Multi-working-condition power supply method, device and equipment based on source load management and control and storage medium

Abstract

The invention is suitable for the technical field of electrical systems, and provides a multi-working-condition power supply method, device and equipment based on source load management and control, and a storage medium, a power distribution network is abstracted into a weighted directed graph and spatial-temporal feature extraction is carried out, so that accurate digital description of a power grid operation state is realized, and a fault point is predicted and fault diffusion is simulated, so that the power distribution network operation state is optimized. And the operation and maintenance mode is converted from passive coping to active defense. According to the embodiment of the invention, the node parameters are dynamically adjusted based on the real-time data, the flexibility, the adaptive capability and the fault processing efficiency of the system for coping with multiple working conditions are remarkably improved, and meanwhile, the global optimization configuration of limited resources is realized through the source-load cooperation and hierarchical supply insurance strategy, so that the power consumption is reduced on the premise of ensuring the power quality. The problems of low demand matching precision and poor reliability are effectively solved.

Description

Technical Field

[0001] This application belongs to the field of electrical system technology, and in particular relates to a multi-condition power supply method, device, equipment and storage medium based on source load control. Background Technology

[0002] With the deepening of energy transition, a large number of new power source devices such as distributed power sources, energy storage, and charging piles have been connected to the distribution network. These devices have significant intermittency, randomness, and uncertainty. Their single-phase and phase-separated distributed connection methods easily lead to serious inconsistencies in the spatiotemporal characteristics of power sources and loads in the distribution network, especially at the distribution substation level. This makes the inherent power quality problems in the substations, such as three-phase imbalance, heavy overload, light load, and voltage exceeding limits, increasingly prominent.

[0003] Currently, the operation and management of power distribution networks face enormous challenges. On the one hand, it is necessary to ensure the continuity and stability of power supply and maintain power quality within acceptable limits; on the other hand, it is also necessary to maximize the absorption of distributed clean energy. However, existing power distribution network management methods mostly rely on static models and local data, which suffer from problems such as low accuracy in demand matching, high energy loss, and insufficient system flexibility, reliability, and stability when dealing with complex and ever-changing operating conditions. Summary of the Invention

[0004] In view of this, embodiments of this application provide a multi-condition power supply method, apparatus, equipment and storage medium based on source and load management, which can solve the problems of low demand matching accuracy, high energy loss and insufficient system flexibility, reliability and stability in related technologies when dealing with complex and changing operating conditions.

[0005] In a first aspect, embodiments of this application provide a multi-condition power supply method based on source-load management, including: Collect multimodal data from heterogeneous data sources in the power distribution network; The distribution network is abstracted into a weighted directed graph based on multimodal data; Spatiotemporal features of the weighted directed graph are extracted to obtain the spatiotemporal features of the weighted directed graph. Based on spatiotemporal characteristics, predict potential fault points in the weighted directed graph; Fault propagation simulation was performed on potential fault points to obtain simulation results; Based on the simulation results, the parameters of the nodes in the distribution network corresponding to the weighted directed graph are adjusted so that power can be supplied through the adjusted distribution network.

[0006] Secondly, embodiments of this application provide a multi-condition power supply device based on source-load management, the device comprising: The acquisition module is used to acquire multimodal data from heterogeneous data sources in the power distribution network; The abstract module is used to abstract the distribution network into a weighted directed graph based on multimodal data; The extraction module is used to extract spatiotemporal features from the weighted directed graph to obtain the spatiotemporal features of the weighted directed graph; The prediction module is used to predict potential fault points in a weighted directed graph based on spatiotemporal characteristics. The simulation module is used to simulate the fault propagation of potential fault points and obtain simulation results. The adjustment module is used to adjust the parameters of the nodes in the distribution network corresponding to the weighted directed graph based on the simulation results, so as to provide power through the adjusted distribution network.

[0007] Thirdly, embodiments of this application provide a terminal device, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the computer program, it implements the steps of the above-described multi-condition power supply method based on source-load control.

[0008] Fourthly, embodiments of this application provide a computer-readable storage medium storing a computer program, which, when executed by a processor, implements the steps of the above-described multi-condition power supply method based on source-load control.

[0009] Fifthly, embodiments of this application provide a computer program product that, when run on a terminal device, causes the terminal device to execute the aforementioned multi-condition power supply method based on source-load management.

[0010] The beneficial effects of this application's embodiments compared to existing technologies are as follows: This application's embodiments, by abstracting the distribution network into a weighted directed graph and extracting spatiotemporal features, achieve a precise digital characterization of the power grid's operating state. Furthermore, by predicting fault points and simulating fault propagation, the operation and maintenance mode is transformed from passive response to proactive defense. This application's implementation dynamically adjusts node parameters based on real-time data, significantly improving the system's flexibility, adaptability, and fault handling efficiency in handling multiple operating conditions. Simultaneously, through source-load coordination and hierarchical supply guarantee strategies, it achieves global optimal allocation of limited resources, thereby effectively solving the problems of low demand matching accuracy and poor reliability while ensuring power quality. Attached Figure Description

[0011] To more clearly illustrate the technical solutions in the embodiments of this application, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of this application. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0012] Figure 1This is a schematic diagram illustrating the implementation process of the multi-condition power supply method based on source-load control provided in the embodiments of this application.

[0013] Figure 2 This is a schematic diagram of the structure of a multi-condition power supply device based on source-load management provided in the embodiments of this application.

[0014] Figure 3 This is a schematic diagram of the structure of the terminal device provided in the embodiments of this application. Detailed Implementation

[0015] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application. All other embodiments obtained by those skilled in the art based on the embodiments of this application without inventive effort are protected by this application.

[0016] It should be noted that the terms "comprising," "including," and "having," and any variations thereof, in the specification, claims, and accompanying drawings of this application, are intended to cover non-exclusive inclusion. For example, a process, method, terminal, product, or device that includes a series of steps or units is not limited to the listed steps or units, but may optionally include steps or units not listed, or may optionally include other steps or units inherent to these processes, methods, products, or devices. Terms such as "first" and "second" in the claims, specification, and accompanying drawings of this application, as well as relational terms, are used merely to distinguish one entity / operation / object from another entity / operation / object, and do not necessarily require or imply any such immediate relationship or order between these entities / operations / objects.

[0017] In this document, the term "embodiment" means that a particular feature, structure, or characteristic described in connection with an embodiment may be included in at least one embodiment of this application. The appearance of this phrase in various places throughout the specification does not necessarily refer to the same embodiment, nor is it a mutually exclusive, independent, or alternative embodiment. It will be explicitly and implicitly understood by those skilled in the art that the embodiments described herein can be combined with other embodiments.

[0018] With the deepening of energy transition, a large number of new power source devices such as distributed power sources, energy storage, and charging piles have been connected to the distribution network. These devices have significant intermittency, randomness, and uncertainty. Their single-phase and phase-separated distributed connection methods easily lead to serious inconsistencies in the spatiotemporal characteristics of power sources and loads in the distribution network, especially at the distribution substation level. This makes the inherent power quality problems in the substations, such as three-phase imbalance, heavy overload, light load, and voltage exceeding limits, increasingly prominent.

[0019] Currently, the operation and management of power distribution networks face enormous challenges. On the one hand, it is necessary to ensure the continuity and stability of power supply and maintain power quality within acceptable limits; on the other hand, it is also necessary to maximize the absorption of distributed clean energy. However, existing power distribution network management methods mostly rely on static models and local data, which suffer from problems such as low accuracy in demand matching, high energy loss, and insufficient system flexibility, reliability, and stability when dealing with complex and ever-changing operating conditions.

[0020] In view of this, this application provides a multi-condition power supply method based on source-load management. By abstracting the distribution network into a weighted directed graph and extracting its spatiotemporal features, it achieves a precise digital characterization of the power grid's operating status. Furthermore, by predicting fault points and simulating fault propagation, it transforms the operation and maintenance mode from passive response to proactive defense. This application's implementation dynamically adjusts node parameters based on real-time data, significantly improving the system's flexibility, adaptability, and fault handling efficiency in handling multiple operating conditions. Simultaneously, through source-load coordination and hierarchical supply guarantee strategies, it achieves global optimal allocation of limited resources, thereby effectively solving the problems of low demand matching accuracy and poor reliability while ensuring power quality.

[0021] To illustrate the technical solution of this application, specific embodiments are described below.

[0022] Figure 1 This illustration shows a schematic diagram of the implementation process of a multi-condition power supply method based on source-load management according to an embodiment of this application. This method can be applied to terminal devices. Terminal devices can be servers, service clusters, mobile phones, tablets, laptops, ultra-mobile personal computers (UMPCs), netbooks, etc.

[0023] Specifically, the above-mentioned multi-condition power supply method based on source and load management may include the following steps S101 to S106.

[0024] Step S101: Collect multimodal data from heterogeneous data sources in the power distribution network.

[0025] Heterogeneous data sources refer to data sets with different types, structures, and origins, such as equipment ledgers from relational databases, current and voltage readings from real-time sensors, coordinate data from geographic information systems (GIS), and data from meteorological API interfaces.

[0026] Multimodal data refers to data collected from the above heterogeneous data sources in various forms, such as structured data (database tables), unstructured data (text, images), time series data (sensor readings), and spatial data (map coordinates).

[0027] In the embodiments of this application, the terminal device can continuously or periodically collect multi-dimensional information covering the physical structure of the power grid, real-time operating status, environmental factors, etc., through sensing devices (such as smart meters, current and voltage transformers), monitoring and data acquisition (SCADA) systems, production management systems, and external data interfaces deployed in various parts of the power distribution network, thereby providing comprehensive and original data input for subsequent modeling and analysis.

[0028] Step S102: Abstract the distribution network into a weighted directed graph based on the multimodal data.

[0029] A weighted directed graph is a mathematical model consisting of "nodes" and "edges" with directions. Weighting refers to assigning numerical values ​​(weights) to nodes and edges. For example, node weights can represent their controllable power generation or energy storage capacity, while edge weights can reflect line impedance, real-time current carrying capacity, etc.

[0030] In the embodiments of this application, the terminal device can utilize multimodal data to map the physical entities of the power distribution network into elements of a graph model. Specifically, it abstracts power supply areas and load points as nodes and assigns weights based on their attributes (such as capacity and type), and abstracts transmission lines as directed edges and assigns weights based on their electrical parameters (such as impedance and current carrying capacity). This constructs a mathematical model that reflects both the network connectivity and quantifies its electrical characteristics. The graph model transforms the complex physical network into structured, computable data objects, providing a foundation for subsequent accurate calculations.

[0031] Step S103: Extract spatiotemporal features from the weighted directed graph to obtain the spatiotemporal features of the weighted directed graph.

[0032] Spatiotemporal features refer to features extracted from data that simultaneously contain information in both time and space dimensions, used to characterize the operating status and change patterns of the power grid at different locations and points in time.

[0033] In the embodiments of this application, the terminal device can clean, align and fuse the multimodal data (such as node voltage and line current) carried by the weighted directed graph that changes with time series, and use feature extraction algorithms (such as statistical analysis, signal processing or machine learning methods) to extract feature indicators (such as voltage deviation trend, load rate change, three-phase imbalance, etc.) from these data that can simultaneously reflect spatial location differences and time evolution laws.

[0034] Step S104: Based on the spatiotemporal characteristics, predict potential fault points in the weighted directed graph.

[0035] In the embodiments of this application, the terminal device can input spatiotemporal features into a pre-trained prediction model (such as a machine learning classifier or regression model). This model can learn the mapping relationship between spatiotemporal features and fault occurrence in historical data to assess the risk probability of each node or edge in the graph under the current state, and identify the parts with risk probabilities exceeding a preset threshold as potential fault points, thereby realizing early warning and location of faults and shifting the operation and maintenance mode from post-processing to pre-prevention.

[0036] Step S105: Perform a fault propagation simulation on the potential fault point to obtain the simulation results.

[0037] In the embodiments of this application, the terminal device can take the potential fault point as the initial fault, and based on the power grid topology and electrical relationships described by the weighted directed graph, use power system simulation algorithms (such as optimal power flow calculation and cascade failure model) to simulate the propagation path and impact range of the fault in the power grid after it occurs, and output simulation results such as system instability range, load loss, and key affected nodes to evaluate the impact of local faults on the overall power grid, and provide an impact assessment basis for formulating control strategies.

[0038] Step S106: Based on the simulation results, adjust the parameters of the nodes in the distribution network corresponding to the weighted directed graph so as to supply power through the adjusted distribution network.

[0039] In the embodiments of this application, the terminal device can generate specific control commands based on the weak links and risk levels of the power grid revealed by the simulation results, and then remotely or automatically operate the corresponding real equipment in the distribution network (such as adjusting the output of distributed power sources, controlling the charging and discharging of energy storage systems, and switching interconnection switches) through the actuator, thereby changing their operating parameters, eliminating potential risks, suppressing the spread of faults, or optimizing the operating status, and achieving the goal of ensuring safe and stable power supply by precisely intervening in the power grid operating parameters.

[0040] The beneficial effects of this application's embodiments compared to existing technologies are as follows: This application's embodiments, by abstracting the distribution network into a weighted directed graph and extracting spatiotemporal features, achieve a precise digital characterization of the power grid's operating state. Furthermore, by predicting fault points and simulating fault propagation, the operation and maintenance mode is transformed from passive response to proactive defense. This application's implementation dynamically adjusts node parameters based on real-time data, significantly improving the system's flexibility, adaptability, and fault handling efficiency in handling multiple operating conditions. Simultaneously, through source-load coordination and hierarchical supply guarantee strategies, it achieves global optimal allocation of limited resources, thereby effectively solving the problems of low demand matching accuracy and poor reliability while ensuring power quality.

[0041] In some specific embodiments of this application, the step of abstracting the distribution network into a weighted directed graph based on the multimodal data may specifically include steps S401 to S403.

[0042] Step S401: Define each power supply area and each power load in the power distribution network as a node of the weighted directed graph, define the transmission lines connecting the nodes as edges of the weighted directed graph, and determine the direction of the edges according to the direction of power flow, thereby forming an initial directed graph structure.

[0043] Among them, the power supply area refers to the power supply area covered by a distribution transformer in the distribution network. It is the basic unit of power distribution and is responsible for converting high-voltage electricity into low-voltage electricity and distributing it to users.

[0044] Electrical loads refer to electrical equipment or users in a power distribution network; they are the consumers of electrical energy.

[0045] In the embodiments of this application, the terminal device can first perform topological modeling of the physical power grid. The specific data processing process can be to abstract each functional entity in the distribution network (such as a transformer substation, a large charging station, or an industrial user) into a graph node, and to abstract the physical lines connecting these entities into edges. At the same time, based on the physical fact that electrical energy flows from the power source (power supply substation) to the load end (electrical load), the edges are assigned directions, thereby establishing a skeleton model that can accurately reflect the electrical connection relationship of the distribution network, laying the foundation for subsequent attribute assignment and quantitative analysis.

[0046] Step S402: Obtain the node attribute data corresponding to each node and the edge attribute data corresponding to each edge from the multimodal data.

[0047] Among them, node attribute data describes the characteristics of nodes (i.e., power supply areas and power loads) in the diagram, which may include physical attributes (such as geographical location and equipment capacity), electrical attributes (such as voltage level and current power) and operating status (such as online, offline, and fault).

[0048] Edge attribute data describes the characteristics of edges (i.e., transmission lines) in a graph. It can include physical attributes (such as line length and material), electrical attributes (such as resistance, reactance, and impedance), and real-time status (such as current current carrying capacity and temperature).

[0049] In the embodiments of this application, the terminal device can inject specific attribute information into the initial graph skeleton. The specific data processing process can be to query and extract the relevant attribute values ​​of each defined node and edge from multimodal data sources (such as obtaining real-time current and voltage from SCADA system, obtaining capacity parameters from equipment management database, and obtaining ambient temperature from meteorological data interface). For example, a transformer rated capacity and current output can be assigned to a certain transformer substation node, and an impedance value and current current can be assigned to a certain line edge. In this way, the abstract graph elements are associated with specific and quantifiable physical parameters, so that the model is upgraded from containing only connection relationships to an attribute graph containing rich semantic information.

[0050] Step S403: Based on the node attribute data, the adjustable resource capacity of the node is quantized into the weight of the node, and based on the edge attribute data, the line impedance and real-time current carrying capacity are quantized into the weight of the edge, thereby constructing the initial directed graph into the weighted directed graph.

[0051] Adjustable resource capacity refers to the ability of a node (especially a power supply area or distributed power source acting as a "source") to adjust its output or absorption power. For example, the adjustable resource capacity of an energy storage node is the range of power it can currently charge and discharge.

[0052] Line impedance refers to the obstruction of alternating current by a transmission line; it is the vector sum of resistance and reactance.

[0053] Real-time current carrying capacity refers to the actual current carried by a transmission line at a certain moment.

[0054] In the embodiments of this application, the terminal device can assign weights to key parameters of a graph model that already has attributes. The specific data processing process can be based on attribute data, selecting indicators that are crucial to power grid analysis and control for quantification. For example, the attribute value of "maximum discharge power" of an energy storage node can be directly used as its node weight, and the "impedance value" and "ratio of current current to rated current" of a line can be calculated (such as weighted summation) and used as the weight of that edge. This upgrades the graph model from "with attributes" to "with weights", so that the nodes and edges of the graph not only have descriptive information, but also have key numerical characteristics for subsequent mathematical calculations and optimization analysis.

[0055] Step S404: Based on the topological connectivity of the weighted directed graph, calculate at least one graph theory index among the degree, clustering coefficient, and betweenness centrality of each node, and calculate the importance index of each node based on the graph theory index.

[0056] Among them, the importance index is a comprehensive quantitative value used to assess the criticality of a node in the power grid.

[0057] In the embodiments of this application, the terminal device can perform topological analysis on the weighted directed graph and use graph theory algorithms to calculate the structural indicators of each node in the network. For example, it can calculate the number of lines directly connected to a node (degree), the density of the local network formed by its surrounding nodes (clustering coefficient), and the frequency of the node appearing in all shortest paths in the entire network (betweenness centrality). Then, it can combine these structural indicators into a comprehensive importance score through weighted summation or other fusion methods, and identify those key and influential hub nodes from the global perspective of network topology, providing a scientific basis for the subsequent formulation of differentiated and key supply guarantee strategies.

[0058] The implementation method of this application first establishes the topological skeleton of the power grid, then injects detailed physical parameters into it, and finally focuses on quantifying and assigning values ​​to the capacity, impedance and load rate indicators that are most critical for power flow calculation and state analysis. This makes the final weighted directed graph not only reflect the physical connection of the distribution network, but also deeply integrate its electrical characteristics and real-time status. This provides a high-quality and computable data foundation for subsequent precise mathematical analysis based on graph theory (such as feature extraction, fault prediction and power flow optimization), and greatly improves the accuracy and operability of the entire method model.

[0059] In some specific embodiments of this application, the step of extracting spatiotemporal features from the weighted directed graph to obtain the spatiotemporal features of the weighted directed graph may specifically include steps S501 to S503.

[0060] Step S501: Map the multimodal data to a unified spatiotemporal framework to generate a spatiotemporally aligned multimodal data sequence.

[0061] In this context, a unified spatiotemporal framework refers to a shared time reference (such as a UTC timestamp) and a spatial reference (such as a unified coordinate system, such as WGS-84 or a local grid coordinate system). All data can be transformed to this reference to ensure that data from different sources can be compared and correlated in time and space.

[0062] Spatiotemporal alignment refers to the process of processing data from different sources with different acquisition time points and spatial reference systems into data that can correspond one-to-one or be correlated in terms of time and spatial location through techniques such as time interpolation and coordinate transformation.

[0063] In the embodiments of this application, the terminal device can first establish a unified time axis (such as generating timestamps at a fixed frequency) and spatial reference system, and then use technologies such as data cleaning, invalid value processing, timestamp alignment, missing value interpolation, and spatial coordinate matching to unify SCADA real-time data, GIS location data, meteorological data, etc., onto this time scale and spatial node, thereby eliminating the spatiotemporal gap between data and generating a set of high-quality data sequences that are strictly aligned in both time and spatial dimensions and can be used for subsequent precise correlation analysis.

[0064] Step S502: Based on the spatiotemporally aligned multimodal data sequence, calculate the power quality index associated with each node at each moment to form the power quality feature sequence of each node.

[0065] Among them, power quality indicators are parameters that quantify the quality of power, and can include voltage deviation, frequency deviation, harmonic distortion rate, voltage sag, three-phase imbalance, etc., used to evaluate the stability and purity of power supply.

[0066] In the embodiments of this application, the terminal device can calculate a series of standard power quality index values ​​for the electrical quantity data (such as three-phase voltage and current waveforms) of each node at each alignment time point using standard power industry algorithms (such as calculating the voltage deviation by calculating the deviation between the effective voltage value and the rated value, and analyzing the harmonic content by fast Fourier transform to obtain the harmonic distortion rate, etc.). It can also arrange these index values ​​in time sequence, thereby transforming the low-level, raw monitoring data into high-level, characteristic trajectories with clear engineering physical significance, thereby capturing the subtle changes and dynamic processes of power quality during power grid operation.

[0067] Step S503: The power quality feature sequence is spliced ​​and fused with the attribute information and importance index of the node to generate the spatiotemporal features.

[0068] In the embodiments of this application, the terminal device can fuse three different dimensions of information: power quality feature sequences, node attribute data, and calculated importance indicators. The specific fusion method can be simple vector concatenation or a more complex attention-based weighted fusion, thereby breaking down data type barriers and organically combining information from three dimensions: the node's real-time operational health, its inherent capabilities, and its global influence in the network, to form a holographic, multi-faceted feature representation.

[0069] The implementation method of this application ensures the spatiotemporal consistency of multi-source data, and then extracts the core power quality dynamic sequence that directly reflects the health of the power grid operation. This dynamic sequence is then deeply integrated with the static attributes of nodes and their topological key indicators in the network to generate spatiotemporal features. This not only captures the real-time state changes of the power grid operation, but also incorporates the inherent capabilities of the equipment and its structural role in the global network, greatly improving the depth of state perception and the scientificity and accuracy of subsequent intelligent analysis model decisions.

[0070] In some specific embodiments of this application, the step of adjusting the parameters of the nodes in the distribution network corresponding to the weighted directed graph based on the simulation results, so as to supply power through the adjusted distribution network, may specifically include steps S601 to S603.

[0071] Step S601: Determine the supply priority sequence of each node based on the failure probability of each node in the simulation results and the importance index.

[0072] The power supply priority sequence refers to an ordered list obtained by sorting all nodes according to their failure risk and importance to the power grid. Nodes at the top of the sequence are those that need to be prioritized for power supply.

[0073] In the embodiments of this application, the terminal device can obtain the failure probability and importance index of each node output by the fault propagation simulation, and calculate a comprehensive supply urgency score through decision-making algorithms such as weighted fusion, fuzzy comprehensive evaluation or multi-objective optimization. Based on the supply urgency score, all nodes are sorted from high to low, thereby transforming the abstract simulation data and topology indicators into a clear and operable action priority order, ensuring that priorities can be distinguished under limited resources, and supply guarantee efforts are used in the most critical places.

[0074] Step S602: According to the supply priority sequence, the node is divided into the backbone layer, the branch layer and the terminal layer, and a supply instruction set is generated to be executed in descending order of level.

[0075] Among them, the backbone layer usually refers to the key nodes (such as hub substations) that connect to the high-voltage power grid and undertake the task of regional power transfer, and its failure has the widest impact range.

[0076] A branch layer refers to an intermediate node (such as a power distribution room or ring main unit) that connects the main layer and the peripheral layer, and is responsible for the power distribution in a certain area.

[0077] The peripheral layer refers to the nodes that are directly connected to the end users (such as low-voltage distribution areas), and the scope of influence is relatively localized.

[0078] In the embodiments of this application, the terminal device can classify nodes into different levels such as trunk, branch, and terminal based on the priority sequence of power supply and the actual topology of the power grid. Then, it can design typical power supply measures for each level (such as configuring mobile emergency power vehicles for the trunk level and enabling distributed energy storage for the branch level) and generate specific sets of instructions to be executed in hierarchical order. This transforms the linear priority list into a structured, batch-executed operational plan, ensuring that the power supply operation can be carried out systematically from the global to the local level, avoiding operational chaos.

[0079] Step S603: According to the power supply instruction set, adjust the operating parameters of the corresponding nodes in the distribution network to supply power through the adjusted distribution network.

[0080] Among them, operating parameters refer to the adjustable settings of controllable equipment in the power grid, such as generator output, energy storage system charging and discharging power, capacitor bank switching status, and transformer tap position. In the embodiments of this application, the terminal equipment can automatically receive and parse the power supply instruction set, and change the operating parameters of the controllable equipment on the corresponding node through remote control or on-site operation. For example, it can start the backup generator to increase power, control the energy storage system to discharge to support voltage, switch the tie switch to change the power supply path, etc. By adjusting the parameters precisely, the power flow distribution and operating status of the power grid can be changed, thereby suppressing the spread of faults or restoring power supply, and achieving the ultimate goal of safe and reliable power supply through the optimized power grid.

[0081] The implementation method of this application transforms the analysis results of fault probability and node importance obtained in the aforementioned steps into a clear, orderly, and executable hierarchical power supply guarantee scheme. This ensures that the response measures can be accurately focused on high-risk and high-critical nodes. Then, through the hierarchical strategy, the systemic optimization of power supply guarantee resources from the global to the local and from the backbone to the end is achieved. Finally, precise adjustment is achieved through automated control, thereby greatly improving the efficiency and reliability of power supply restoration in the event of a fault, avoiding resource waste and blind operation, effectively solving the problems of low power supply flexibility and poor reliability mentioned in the background technology, and maximizing the utilization of limited power supply guarantee resources.

[0082] In some specific embodiments of this application, the step of predicting potential fault points in the weighted directed graph based on the spatiotemporal characteristics may specifically include steps S701 and S702.

[0083] Step S701: Input the spatiotemporal features into the trained supply guarantee strategy prediction model to obtain the predicted label data for each node.

[0084] The predicted label data consists of one or more quantitative indicators used to characterize the future stability of the node. For example, it can be a numerical value representing the probability of failure (such as between 0 and 1) or a label representing the stability level.

[0085] In the embodiments of this application, the terminal device can batch-feed spatiotemporal feature vectors as input data into a pre-trained power supply guarantee strategy prediction model. This model can internally calculate the input features using complex nonlinear functions it has learned, and ultimately output one or more quantified predicted values ​​for each node in the weighted directed graph. By leveraging a data-driven model to surpass traditional threshold judgments, a forward-looking and refined quantitative assessment of node operational stability is achieved, transforming the abstract power grid state into concrete and comparable numerical indicators.

[0086] Step S702: Nodes whose stability level represented by the predicted label data is lower than a preset threshold are identified as potential fault points.

[0087] The preset threshold is a pre-defined critical value used to determine whether the stability of a node is low enough to be identified as a potential failure point.

[0088] In the embodiments of this application, the terminal device can acquire the predicted label data (e.g., failure probability values) generated by all nodes, compare it with a preset threshold, filter out all nodes whose stability represented by the predicted label data is lower than the threshold (e.g., nodes with failure probability higher than 0.7), and formally mark them as potential failure points, thereby transforming continuous predicted probability values ​​into a clear and operable list of failure risks, providing a clear initial target set for subsequent failure propagation simulation.

[0089] The implementation method of this application first uses a trained prediction model to perform in-depth analysis of spatiotemporal features containing rich information, and outputs a quantitative assessment of the stability of each node, thereby transforming the rough judgment based on experience into a precise prediction based on data. Then, through the key step of setting a preset threshold, the prediction results are transformed into a clear and actionable list of potential fault points, enabling the operation and maintenance strategy to shift from "passively responding to faults that have occurred" to "actively targeting high-risk nodes", which greatly improves the accuracy and foresight of fault early warning, thereby enhancing the overall proactive defense capability and power supply reliability of the power grid.

[0090] In some specific embodiments of this application, before predicting potential fault points in the weighted directed graph based on the spatiotemporal characteristics, the above method may further include steps S801 to S804.

[0091] Step S801: Obtain historical operating data for each node, and generate real label data for quantifying the stability of the nodes based on the historical operating data.

[0092] The real-name data may include at least one of the following: number of equipment outages, historical power outage frequency, average power outage duration, average power outage duration for users, and average system power availability.

[0093] Historical operational data refers to data recorded from power grid monitoring systems (such as SCADA and production management systems) that reflects the past operational status of nodes, such as power outage records, equipment operation and maintenance logs, and user complaint data.

[0094] In the embodiments of this application, the terminal device can extract long-term series operation records related to each node from the power grid historical database. Then, it performs statistical calculations on these raw records according to predefined rules and formulas to generate a series of quantitative indicators that can directly characterize the historical stability of nodes as real label data. For example, it can count the number of outages of a transformer in the past year or calculate the average power availability of a power supply area. This provides reliable supervision signals for model training, enabling the model to learn the potential mapping relationship between spatiotemporal characteristics and real stability. In summary, this step is the data preparation stage for model training, extracting standard answers from historical data to provide targets for subsequent learning.

[0095] Step S802: Construct an initial prediction model that takes the spatiotemporal features of the nodes as input and the label data of the nodes as the prediction output.

[0096] In the embodiments of this application, the terminal device can select a suitable machine learning algorithm (such as a deep learning model) and determine that the dimension of its input layer matches the dimension of the spatiotemporal features and the dimension of its output layer matches the dimension of the real label data, thereby constructing a model framework that has not yet started learning to handle the complex nonlinear relationship between input and output.

[0097] Step S803: Input the spatiotemporal features into the initial prediction model to obtain the predicted label data for each node.

[0098] In the embodiments of this application, the terminal device can input a batch of samples with existing spatiotemporal features and corresponding real label data into the initial prediction model. The initial prediction model can calculate and generate a set of prediction outputs for each input sample according to its current random or initialized internal parameters, thereby obtaining the prediction results of the model under the current parameters in the training iteration, so as to compare with the real labels and calculate the error.

[0099] Step S804: Calculate the model loss value based on the node importance index, the predicted label data, and the real label data, and update the model parameters through the backpropagation algorithm until the model converges, thus obtaining the trained supply guarantee strategy prediction model.

[0100] In the embodiments of this application, the terminal device can first design a loss function that not only calculates the difference between the predicted label and the true label, but also introduces the importance index of the node as a weight, so that the model can pay more attention to the accuracy of predicting important nodes during training. Then, the gradient of the loss function with respect to each parameter of the model is calculated through the backpropagation algorithm, and the optimizer (e.g., Adam) is used to update the parameters according to the gradient direction. This process is repeated until the model performance tends to stabilize, thereby driving the continuous adjustment of the model parameters, and finally enabling the model to learn the ability to accurately predict the stability of nodes from complex spatiotemporal features.

[0101] In some specific embodiments of this application, the step of performing fault propagation simulation on the potential fault point to obtain simulation results may specifically include steps S901 to S904.

[0102] Step S901: Calculate the betweenness centrality of each edge in the weighted directed graph, and identify vulnerable branches in the network based on the betweenness centrality.

[0103] Betweenness centrality is a centrality metric in graph theory that measures how often an edge appears in all shortest paths throughout the entire network.

[0104] Vulnerable branches are transmission lines that play an important role in the network (i.e., have high betweenness centrality) and whose failure would have a significant impact on network connectivity.

[0105] In the embodiments of this application, the terminal device can apply graph theory algorithms (such as Brandes' algorithm) to traverse the weighted directed graph, calculate the betweenness centrality value of each edge (i.e., transmission line) in the graph, and then identify the edges with the highest betweenness centrality values ​​as vulnerable branches according to a preset threshold or sorting. The embodiments of this application can, from the perspective of the global network topology, pre-identify critical lines that are highly likely to trigger large-scale power flow shifts and chain reactions once a fault occurs, providing key propagation paths for subsequent fault simulation.

[0106] Step S902: Set the potential fault point as the initial fault point, simulate N-1 fault conditions, and generate the initial fault scenario.

[0107] The N-1 fault condition is a standard criterion in power system security analysis, which means that after any independent component in the system (such as a generator, a line, or a transformer) fails and is disconnected, the system should be able to maintain stable operation without causing overload of other components or system collapse.

[0108] In the embodiments of this application, the terminal device can set the potential fault point as the initial fault and remove the element from the network in the weighted directed graph model (simulating its exit from operation due to fault), thereby creating an initial fault scenario that conforms to the N-1 safety verification criterion. This provides a starting state that conforms to engineering reality and has analytical significance for subsequent dynamic simulation, transforming the abstract "potential risk" into a specific and simulable fault event.

[0109] Step S903: Combine the initial fault scenario with the vulnerable branch, and simulate the process of fault propagation along the topology of the weighted directed graph using a cascaded failure model to obtain the initial simulation results.

[0110] In the embodiments of this application, the terminal device can start from the initial fault scenario and run a cascaded failure model. This cascaded failure model can be based on the power flow calculation principle of the power grid to simulate the power redistribution caused by the exit of the initial component, and pay special attention to whether the vulnerable branch exceeds its capacity limit due to the large power flow shift. Once overloaded, the protection device can be simulated to cut it off, and then a new round of power flow shift can be simulated until the system is stable and no new overload occurs. In this way, the entire process of the fault from the point of occurrence to the spread and stabilization can be dynamically and realistically reproduced, and the detailed evolution path can be recorded.

[0111] Step S904: Based on the initial simulation results, calculate the probability of each node failing during the fault propagation process, and generate simulation results containing the failure probabilities of each node.

[0112] In the embodiments of this application, the terminal device can analyze the initial simulation results (i.e., a detailed record of fault propagation), and statistically determine whether each node ultimately loses power during the simulation process. If multiple simulations are performed (e.g., considering different load levels or fault points), the frequency of power loss for each node can be calculated as its fault probability. If it is a deterministic simulation, the probability of a node that ultimately loses power can be marked as 1, and that of a node that does not lose power can be marked as 0, or a probability value can be assigned based on the degree of impact. This condenses the detailed dynamic simulation record into a concise, quantified risk distribution map, intuitively showing the probability of each node being affected by the initial fault.

[0113] The implementation method of this application first identifies key vulnerable branches from the perspective of network topology, providing a focus for simulation. Then, it generates an initial scenario that conforms to engineering reality by simulating N-1 fault conditions. Next, it uses a cascading failure model to dynamically deduce the chain reaction of the fault, which can realistically reflect the propagation path and impact range of the fault. Finally, it quantifies the complex simulation process into the fault probability of each node, making the output results clear and intuitive. This greatly improves the depth of power grid fault prediction analysis and the pertinence of decision support, effectively avoiding the blindness of relying solely on experience.

[0114] Figure 3 This illustration shows a structural diagram of a multi-condition power supply device based on source-load management according to an embodiment of this application. The aforementioned multi-condition power supply device 2 based on source-load management can be configured on a terminal device. Specifically, the aforementioned multi-condition power supply device 2 based on source-load management may include: Acquisition module 201 is used to acquire multimodal data from heterogeneous data sources in the power distribution network; Abstraction module 202 is used to abstract the distribution network into a weighted directed graph based on the multimodal data; Extraction module 203 is used to extract spatiotemporal features from the weighted directed graph to obtain the spatiotemporal features of the weighted directed graph; Prediction module 204 is used to predict potential fault points in the weighted directed graph based on the spatiotemporal characteristics. Simulation module 205 is used to simulate the fault propagation of the potential fault points and obtain simulation results; The adjustment module 206 is used to adjust the parameters of the nodes in the distribution network corresponding to the weighted directed graph according to the simulation results, so as to supply power through the adjusted distribution network.

[0115] The beneficial effects of this application's embodiments compared to existing technologies are as follows: This application's embodiments, by abstracting the distribution network into a weighted directed graph and extracting spatiotemporal features, achieve a precise digital characterization of the power grid's operating state. Furthermore, by predicting fault points and simulating fault propagation, the operation and maintenance mode is transformed from passive response to proactive defense. This application's implementation dynamically adjusts node parameters based on real-time data, significantly improving the system's flexibility, adaptability, and fault handling efficiency in handling multiple operating conditions. Simultaneously, through source-load coordination and hierarchical supply guarantee strategies, it achieves global optimal allocation of limited resources, thereby effectively solving the problems of low demand matching accuracy and poor reliability while ensuring power quality.

[0116] In some embodiments of this application, the abstract module 202 is further used for: Each power distribution area and each power load in the power distribution network is defined as a node of the weighted directed graph, and the transmission lines connecting the nodes are defined as edges of the weighted directed graph. The direction of the edges is determined according to the direction of power flow, thereby forming an initial directed graph structure. Obtain the node attribute data corresponding to each node and the edge attribute data corresponding to each edge from the multimodal data; Based on the node attribute data, the adjustable resource capacity of the node is quantified into the weight of the node, and based on the edge attribute data, the line impedance and real-time current carrying capacity are quantified into the weight of the edge, thereby constructing the initial directed graph into the weighted directed graph. Based on the topological connectivity of the weighted directed graph, calculate at least one graph theory metric among degree, clustering coefficient, and betweenness centrality for each node, and calculate the importance metric for each node based on the graph theory metric.

[0117] In some embodiments of this application, the extraction module 203 is further configured to: The multimodal data is mapped to a unified spatiotemporal framework to generate a spatiotemporally aligned multimodal data sequence. Based on the spatiotemporally aligned multimodal data sequence, power quality indicators associated with each node at each time step are calculated to form a power quality feature sequence for each node. The power quality feature sequence is spliced ​​and fused with the node attribute data and the importance index to generate the spatiotemporal features.

[0118] In some embodiments of this application, the adjustment module 206 is further configured to: Based on the failure probability of each node in the simulation results and the importance index, the supply priority sequence of each node is determined. Based on the supply priority sequence, the nodes are divided into the backbone layer, the branch layer and the terminal layer, and a supply instruction set is generated to be executed in descending order of level. According to the power supply instruction set, the operating parameters of the corresponding nodes in the distribution network are adjusted so as to supply power through the adjusted distribution network.

[0119] In some embodiments of this application, the prediction module 204 is further configured to: The spatiotemporal features are input into the trained supply guarantee strategy prediction model to obtain the predicted label data for each node; Nodes whose stability level represented by the predicted label data is lower than a preset threshold are identified as potential fault points.

[0120] In some embodiments of this application, the above-mentioned multi-condition power supply device 2 based on source-load control further includes a training module for: Historical operating data of each node is obtained, and real label data for quantifying the stability of the node is generated based on the historical operating data. The real label data includes at least one of the following: number of equipment outages, historical power outage frequency, average power outage duration, average power outage duration for users, and average power supply availability of the system. Construct an initial prediction model that takes the spatiotemporal features of nodes as input and the label data of nodes as prediction output; The spatiotemporal features are input into the initial prediction model to obtain the predicted label data for each node; The model loss value is calculated based on the node importance index, the predicted label data, and the real label data. The model parameters are then updated using the backpropagation algorithm until the model converges, resulting in a trained supply guarantee strategy prediction model.

[0121] In some embodiments of this application, the simulation module 205 is further configured to: Calculate the betweenness centrality of each edge in the weighted directed graph, and identify vulnerable branches in the network based on the betweenness centrality; The potential fault points are set as the initial fault points, and N-1 fault conditions are simulated to generate the initial fault scenario. By combining the initial fault scenario with the vulnerable branch, the process of fault propagation along the topology of the weighted directed graph is simulated using a cascaded failure model to obtain initial simulation results. Based on the initial simulation results, the probability of each node failing during the fault propagation process is calculated, and simulation results containing the failure probabilities of each node are generated.

[0122] like Figure 3 The diagram shown is a schematic of a terminal device provided in an embodiment of this application. The terminal device 3 may include: a processor 301, a memory 302, and a computer program 303 stored in the memory 302 and executable on the processor 301, such as a multi-condition power supply program based on source-load management. When the processor 301 executes the computer program 303, it implements the steps in the various multi-condition power supply embodiments based on source-load management described above, for example... Figure 1 Steps S101 to S106 are shown.

[0123] A computer program can be divided into one or more modules / units. One or more modules / units are stored in memory 302 and executed by processor 301 to complete this application. One or more modules / units can be a series of computer program instruction segments capable of performing a specific function, which describe the execution process of the computer program in a terminal device.

[0124] The terminal device may include, but is not limited to, processor 301 and memory 302. Those skilled in the art will understand that... Figure 3 This is merely an example of a terminal device and does not constitute a limitation on the terminal device. It may include more or fewer components than shown in the figure, or combine certain components, or different components. For example, a terminal device may also include input / output devices, network access devices, buses, etc.

[0125] The processor 301 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. A general-purpose processor may be a microprocessor or any conventional processor.

[0126] The memory 302 can be an internal storage unit of the terminal device, such as the hard drive or RAM of the terminal device. The memory 302 can also be an external storage device of the terminal device, such as a plug-in hard drive, Smart Media Card (SMC), Secure Digital (SD) card, or Flash Card. Furthermore, the memory 302 can include both internal and external storage units of the terminal device. The memory 302 is used to store computer programs and other programs and data required by the terminal device. The memory 302 can also be used to temporarily store data that has been output or will be output.

[0127] It should be noted that, for the sake of convenience and brevity, the structure of the terminal device described above can also be referred to the specific description of the structure in the method embodiment, which will not be repeated here.

[0128] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the above-described division of functional units and modules is merely an example. In practical applications, the above functions can be assigned to 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 embodiments can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit. Furthermore, the specific names of the functional units and modules are only for easy differentiation and are not intended to limit the scope of protection of this application. The specific working process of the units and modules in the above system can be referred to the corresponding process in the foregoing method embodiments, and will not be repeated here.

[0129] This application also provides a computer-readable storage medium storing a computer program, which, when executed by a processor, can implement the steps in the above-described multi-condition power supply method based on source-load control.

[0130] This application provides a computer program product that, when run on a mobile terminal, enables the mobile terminal to implement the steps in the above-mentioned multi-condition power supply method based on source-load control.

[0131] In the above embodiments, the descriptions of each embodiment have different focuses. For parts that are not described in detail or recorded in a certain embodiment, please refer to the relevant descriptions of other embodiments.

[0132] Those skilled in the art will recognize that the units and algorithm steps of the various examples described in conjunction with the embodiments disclosed herein can be implemented in electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for various specific applications, but such implementations should not be considered beyond the scope of this application.

[0133] In the embodiments provided in this application, it should be understood that the disclosed apparatus / terminal devices and methods can be implemented in other ways. For example, the apparatus / terminal device embodiments described above are merely illustrative. For instance, the division of modules or units is only a logical functional division, and in actual implementation, there may be other division methods. For example, multiple units or components may be combined or integrated into another system, or some features may be ignored or not executed. Furthermore, the coupling or direct coupling or communication connection shown or discussed may be through some interfaces; the indirect coupling or communication connection between devices or units may be electrical, mechanical, or other forms.

[0134] The units described as separate components may or may not be physically separate. The components shown as units may or may not be physical units; that is, they may be located in one place or distributed across multiple network units. Some or all of the units can be selected to achieve the purpose of this embodiment according to actual needs.

[0135] Furthermore, the functional units in the various embodiments of this application can be integrated into one processing unit, or each unit can exist physically separately, or two or more units can be integrated into one unit. The integrated unit can be implemented in hardware or as a software functional unit.

[0136] If the integrated module / unit is implemented as 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, all or part of the processes in the methods of the above embodiments can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying the computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, read-only memory (ROM), random access memory (RAM), electrical carrier signals, telecommunication signals, and software distribution media, etc. It should be noted that the content included in the computer-readable medium can be appropriately added or removed 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 do not include electrical carrier signals and telecommunication signals.

[0137] The embodiments described above are merely illustrative of the technical solutions of this application and are not intended to limit it. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. These modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A multi-condition power supply method based on source-load management, characterized in that, include: Collect multimodal data from heterogeneous data sources in the power distribution network; The distribution network is abstracted into a weighted directed graph based on the multimodal data; Spatiotemporal features of the weighted directed graph are extracted to obtain the spatiotemporal features of the weighted directed graph. Based on the spatiotemporal characteristics, predict potential fault points in the weighted directed graph; A fault propagation simulation was performed on the potential fault points to obtain simulation results; Based on the simulation results, the parameters of the nodes in the distribution network corresponding to the weighted directed graph are adjusted so that power can be supplied through the adjusted distribution network.

2. The multi-condition power supply method based on source-load management as described in claim 1, characterized in that, The step of abstracting the distribution network into a weighted directed graph based on the multimodal data includes: Each power distribution area and each power load in the power distribution network is defined as a node of the weighted directed graph, and the transmission lines connecting the nodes are defined as edges of the weighted directed graph. The direction of the edges is determined according to the direction of power flow, thereby forming an initial directed graph structure. Obtain the node attribute data corresponding to each node and the edge attribute data corresponding to each edge from the multimodal data; Based on the node attribute data, the adjustable resource capacity of the node is quantified into the weight of the node, and based on the edge attribute data, the line impedance and real-time current carrying capacity are quantified into the weight of the edge, thereby constructing the initial directed graph into the weighted directed graph. Based on the topological connectivity of the weighted directed graph, calculate at least one graph theory metric among degree, clustering coefficient, and betweenness centrality for each node, and calculate the importance metric for each node based on the graph theory metric.

3. The multi-condition power supply method based on source-load management as described in claim 2, characterized in that, The process of extracting spatiotemporal features from the weighted directed graph to obtain its spatiotemporal features includes: The multimodal data is mapped to a unified spatiotemporal framework to generate a spatiotemporally aligned multimodal data sequence. Based on the spatiotemporally aligned multimodal data sequence, power quality indicators associated with each node at each time step are calculated to form a power quality feature sequence for each node. The power quality feature sequence is spliced ​​and fused with the node attribute data and the importance index to generate the spatiotemporal features.

4. The multi-condition power supply method based on source-load management as described in claim 2, characterized in that, The step of adjusting the parameters of nodes in the distribution network corresponding to the weighted directed graph based on the simulation results, so as to provide power through the adjusted distribution network, includes: Based on the failure probability of each node in the simulation results and the importance index, the supply priority sequence of each node is determined. Based on the supply priority sequence, the nodes are divided into the backbone layer, the branch layer and the terminal layer, and a supply instruction set is generated to be executed in descending order of level. According to the power supply instruction set, the operating parameters of the corresponding nodes in the distribution network are adjusted so as to supply power through the adjusted distribution network.

5. The multi-condition power supply method based on source-load management as described in claim 1, characterized in that, The step of predicting potential fault points in the weighted directed graph based on the spatiotemporal characteristics includes: The spatiotemporal features are input into the trained supply guarantee strategy prediction model to obtain the predicted label data for each node; Nodes whose stability level represented by the predicted label data is lower than a preset threshold are identified as potential fault points.

6. The multi-condition power supply method based on source-load management as described in claim 5, characterized in that, Before predicting potential fault points in the weighted directed graph based on the spatiotemporal characteristics, the process includes: Historical operating data of each node is obtained, and real label data for quantifying the stability of the node is generated based on the historical operating data. The real label data includes at least one of the following: number of equipment outages, historical power outage frequency, average power outage duration, average power outage duration for users, and average power supply availability of the system. Construct an initial prediction model that takes the spatiotemporal features of nodes as input and the label data of nodes as prediction output; The spatiotemporal features are input into the initial prediction model to obtain the predicted label data for each node; The model loss value is calculated based on the node importance index, the predicted label data, and the real label data. The model parameters are then updated using the backpropagation algorithm until the model converges, resulting in a trained supply guarantee strategy prediction model.

7. The multi-condition power supply method based on source-load management as described in claim 1, characterized in that, The simulation of fault propagation at the potential fault points, and the resulting simulation results, include: Calculate the betweenness centrality of each edge in the weighted directed graph, and identify vulnerable branches in the network based on the betweenness centrality; The potential fault points are set as the initial fault points, and N-1 fault conditions are simulated to generate the initial fault scenario. By combining the initial fault scenario with the vulnerable branch, the process of fault propagation along the topology of the weighted directed graph is simulated using a cascaded failure model to obtain initial simulation results. Based on the initial simulation results, the probability of each node failing during the fault propagation process is calculated, and simulation results containing the failure probabilities of each node are generated.

8. A multi-condition power supply device based on source-load management, characterized in that, The device includes: The acquisition module is used to acquire multimodal data from heterogeneous data sources in the power distribution network; An abstraction module is used to abstract the distribution network into a weighted directed graph based on the multimodal data; The extraction module is used to extract spatiotemporal features from the weighted directed graph to obtain the spatiotemporal features of the weighted directed graph; The prediction module is used to predict potential fault points in the weighted directed graph based on the spatiotemporal characteristics. The simulation module is used to simulate the fault propagation of the potential fault points and obtain simulation results; The adjustment module is used to adjust the parameters of the nodes in the distribution network corresponding to the weighted directed graph based on the simulation results, so as to supply power through the adjusted distribution network.

9. A terminal device, comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor executes the computer program to implement the steps of the multi-condition power supply method based on source-load management as described in any one of claims 1 to 7.

10. A computer-readable storage medium storing a computer program, characterized in that, When the computer program is executed by the processor, it implements the steps of the multi-condition power supply method based on source-load management as described in any one of claims 1 to 7.

Images