grid. For example, for the monitoring data set of a substation, the spatial location, connection mode and network hierarchical information of the transmission lines, transformers, switches and other equipment connected thereto can be extracted. In addition, the life cycle features in the monitoring data set are extracted through the full life cycle archive of the power equipment, including the factory date, commissioning time, maintenance records, technical parameter change history, etc. of the equipment. For the same substation, the operating years, cumulative operating time, historical fault times and maintenance conditions of each equipment can be extracted. On this basis, semantic associations are set between topological features and lifecycle features, such as associating the spatial location of the equipment with its operating status, and associating the network connection relationship of the equipment with its maintenance history, so as to construct a semantic constraint relationship network.

[0046] The realization process of generating semantic constraint relationships is achieved by establishing an association mapping between topological features and lifecycle features. Specifically, a multidimensional feature association matrix is ​​first constructed, which takes topological features as the horizontal axis and lifecycle features as the vertical axis. The relationship between different features is quantified by setting semantic association rules, thereby describing the semantic association strength between features. For example, for a 110kV substation device, its topological features (such as device location, network connection type) can be associated with lifecycle features (such as equipment age, cumulative operating time, and historical fault frequency). By defining semantic association rules, for example, "equipment with an age of more than 15 years and connected to important transmission lines at key nodes" will be given a higher risk semantic constraint weight. In specific implementation, a semantic association scoring algorithm can be used to generate a multidimensional semantic constraint relationship vector for each device based on the predefined multidimensional feature association matrix and semantic association rules. This vector not only reflects the topological features and lifecycle features of the device, but also captures its potential operating risks and importance, providing a semantic constraint basis for subsequent ontology modeling and knowledge graph construction. In this way, abstract topological and lifecycle features can be transformed into specific and quantifiable semantic association relationships.

[0047] The ontology unit is used to construct a semantic ontology model according to the semantic constraint relationship by adopting the ontology semantic mapping method. The semantic ontology model is used to normalize the data concept domain, standardize the attribute features, and eliminate semantic ambiguity. Specifically, the ontology unit is a module specially used to process semantic constraint relationships. Its main function is to convert complex semantic constraint relationships into a structured and standardized ontology model through the semantic ontology mapping method. This unit is used to realize the normalization of the data concept domain, standardize the attribute features, and effectively eliminate semantic ambiguity, thereby providing a unified semantic expression and understanding framework. The ontology unit consists of three modules: a semantic mapping submodule, a concept domain standardization submodule, and a semantic disambiguation submodule. The semantic mapping submodule is used to convert the input semantic constraint relationship vector into a normalized ontology model; the concept domain standardization submodule is used to define the boundaries of the data concept domain, which helps to build a standardized representation of power grid-related equipment and parameters, determine the attribute feature mapping rules, which ensure the uniformity between physical equipment and virtual models, and perform semantic consistency verification. The semantic consistency verification is used to ensure the consistency between the various parts of the data in the ontology model (such as concepts, attributes, relationships, etc.); the semantic disambiguation submodule is used to ensure the semantic accuracy and consistency of the model through at least one of multidimensional semantic association comparison, ambiguity identification algorithm and semantic conflict resolution method.

[0048] The data concept domain refers to the scope of various grid-related concepts and data structures involved in the system. In the grid intelligent monitoring and dynamic dispatching system, the data concept domain includes multiple key elements. For example, grid equipment such as substations, distribution networks, generators and circuit breakers constitute the physical basis of the grid; electrical parameters such as voltage, current, power and frequency are key indicators used to describe the operation status of the grid; operating status includes normal operation, fault, overload, etc., which are used to indicate the health status of various parts of the grid; and dispatching decisions include load distribution, fault isolation and energy transmission optimization, etc., to ensure the efficient and safe operation of the grid.

[0049] Taking the ontology modeling of power equipment as an example, the implementation process includes four steps. First, extract semantic constraint relationships, including topological features (such as location, network connection) and life cycle features (such as service life, operating time) of the equipment. Secondly, perform semantic mapping, convert discrete features into standardized semantic descriptions, and establish a feature association rule base. Third, perform conceptual domain normalization, clearly define the boundaries of equipment types, and standardize attribute descriptions, such as accurately locating the geographic coordinates and network connection type of "110kV substation equipment." Finally, identify and resolve existing semantic conflicts through the semantic disambiguation submodule to ensure the consistency and accuracy of the description.

[0050] The technical implementation of the ontology unit includes the semantic mapping algorithm and the semantic disambiguation algorithm. The semantic mapping algorithm takes the feature vector as input and generates a standardized ontology model through feature standardization, semantic association scoring and ontology mapping. The semantic disambiguation algorithm dynamically adjusts the feature vector by identifying semantic conflicts, applying predefined disambiguation rules, and finally outputs a standardized feature representation.

[0051] The semantic reasoning unit is used to generate multidimensional semantic associations using a relational deduction algorithm based on the semantic ontology model. The relational deduction algorithm is used to deeply analyze the structural associations and evolutionary relationships between data entities, and to mine potential logical associations and semantic dependencies, thereby obtaining multidimensional semantic associations.

[0052] Specifically, in the knowledge graph building module, the semantic reasoning unit first relies on the semantic association unit to extract key features from the power grid topology and equipment life cycle data, and generates corresponding semantic constraint relationships. Through these semantic constraint relationships, the semantic ontology model is constructed using the ontology semantic mapping method. The semantic ontology model is the core structure in the power grid monitoring system. It defines the elements such as various devices, topological structures, operating states, and their relationships in the power grid in a standardized way, providing the necessary knowledge basis for subsequent reasoning.

[0053] The task of the semantic reasoning unit is to use the relational deduction algorithm to reason about the semantic ontology model. The relational deduction algorithm is based on the operation rules of the power grid, the characteristics of the equipment and their mutual relationships, combined with the monitoring data for reasoning and analysis. Through this reasoning method, the system can identify the potential connections between different power grid elements based on known knowledge, and obtain new information or inference results through reasoning, thereby enhancing the intelligence and adaptability of the power grid monitoring system.

[0054] Through the application of relational deduction algorithms, the semantic reasoning unit can generate multi-dimensional semantic associations. These multi-dimensional semantic associations involve multiple levels, such as changes in power grid topology, dynamic changes in equipment life cycle, and prediction of power grid operation status. Specifically, the association of power grid topology can show the connection relationship between devices and the changes in the topological structure of the power grid; the association of equipment life cycle involves the health status of the equipment and the expected failure period; and the association of operation status helps the system predict potential overload, failure and other problems of the power grid.

[0055] The generated multi-dimensional semantic associations provide more in-depth and comprehensive data support for the power grid monitoring system, helping to timely discover potential operational risks. Through these associations, the system can identify and predict possible anomalies or failures in the power grid in advance, providing more accurate data support for the risk analysis module.

[0056] The knowledge graph establishment unit is used to establish a knowledge graph according to the multi-dimensional semantic association. The knowledge graph represents the association relationship between the semantics of power grid devices.

[0057] Specifically, Figure 5 As shown, the mapping module includes: an association matching unit, a mapping unit and a data set output unit.

[0058] The association matching unit is used to extract power equipment graph data from the knowledge graph, and perform identifier association matching between the power equipment graph data and the digital twin model to obtain an identifier association relationship. The power equipment graph data is used to represent the relationship between the semantics of power equipment. The implementation steps of the association matching unit are as follows: a twin is obtained according to the digital twin model. The twin is a simulation of the operating state of the power grid, and the twin includes the connection relationship and operating state of the power grid equipment.

[0059] The mapping unit is used to perform data mapping processing on the knowledge graph and the digital twin model according to the identifier association relationship to obtain a mapping model; The data set output unit is used to input the power grid operation status data into the mapping model to obtain a mapping data set. The mapping data set can be calculated by edge computing method, by deploying the mapping model on different servers, each server obtains the power grid operation status data from different sources, and the process of inputting the power grid operation status data into the mapping model is performed in each server, and then the output of the mapping model in each server is summarized to obtain a mapping data set, thereby realizing parallel processing of data and improving data processing speed.

[0060] Specifically, Figure 6 As shown, the risk analysis module includes: a load forecasting unit, an equipment failure analysis unit, an extreme weather analysis unit and a risk analysis unit.

[0061] The load prediction unit is used to obtain several load prediction results of the power grid by adopting a hybrid ensemble learning algorithm according to the mapping data set.

[0062] As a preferred implementation of this embodiment, the number of the several load forecasting results is the same as the number of algorithms used in the hybrid ensemble learning algorithm. This embodiment takes two algorithms as an example, and other algorithms can be added for prediction according to actual conditions. In the load forecasting unit, it includes: a first prediction subunit, a second prediction subunit and a prediction output subunit. The first prediction subunit is used to obtain a first prediction result by using a long short-term memory network according to the mapping data set. The long short-term memory network captures the change law of the power grid load in the time dimension through the memory and forgetting mechanism, so as to obtain the load forecast result at a specified time as the first prediction result. The second prediction subunit is used to obtain a second prediction result by using a random forest algorithm according to the mapping data set. The implementation steps of the second prediction subunit include: first, dividing the mapping data set into several random forest training sets. Secondly, using a random forest algorithm to establish several random forest decision trees according to the several random forest training sets, the number of the random forest decision trees is the same as the number of the random forest training sets. Then, generating several decision results according to the several random forest decision trees. Finally, using a voting algorithm to vote on several decision results to obtain a second prediction result. The first prediction result and the second prediction result are prediction results for the same time period. The prediction output subunit is used to calculate a comprehensive prediction result based on the first prediction result and the second prediction result, select different prediction time periods, calculate several comprehensive prediction results, obtain a load change trend based on the several comprehensive prediction results, and generate a load prediction result based on the load change trend. When calculating the comprehensive prediction result based on the first prediction result and the second prediction result, the calculation can be performed by average value or weighted sum. The different prediction time periods include at least two time periods, wherein the first time period is a day, a week or a month, and the second time period is a quarter or a year. The long short-term memory network and the random forest algorithm are trained by historical operation data, and the historical operation data includes loads at different times during the operation of the power grid.

[0063] The equipment failure analysis unit is used to obtain an equipment health analysis report by using a Bayesian algorithm and a Markov transition probability model based on a number of the load forecasting results, the full life cycle archive of the power equipment and the mapping data set.

[0064] Specifically, the equipment fault analysis unit includes: a first feature extraction subunit, a Bayesian calculation subunit, a Markov calculation subunit and an equipment health analysis report generation subunit.

[0065] The first feature extraction subunit is used to extract features from the load forecast results, the life cycle archive of the power equipment and the mapping data set to obtain a number of key feature vectors. In the first feature extraction subunit, the load forecast results, the life cycle archive of the power equipment and the mapping data set are first standardized to unify the formats of different data. Wavelet transform and principal component analysis are used to remove noise from the data in the unified format. Finally, feature extraction is performed on the standardized and denoised data to obtain a number of key feature vectors. The key feature vectors refer to features that reflect the functional status of the equipment.

[0066] The Bayesian calculation subunit is used to calculate the conditional probabilities of different fault modes using the Bayesian algorithm based on the key feature vectors and the mapping data set. In the Bayesian calculation subunit, historical fault data is first obtained, and then the conditional probabilities of faults under different key features are calculated using the Bayesian algorithm based on the historical fault data. Then, real-time monitoring data is obtained based on the mapping data set, and the current fault probability is calculated using the Bayesian algorithm based on the real-time monitoring data. The conditional probabilities of different fault modes are obtained based on the current fault probability and the conditional probabilities of faults under different key features.

[0067] The Markov calculation subunit is used to establish a Markov transition probability model according to the conditional probabilities of the different fault modes, generate a device performance degradation curve according to the mapping data set and the Markov transition probability model, and obtain the device degradation inflection point and the device degradation critical state according to the device performance degradation curve. In the Markov calculation subunit, a Markov transition probability model is first established according to the conditional probabilities of the different fault modes, and then the real-time operation data and the fault probability are obtained according to the mapping data set. According to the Markov transition probability model, the real-time operation data and the fault probability, the device performance degradation curve is generated, and the device degradation inflection point and the device degradation critical state are obtained according to the device performance degradation curve.

[0068] The equipment health analysis report generation subunit is used to process the conditional probabilities of the different failure modes, the degradation inflection points and the degradation critical states by a multi-dimensional comprehensive evaluation method to obtain an equipment health analysis report. In the equipment health analysis report generation subunit, a number of equipment health status indicators are first established, and then each equipment health indicator is assigned a value according to the conditional probabilities of the different failure modes, the equipment degradation inflection points and the equipment degradation critical states. Finally, a weighted sum is performed according to the assigned values ​​of each equipment health indicator to obtain an equipment health status index, and an equipment health analysis report is generated according to the equipment health status index.

[0069] The extreme weather analysis unit is used to obtain extreme weather warning data, and obtain the extreme weather analysis report by using the Monte Carlo algorithm according to the extreme weather warning data and the equipment health analysis report. The extreme weather includes heavy rainstorms, extreme low temperatures and strong typhoons. First, the equipment health analysis report and the extreme weather warning data are input into a pre-trained multi-dimensional risk identification model, and the multi-dimensional risk identification model simulates the equipment health analysis reports under different extreme weather conditions by using the Monte Carlo simulation method. The impact of extreme weather on equipment health is obtained according to the equipment health analysis reports under different extreme weather conditions, and the impact of extreme weather on equipment health is quantified as a risk score, and the risk is divided into several risk levels, each risk level corresponds to a risk score interval, and the corresponding risk level is found according to the risk score, and an extreme weather analysis report is generated according to the risk score and the risk level. The pre-trained multi-dimensional risk identification model is trained by historical weather data and historical equipment health analysis reports.

[0070] The risk analysis unit is used to construct a multidimensional risk indicator system using the entropy weight method according to the several load forecast results of the power grid, the equipment health analysis report and the extreme weather analysis report, and generate a comprehensive risk analysis report according to the multidimensional risk indicator system. In the risk analysis module, several risk indicators are first generated according to the load forecast results, the equipment health analysis report and the extreme weather analysis report. The entropy weight method is used to assign weights to each of the risk indicators, and then each of the risk indicators is assigned a value according to the several load forecast results of the power grid, the equipment health analysis report and the extreme weather analysis report. A weighted calculation is performed according to the value and weight of each risk indicator to obtain a multidimensional risk score, and finally a comprehensive risk analysis report is generated according to the multidimensional risk score. The comprehensive risk analysis report includes the score of each risk indicator, the score weight of the risk indicator, several load forecast results of the power grid, the equipment health analysis report, the extreme weather analysis report, the multidimensional risk score and the impact of each item in the comprehensive risk analysis report on the operation of the power grid.

[0071] Specifically, Figure 7As shown, in the scheduling module, it includes: an initial strategy generation unit, a resource configuration unit, an emergency plan construction unit and a strategy optimization unit. The initial strategy generation unit is used to generate an initial power grid scheduling strategy based on the mapping data set and the comprehensive risk analysis report by adopting a multi-objective adaptive optimization algorithm. The multi-objective adaptive optimization algorithm is used to use the mapping data set and the comprehensive risk analysis report as input values, and solve according to a plurality of preset objective functions to obtain an initial power grid scheduling strategy. The plurality of objective functions include the lowest failure rate, the lowest operating cost, the lowest carbon emissions or the highest stability. The resource configuration unit is used to obtain a power resource configuration plan based on the comprehensive risk analysis report and the initial power grid scheduling strategy by adopting a load balancing algorithm. First, the initial power resource configuration plan corresponding to the initial scheduling strategy is obtained, and the initial power resource configuration plan is adjusted to load balance to obtain a power resource configuration plan.

[0072] Specifically, the emergency plan construction unit includes: a scenario generation subunit, a second feature extraction subunit, a similarity calculation subunit and an emergency plan output subunit. The scenario generation subunit is used to construct a number of potential risk scenarios using natural language processing technology according to the comprehensive risk analysis report. The second feature extraction subunit is used to extract features from the potential risk scenarios to obtain a number of risk semantic feature vectors. In the second feature extraction subunit, the element gradient is firstly performed on the potential risk scenarios to obtain a number of key elements, which include: risk type, impact scope and severity, etc. Then, mapping is performed according to the key elements to obtain a number of risk semantic feature vectors. The potential risk scenarios are digitally represented by the risk semantic feature vectors. The similarity calculation subunit is used to calculate the similarity between each emergency plan set in the emergency disposal knowledge base and the risk semantic feature vectors. Specifically, in the similarity calculation subunit, the cosine similarity algorithm is used to calculate the similarity, and the similarity is used to reflect the similarity between the plans in the emergency plan set and the potential risk scenarios. The greater the similarity, the more similar the two are, and the better the effect of the emergency plan. The emergency plan output subunit is used to select the emergency plan set with the greatest similarity as the pre-selected plan set, and select the scenario emergency plan from the pre-selected plan set according to the power resource allocation plan. The cosine similarity algorithm introduces a weight adaptive adjustment mechanism of machine learning, and optimizes the parameters in the cosine similarity algorithm according to historical emergency plans and historical potential risk scenarios. It also includes a feasibility verification subunit, which is used to perform feasibility verification on the scenario emergency plan according to the resource allocation plan and the comprehensive risk analysis report. If the feasibility verification fails, the current pre-selected plan is deleted and re-selected until it passes the feasibility verification.

[0073] Specifically, the strategy optimization unit is used to optimize the initial power grid dispatching strategy using an adaptive closed-loop optimization algorithm according to the emergency plans for the several scenarios and the comprehensive risk analysis report to obtain a power grid dispatching strategy. The adaptive closed-loop optimization algorithm first obtains a closed-loop optimization target by quantifying the performance deviation between the initial power grid dispatching strategy and the target dispatching strategy, and sets a reward function according to the closed-loop optimization target. During the adaptive closed-loop optimization process, the initial power grid dispatching strategy is optimized by the reward function until the closed-loop optimization target is met. The initial power grid dispatching strategy at this time is subjected to a convergence check. If it fails, the iteration is continued. If it passes the convergence check, the initial power grid dispatching strategy at this time is output as the power grid dispatching strategy.

[0074] Embodiment 2 like Figure 8 As shown, an embodiment of the present invention also provides a method for intelligent monitoring and dynamic scheduling of a power grid based on digital twins, comprising the following steps: S10: Acquire power grid geographic data and power equipment basic data, and generate power grid topology and power equipment full life cycle archives based on the power grid geographic data and the power equipment basic data; S20: Acquire power grid operation status data, and establish a digital twin model according to the power grid operation status data, the power grid topology, and the full life cycle archive of the power equipment; S30: Establishing a knowledge graph based on the power grid operation status data, the power grid topology and the power equipment life cycle archive; S40: establishing a mapping model according to the knowledge graph and the digital twin model, inputting the power grid operation status data into the mapping model, and obtaining a mapping data set; S50: Performing multi-risk analysis according to the mapping data set to generate a comprehensive risk analysis report; S60: Obtaining a power grid dispatching strategy using a multi-objective adaptive optimization algorithm and a load balancing algorithm according to the comprehensive risk analysis report.

[0075] Specifically, in step S10, the following steps are included: S11: Obtaining power grid geographic data and power equipment basic data; S12: extracting a power grid geographic map from the power grid geographic data, extracting a device connection relationship from the power equipment basic data, and embedding the device connection relationship into the power grid geographic map to obtain a power grid topology; S13: Perform a full life cycle analysis on the basic data of the power equipment to obtain a full life cycle file of the power equipment.

[0076] Specifically, in step S20, the following steps are included: S21: Obtaining power grid operation status data; S22: constructing an initial digital twin model according to the power grid topology and the full life cycle archive of the power equipment; S23: obtaining a number of simulation results using a cross-scale multi-physics field coupling modeling algorithm according to the full life cycle archive of the power equipment and the power grid topology; S24: Based on the simulation results and the power grid operation status data, a deep reinforcement learning algorithm is used to optimize the initial digital twin model to obtain a digital twin model.

[0077] Specifically, in step S30, the following steps are included: S31: preprocessing the power grid operation status data to obtain a monitoring data set; S32: extracting the topological features and life cycle features of the monitoring data set according to the power grid topology and the full life cycle archive of the power equipment, and generating semantic constraint relationships according to the topological features and the life cycle features; S33: constructing a semantic ontology model using an ontology semantic mapping method according to the semantic constraint relationship; S34: Generate multidimensional semantic associations using a relational deduction algorithm according to the semantic ontology model; S35: Establish a knowledge graph based on the multi-dimensional semantic association.

[0078] Specifically, in step S40, the following steps are included: S41: extracting power equipment graph data from the knowledge graph, and performing identifier association matching between the power equipment graph data and the digital twin model to obtain an identifier association relationship; S42: Performing data mapping processing on the knowledge graph and the digital twin model according to the identifier association relationship to obtain a mapping model; S43: Inputting the power grid operation status data into the mapping model to obtain a mapping data set.

[0079] Specifically, in step S50, the following steps are included: S51: obtaining several load forecasting results of the power grid by using a hybrid ensemble learning algorithm according to the mapping data set; S52: Obtaining an equipment health analysis report using a Bayesian algorithm and a Markov transition probability model according to the load forecast results, the power equipment life cycle archive, and the mapping data set; S53: Acquire extreme weather warning data, and obtain an extreme weather analysis report using a Monte Carlo algorithm according to the extreme weather warning data and the equipment health analysis report; S54: Based on the several load forecast results of the power grid, the equipment health analysis report and the extreme weather analysis report, an entropy weight method is used to construct a multidimensional risk indicator system, and a comprehensive risk analysis report is generated based on the multidimensional risk indicator system.

[0080] Specifically, in step S60, the following steps are included: S61: Generate an initial power grid dispatching strategy using a multi-objective adaptive optimization algorithm according to the mapping data set and the comprehensive risk analysis report; S62: According to the comprehensive risk analysis report and the initial power grid dispatching strategy, a load balancing algorithm is used to obtain a power resource configuration plan; S63: Identify the comprehensive risk analysis report, obtain potential risk scenarios, and construct scenario emergency plans using a scenario matching algorithm according to the potential risk scenarios, a preset emergency response knowledge base, and the power resource configuration plan; S64: According to the emergency plans for the several scenarios and the comprehensive risk analysis report, an adaptive closed-loop optimization algorithm is used to optimize the initial power grid dispatching strategy to obtain a power grid dispatching strategy.

[0081] Specifically, in step S23, the following steps are included: S231: constructing an electromagnetic field distribution model of the electric power equipment according to the power grid topology and the full life cycle archive of the electric power equipment, and obtaining an electromagnetic field simulation result by using a finite element analysis method according to the electromagnetic field distribution model; S232: constructing a temperature field distribution model of the equipment according to the full life cycle archive of the power equipment, and obtaining a thermal field simulation result; S233: According to the life cycle archive of the power equipment, a mechanical stress simulation result is obtained by using a stress-strain curve and cumulative damage theory.

[0082] Specifically, in step S51, the following steps are included: S511: Obtaining a first prediction result using a long short-term memory network according to the mapping data set; S512: The mapping data set uses a random forest algorithm to obtain a second prediction result; S513: Calculate a comprehensive prediction result based on the first prediction result and the second prediction result, select different prediction time periods, calculate and obtain a number of comprehensive prediction results, obtain a load change trend based on the several comprehensive prediction results, and generate a load prediction result based on the load change trend.

[0083] Specifically, in step S52, the following steps are included: S521: extracting features from the load forecast result, the life cycle archive of the power equipment and the mapping data set to obtain a number of key feature vectors; S522: Calculating conditional probabilities of different failure modes using a Bayesian algorithm according to the key feature vectors and the mapping data set; S523: establishing a Markov transition probability model according to the conditional probabilities of the different failure modes, generating a device performance degradation curve according to the mapping data set and the Markov transition probability model, and obtaining a device degradation inflection point and a device degradation critical state according to the device performance degradation curve; S524: Using a multi-dimensional comprehensive evaluation method to process the conditional probabilities of the different failure modes, the degradation inflection points, and the degradation critical states, to obtain an equipment health analysis report.

[0084] Specifically, in step S63, the following steps are included: S631: construct several potential risk scenarios using natural language processing technology based on the comprehensive risk analysis report; S632: Extracting features from the potential risk scenarios to obtain risk semantic feature vectors; S633: Calculate the similarity between each emergency plan set in the emergency disposal knowledge base and the risk semantic feature vectors; S634: Select the emergency plan set with the greatest similarity as the pre-selected plan set, and select a scenario emergency plan from the pre-selected plan set according to the power resource configuration plan.

[0085] The technicians in the relevant field can clearly understand that for the convenience and simplicity of description, only the division of the above-mentioned functional units and modules is used as an example for illustration. In practical applications, the above-mentioned function allocation can be completed by different functional units and modules as needed, that is, the internal structure of the device can be divided into different functional units or modules to complete all or part of the functions described above. The functional units and modules in the embodiment can be integrated in a processing unit, or each unit can exist physically separately, or two or more units can be integrated in one unit. The above-mentioned integrated unit can be implemented in the form of hardware or in the form of software functional units. In addition, the specific names of the functional units and modules are only for the convenience of distinguishing each other, and are not used to limit the scope of protection of this application. The specific working process of the units and modules in the above-mentioned system can refer to the corresponding process in the aforementioned method embodiment, which will not be repeated here.

[0086] An embodiment of the present invention also provides a computer-readable storage medium having a computer program stored thereon. When the program is executed by a processor, the program implements any of the above-mentioned digital twin-based intelligent monitoring and dynamic scheduling methods for power grids.

[0087] If the integrated module / unit is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the present invention implements all or part of the processes in the above-mentioned embodiment method, and can also be completed by instructing the relevant hardware through a computer program. The computer program can be stored in a computer-readable storage medium, and the computer program can implement the steps of the above-mentioned various method embodiments when executed by the processor. Among them, the computer program includes computer program code, and the computer program code can be in source code form, object code form, executable file or some intermediate form. The computer-readable medium may include: any entity or device capable of carrying the computer program code, recording medium, U disk, mobile hard disk, disk, optical disk, computer memory, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), electric carrier signal, telecommunication signal and software distribution medium. Of course, there are other forms of readable storage media, such as quantum memory, graphene memory, etc. It should be noted that the content contained in the computer-readable medium can be appropriately increased or decreased according to the requirements of legislation and patent practice in the jurisdiction. For example, in some jurisdictions, according to legislation and patent practice, computer-readable media does not include electrical carrier signals and telecommunication signals.

[0088] The present invention also provides an electronic device. The electronic device of an embodiment of the present invention includes: one or more processors; a storage device for storing one or more programs, and when the one or more programs are executed by the one or more processors, the one or more processors implement a digital twin-based power grid intelligent monitoring and dynamic scheduling method provided by the present invention.

[0089] Reference below Fig. 9 , which shows a schematic diagram of the structure of a computer system 800 of an electronic device suitable for implementing an embodiment of the present invention. Fig. 9 The electronic device shown is only an example and should not bring any limitation to the functions and scope of use of the embodiments of the present invention.

[0090] like Fig. 9As shown, the computer system 800 includes a central processing unit (CPU) 801, which can perform various appropriate actions and processes according to a program stored in a read-only memory (ROM) 802 or a program loaded from a storage part 808 to a random access memory (RAM) 803. In the RAM 803, various programs and data required for the operation of the computer system 800 are also stored. The CPU 801, the ROM 802, and the RAM 803 are connected to each other through a bus 804. An input / output (I / O) interface 805 is also connected to the bus 804.

[0091] The following components are connected to the I / O interface 805: an input section 806 including a keyboard, a mouse, etc.; an output section 807 including a cathode ray tube (CRT), a liquid crystal display (LCD), etc., and a speaker, etc.; a storage section 808 including a hard disk, etc.; and a communication section 809 including a network interface card such as a LAN card, a modem, etc. The communication section 809 performs communication processing via a network such as the Internet. A drive 810 is also connected to the I / O interface 805 as needed. A removable medium 811, such as a magnetic disk, an optical disk, a magneto-optical disk, a semiconductor memory, etc., is installed on the drive 810 as needed so that a computer program read therefrom is installed into the storage section 808 as needed.

[0092] In particular, according to the embodiments disclosed in the present invention, the process described in the main step diagram above can be implemented as a computer software program. For example, an embodiment of the present invention includes a computer program product, which includes a computer program carried on a computer-readable medium, and the computer program contains program code for executing the method shown in the main step diagram. In the above embodiment, the computer program can be downloaded and installed from the network through the communication part 809, and / or installed from the removable medium 811. When the computer program is executed by the central processing unit 801, the above functions defined in the system of the present invention are executed.

[0093] It should be noted that the computer-readable medium shown in the present invention may be a computer-readable signal medium or a computer-readable storage medium or any combination of the above two. The computer-readable storage medium may be, for example, but not limited to, an electrical, magnetic, optical, electromagnetic, infrared, or semiconductor system, device or device, or any combination of the above. More specific examples of computer-readable storage media may include, but are not limited to: an electrical connection with one or more wires, a portable computer disk, a hard disk, a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber, a portable compact disk read-only memory (CD-ROM), an optical storage device, a magnetic storage device, or any suitable combination of the above. In the present invention, a computer-readable storage medium may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, device or device. In the present invention, a computer-readable signal medium may include a data signal propagated in a baseband or as part of a carrier wave, which carries a computer-readable program code. This propagated data signal may take a variety of forms, including but not limited to electromagnetic signals, optical signals, or any suitable combination of the above. Computer readable signal media may also be any computer readable medium other than computer readable storage media, which may send, propagate or transmit a program for use by or in conjunction with an instruction execution system, apparatus or device. The program code contained on the computer readable medium may be transmitted using any appropriate medium, including but not limited to: wireless, wire, optical cable, RF, etc., or any suitable combination of the above.

[0094] The flow chart and block diagram in the accompanying drawings illustrate the possible architecture, function and operation of the system, method and computer program product according to various embodiments of the present invention. In this regard, each box in the flow chart or block diagram can represent a module, a program segment or a part of a code, and the above-mentioned module, program segment or a part of a code contains one or more executable instructions for realizing the specified logical function. It should also be noted that in some alternative implementations, the functions marked in the box can also occur in a different order from the order marked in the accompanying drawings. For example, two boxes represented in succession can actually be executed substantially in parallel, and they can sometimes be executed in the opposite order, depending on the functions involved. It should also be noted that each box in the block diagram or flow chart, and the combination of the boxes in the block diagram or flow chart can be implemented with a dedicated hardware-based system that performs a specified function or operation, or can be implemented with a combination of dedicated hardware and computer instructions.

[0095] The units involved in the embodiments of the present invention may be implemented by software or hardware. The units described may also be arranged in a processor. The names of these units do not, in some cases, limit the units themselves.

[0096] The above specific implementations do not constitute a limitation on the protection scope of the present invention. It should be understood by those skilled in the art that various modifications, combinations, sub-combinations and substitutions may occur depending on design requirements and other factors. Any modification, equivalent substitution and improvement made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. A power grid intelligent monitoring and dynamic dispatching system based on digital twins, characterized in that: include: A power grid topology generation module, used to obtain power grid geographic data and power equipment basic data, and generate power grid topology and power equipment full life cycle archives based on the power grid geographic data and the power equipment basic data; A modeling module, used to obtain power grid operation status data, and establish a digital twin model based on the power grid operation status data, the power grid topology and the full life cycle archive of the power equipment; A knowledge graph establishment module, used to establish a knowledge graph based on the power grid operation status data, the power grid topology and the power equipment life cycle archive; A mapping module, used to establish a mapping model according to the knowledge graph and the digital twin model, input the power grid operation status data into the mapping model, and obtain a mapping data set; A risk analysis module, used for performing multi-risk analysis based on the mapping data set to generate a comprehensive risk analysis report; The dispatching module is used to obtain a power grid dispatching strategy by adopting a multi-objective adaptive optimization algorithm and a load balancing algorithm according to the comprehensive risk analysis report.

2. According to the digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1, it is characterized in that: The power grid topology generation module specifically includes: A first data acquisition unit, used to acquire power grid geographic data and power equipment basic data; A power grid topology construction unit, configured to extract a power grid geographic map from the power grid geographic data, extract a device connection relationship from the power equipment basic data, and embed the device connection relationship into the power grid geographic map to obtain a power grid topology; The archive generation unit is used to perform a full life cycle analysis on the basic data of the power equipment to obtain a full life cycle archive of the power equipment.

3. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1 is characterized in that: The modeling module specifically includes: A second data acquisition unit, used to acquire power grid operation status data; An initial modeling unit, used for constructing an initial digital twin model according to the power grid topology and the full life cycle archive of the power equipment; A simulation unit, used for obtaining a number of simulation results by adopting a cross-scale multi-physics field coupling modeling algorithm according to the full life cycle archive of the power equipment and the power grid topology; The model optimization unit is used to optimize the initial digital twin model by using a deep reinforcement learning algorithm according to the simulation results and the power grid operation status data to obtain a digital twin model.

4. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1 is characterized in that: The knowledge graph establishment module specifically includes: A preprocessing unit, used for preprocessing the power grid operation status data to obtain a monitoring data set; A semantic association unit, configured to extract the topological features and life cycle features of the monitoring data set according to the power grid topology and the full life cycle archive of the power equipment, and generate a semantic constraint relationship according to the topological features and the life cycle features; An ontology unit, used for constructing a semantic ontology model by adopting an ontology semantic mapping method according to the semantic constraint relationship; A semantic reasoning unit, used for generating multi-dimensional semantic associations by using a relational deduction algorithm according to the semantic ontology model; A knowledge graph establishing unit is used to establish a knowledge graph based on the multi-dimensional semantic association.

5. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1 is characterized in that: The mapping module specifically includes: An association matching unit, used for extracting electric power equipment graph data from the knowledge graph, and performing identifier association matching between the electric power equipment graph data and the digital twin model to obtain an identifier association relationship; A mapping unit, used to perform data mapping processing on the knowledge graph and the digital twin model according to the identifier association relationship to obtain a mapping model; The data set output unit is used to input the power grid operation status data into the mapping model to obtain a mapping data set.

6. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1 is characterized in that: The risk analysis module specifically includes: A load forecasting unit, used to obtain several load forecasting results of the power grid by adopting a hybrid ensemble learning algorithm according to the mapping data set; An equipment failure analysis unit, configured to obtain an equipment health analysis report by using a Bayesian algorithm and a Markov transition probability model according to the load forecast results, the life cycle archive of the power equipment and the mapping data set; An extreme weather analysis unit, used to obtain extreme weather warning data, and obtain an extreme weather analysis report using a Monte Carlo algorithm based on the extreme weather warning data and the equipment health analysis report; The risk analysis unit is used to construct a multidimensional risk indicator system using the entropy weight method according to the several load forecast results of the power grid, the equipment health analysis report and the extreme weather analysis report, and generate a comprehensive risk analysis report according to the multidimensional risk indicator system.

7. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 1 is characterized in that: The scheduling module specifically includes: An initial strategy generating unit, configured to generate an initial power grid dispatching strategy by adopting a multi-objective adaptive optimization algorithm according to the mapping data set and the comprehensive risk analysis report; A resource configuration unit, configured to obtain a power resource configuration plan by using a load balancing algorithm according to the comprehensive risk analysis report and the initial power grid dispatching strategy; An emergency plan construction unit, used to identify the comprehensive risk analysis report, obtain potential risk scenarios, and construct a scenario emergency plan using a scenario matching algorithm according to the potential risk scenarios, a preset emergency response knowledge base and the power resource allocation plan; A strategy optimization unit is used to optimize the initial power grid dispatching strategy by adopting an adaptive closed-loop optimization algorithm according to the emergency plans for the several scenarios and the comprehensive risk analysis report to obtain a power grid dispatching strategy.

8. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 6 is characterized in that: The equipment failure analysis unit specifically includes: A first feature extraction subunit is used to extract features from the load forecast result, the power equipment life cycle archive and the mapping data set to obtain a number of key feature vectors; A Bayesian calculation subunit, used to calculate the conditional probabilities of different fault modes using a Bayesian algorithm according to the key feature vectors and the mapping data set; A Markov calculation subunit, used to establish a Markov transition probability model according to the conditional probabilities of the different failure modes, generate a device performance degradation curve according to the mapping data set and the Markov transition probability model, and obtain a device degradation inflection point and a device degradation critical state according to the device performance degradation curve; The equipment health analysis report generation subunit is used to process the conditional probabilities of the different failure modes, the degradation inflection points and the degradation critical states by adopting a multi-dimensional comprehensive evaluation method to obtain an equipment health analysis report.

9. The digital twin-based power grid intelligent monitoring and dynamic dispatching system according to claim 7 is characterized in that: The emergency plan building unit specifically includes: A scenario generation subunit, configured to construct a number of potential risk scenarios using a natural language processing algorithm according to the comprehensive risk analysis report; A second feature extraction subunit is used to extract features of the potential risk scenarios to obtain a number of risk semantic feature vectors; A similarity calculation subunit is used to calculate the similarity between each emergency plan set in the emergency disposal knowledge base and a plurality of risk semantic feature vectors; The emergency plan output subunit is used to select the emergency plan set with the greatest similarity as the pre-selected plan set, and select a scenario emergency plan from the pre-selected plan set according to the power resource configuration plan.

10. A method for intelligent monitoring and dynamic dispatching of power grid based on digital twin, characterized in that: The following steps are involved: S10: Acquire power grid geographic data and power equipment basic data, and generate power grid topology and power equipment full life cycle archives based on the power grid geographic data and the power equipment basic data; S20: Acquire power grid operation status data, and establish a digital twin model according to the power grid operation status data, the power grid topology, and the full life cycle archive of the power equipment; S30: Establishing a knowledge graph based on the power grid operation status data, the power grid topology and the power equipment life cycle archive; S40: establishing a mapping model according to the knowledge graph and the digital twin model, inputting the power grid operation status data into the mapping model, and obtaining a mapping data set; S50: Performing multi-risk analysis according to the mapping data set to generate a comprehensive risk analysis report; S60: Obtaining a power grid dispatching strategy using a multi-objective adaptive optimization algorithm and a load balancing algorithm according to the comprehensive risk analysis report.

Images