Panoramic situation awareness-based power grid fault prediction processing method and system

Abstract

The invention discloses a power grid fault prediction processing method and system based on panoramic situation awareness, and belongs to the field of power grid fault prediction, and the method comprises the steps: carrying out the data collection of a target power grid through a panoramic situation awareness system, and obtaining a multi-source perception data set; building a power grid space topological structure; clustering each node in the power grid spatial topological structure, and outputting a multi-layer topological network; constructing an intra-layer fault propagation model and a cross-layer fault transfer model; connecting a panoramic situation awareness system, and acquiring real-time power grid state data of the target power grid; and performing fault prediction on the real-time power grid state data, and outputting a plurality of pieces of fault early warning information. According to the method and the device, the technical problem of insufficient fault early warning accuracy and timeliness caused by lack of multi-level dynamic analysis and cross-layer influence consideration in power grid fault prediction in the prior art is solved, and the technical effect of improving the accuracy and timeliness of power grid fault prediction through intra-layer fault propagation and cross-layer fault transfer of a multi-layer topology network is achieved.

Description

Technical Field

[0001] The present invention relates to the field of power grid fault prediction, and in particular to a power grid fault prediction processing method and system based on panoramic situation awareness. Background Art

[0002] Existing power grid fault prediction methods are mainly based on single-level data analysis and simple network models, usually only considering local network structures and static fault modes, and it is difficult to fully reflect the multi-level structural characteristics and dynamic fault propagation process of the power grid. For example, existing power grid fault prediction methods focus on faults at the distribution network level, but ignore the mutual influence between the power grid and the transmission network; or although the network topology is considered, it fails to effectively simulate the fault transfer mechanism between different levels. Therefore, the lack of multi-level dynamic analysis and cross-layer impact considerations leads to insufficient accuracy and timeliness of power grid fault warnings. Summary of the invention

[0003] This application provides a power grid fault prediction and processing method and system based on panoramic situational awareness, aiming to solve the technical problem that power grid fault prediction in the prior art lacks multi-level dynamic analysis and cross-layer impact consideration, resulting in insufficient accuracy and timeliness of fault warning.

[0004] In view of the above problems, the present application provides a power grid fault prediction and processing method and system based on panoramic situational awareness.

[0005] The first aspect disclosed in the present application provides a power grid fault prediction and processing method based on panoramic situational awareness, the method comprising: collecting data on the target power grid through a panoramic situational awareness system to obtain a multi-source perception data set; building a power grid spatial topology structure based on the multi-source perception data set; clustering each node in the power grid spatial topology structure to output a multi-layer topology network; constructing an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network, and a cross-layer fault transfer model between layers in the multi-layer topology network; connecting the panoramic situational awareness system to obtain real-time power grid status data of the target power grid; performing fault prediction on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and outputting multiple fault warning information.

[0006] Another aspect disclosed in the present application provides a power grid fault prediction and processing system based on panoramic situational awareness, the system comprising: a data acquisition module, used to collect data on the target power grid through the panoramic situational awareness system to obtain a multi-source perception data set; a topology construction module, used to build a power grid space topology structure according to the multi-source perception data set; a multi-layer clustering module, used to cluster each node in the power grid space topology structure, and output a multi-layer topology network; a model construction module, used to construct an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network, and a cross-layer fault transfer model between layers in the multi-layer topology network; a real-time data acquisition module, used to connect to the panoramic situational awareness system to obtain real-time power grid status data of the target power grid; a fault prediction module, used to perform fault prediction on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and output multiple fault warning information.

[0007] One or more technical solutions provided in this application have at least the following technical effects or advantages:

[0008] By using a panoramic situational awareness system to collect data from the target power grid and obtain multi-source perception data sets, we can achieve a comprehensive perception of the power grid's operating status, providing a rich data foundation for subsequent analysis; based on the multi-source perception data sets, we build a spatial topological structure of the power grid, laying the foundation for subsequent multi-level analysis; cluster each node in the spatial topological structure of the power grid, output a multi-layer topological network, and achieve a multi-level division of the power grid structure, reflecting the hierarchical characteristics of the power grid, creating conditions for more detailed fault analysis; construct an intra-layer fault propagation model corresponding to each layer of the multi-layer topological network, as well as a cross-layer fault transfer model between layers in the multi-layer topological network, to simulate the fault propagation within a single layer, while considering different layers Inter-level fault transfer comprehensively describes the dynamic propagation characteristics of faults; connects to the panoramic situation awareness system to obtain real-time grid status data of the target power grid, so that fault prediction can be based on the latest grid operation status; based on the intra-layer fault propagation model and cross-layer fault transfer model, fault prediction is performed on real-time grid status data, and multiple fault warning information is output, realizing a technical solution for accurate and timely fault prediction and warning, solving the technical problem that the power grid fault prediction in the existing technology lacks multi-level dynamic analysis and cross-layer impact considerations, resulting in insufficient accuracy and timeliness of fault warning, and achieving the technical effect of improving the accuracy and timeliness of power grid fault prediction through intra-layer fault propagation and cross-layer fault transfer of multi-layer topology networks.

[0009] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0010] Figure 1 A flowchart of a method for predicting and processing power grid faults based on panoramic situational awareness is provided for an embodiment of the present application;

[0011] Figure 2 A structural schematic diagram of a power grid fault prediction and processing system based on panoramic situational awareness is provided for an embodiment of the present application.

[0012] Explanation of the reference numerals: data collection module 11 , topology construction module 12 , multi-layer clustering module 13 , model construction module 14 , real-time data acquisition module 15 , fault prediction module 16 . DETAILED DESCRIPTION

[0013] The overall idea of ​​the technical solution provided by this application is as follows:

[0014] The embodiments of the present application provide a method and system for predicting and processing power grid faults based on panoramic situational awareness. First, a panoramic situational awareness system is used to collect multi-source data to construct the spatial topological structure of the power grid. Subsequently, the power grid structure is divided into multiple levels to form a multi-layer topological network. On this basis, an intra-layer fault propagation model and a cross-layer fault transfer model are simultaneously constructed, so that fault prediction can fully consider the complex structure and dynamic characteristics of the power grid. Subsequently, the power grid status data is acquired in real time and combined with the aforementioned model to perform fault prediction and early warning. Through a systematic and multi-level analysis method, the problem of lack of consideration of the multi-layer structure of the power grid and the inability to effectively simulate the dynamic propagation process of the fault in the traditional power grid fault prediction method is effectively solved, thereby improving the accuracy and timeliness of fault prediction.

[0015] After introducing the basic principles of the present application, various non-limiting implementation methods of the present application will be specifically described below in conjunction with the drawings in the specification.

[0016] Embodiment 1, as Figure 1 As shown, the embodiment of the present application provides a power grid fault prediction and processing method based on panoramic situation awareness, the method comprising:

[0017] S1: Collect data on the target power grid through the panoramic situational awareness system to obtain a multi-source perception data set.

[0018] Specifically, the panoramic situational awareness system is used to collect comprehensive data on the target power grid and obtain a multi-source perception data set. Among them, the panoramic situational awareness system is a comprehensive data acquisition platform that integrates a variety of sensor devices distributed throughout the power grid, including high-precision measuring instruments installed in substations, smart meters connected to the distribution network, and online monitoring devices for monitoring transmission lines, etc., forming a widely covered perception network that can capture various operating parameters and status information of the power grid in real time. The multi-source perception data set collected by the panoramic situational awareness system contains rich and diverse information, covering basic electrical parameters such as voltage, current, and power, equipment operating conditions such as transformer temperature and switch status, and environmental factors such as ambient temperature and humidity. In addition, it also includes the precise geographic location information of each device, providing a basis for the subsequent construction of the spatial topology of the power grid.

[0019] Through comprehensive data collection, a data foundation is laid for subsequent power grid analysis and fault prediction, so as to more accurately grasp the overall operating status of the power grid, thereby improving the accuracy and reliability of fault prediction.

[0020] S2: Build a power grid spatial topology structure based on the multi-source sensing data set.

[0021] Specifically, first, the multi-source sensing data set is preprocessed and fused. Preprocessing includes but is not limited to data cleaning, outlier processing, time series alignment and other operations to ensure the quality and consistency of the data; data fusion utilizes the complementarity of multi-source data to integrate information from different sources to form a more comprehensive and accurate description of the power grid state. Subsequently, based on the processed data, graph theory and complex network theory are used to abstract the power grid into a topological structure composed of nodes and edges. Among them, nodes represent key power equipment or facilities such as substations, distribution rooms, and important users, while edges represent the power transmission paths between these nodes, such as transmission lines, distribution lines, etc. In addition, the dynamic characteristics of the power grid are taken into account. By analyzing the topological changes in historical data, such as switch operations, line switching, etc., a dynamic topological model that can evolve over time is established, and a spatial topological structure of the power grid that combines electrical characteristics, geographic information and time dynamics is obtained, which lays a structural foundation for achieving high-precision power grid fault prediction.

[0022] S3: Clustering each node in the spatial topological structure of the power grid, and outputting a multi-layer topological network.

[0023] Specifically, first, the node characteristics of each node in the spatial topological structure of the power grid are obtained, including geographical location distance characteristics, electrical similarity characteristics and connection edge length characteristics. Then, based on these characteristics, a multi-dimensional similarity measurement method is used to calculate a comprehensive similarity index for each pair of nodes. For example, weighted Euclidean distance, cosine similarity or other suitable measurement methods are used to fully reflect the similarity between nodes. Subsequently, the nodes are grouped using a clustering algorithm, including but not limited to K-means, spectral clustering, hierarchical clustering, etc. During the clustering process, an adaptive parameter adjustment mechanism is introduced to ensure that the stability of each layer of the topological network meets the preset threshold. The clustering process is iterated until all nodes of the spatial topological structure of the power grid are traversed, thereby obtaining a multi-layer topological network structure, in which each layer represents a set of nodes with similar characteristics.

[0024] By performing multi-level clustering on the spatial topological structure of the power grid, a structured framework is provided for subsequent fault propagation analysis, which helps to reduce complexity, improve computational efficiency, and more accurately characterize the propagation characteristics of faults at different scales and ranges, thus laying the foundation for achieving high-precision, multi-scale power grid fault prediction.

[0025] S4: constructing an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network and a cross-layer fault transfer model between layers in the multi-layer topology network.

[0026] Specifically, two key models in multi-layer topology networks are constructed: a fault propagation model within each layer of the topology network, and a cross-layer fault transfer model between layers, to comprehensively describe how power grid faults propagate within and between different layers.

[0027] For the intra-layer fault propagation model, simulation data containing fault propagation paths and probabilities were first collected. Based on this data, a random walk model was used to describe the spread of faults in the network. By calculating the propagation probability between nodes and considering the initial state and evolution speed of the fault, a mathematical model that can accurately describe the propagation characteristics of intra-layer faults was finally obtained. When constructing the cross-layer fault transfer model, the connection nodes between different layers were first identified, and then a probabilistic model describing how the fault propagates from one layer to another was established based on the cross-layer connection relationship and fault transfer data. The random walk method was also used to consider the starting conditions and evolution process of cross-layer propagation to form a mathematical model for cross-layer fault transfer.

[0028] Through a multi-level, cross-scale modeling approach, the complexity and hierarchy of the power grid system are taken into account, and both the micro-mechanism and macro-trend of fault propagation can be captured simultaneously, laying a solid theoretical and model foundation for subsequent high-precision fault prediction, which helps to predict fault propagation in the power grid more comprehensively and accurately.

[0029] S5: Connecting to the panoramic situation awareness system to obtain real-time grid status data of the target grid.

[0030] Specifically, by connecting to the panoramic situational awareness system, the status data of the target power grid is obtained in real time. First, a real-time connection with the panoramic situational awareness system is established to ensure that the latest power grid operation information can be continuously obtained. Then, through the panoramic situational awareness system, the real-time power grid status data of the target power grid is obtained, including various electrical parameters, equipment status, environmental information, and user-side data, etc., to timely capture the dynamic changes in the power grid operation status, and provide the latest and most relevant input information for subsequent fault prediction, thereby improving the timeliness and accuracy of the prediction.

[0031] S6: Perform fault prediction on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and output multiple fault warning information.

[0032] Specifically, first, the acquired real-time grid status data is input into the intra-layer fault propagation model and the cross-layer fault transfer model for parallel calculation. During the calculation process, a dynamic weight adjustment mechanism is used to adaptively optimize the weights of different models according to the real-time data characteristics and historical prediction accuracy. Based on the model output, a comprehensive analysis and risk assessment are carried out, including calculating the failure probability of each node, evaluating the potential propagation path and impact range, and estimating the time window for the occurrence of the failure. Afterwards, according to the predefined risk threshold, multiple hierarchical fault warning information is generated, including fault location, type, potential impact and recommended measures, etc., to achieve comprehensive and accurate prediction of power grid faults and provide accurate decision support for power grid operation and maintenance.

[0033] Furthermore, the embodiment of the present application also includes:

[0034] Defining each power device component in the target power grid and a transmission path between the power device components according to the multi-source sensing data set;

[0035] The various power equipment components and the transmission paths are mapped into a panoramic three-dimensional space using a geographic information system to construct a power grid space topology structure, wherein the nodes of the power grid space topology structure are power equipment components, and the lengths of the sides of the power grid space topology structure correspond to the sizes of the transmission paths.

[0036] In a feasible implementation, first, the information from different data sources in the multi-source sensing data set is analyzed and integrated, including but not limited to equipment ledger data, real-time measurement data, historical operation records, etc. Through these data, key components in the power grid, such as substations, distribution rooms, important user terminals, etc., are identified, and the power transmission relationship between these components is determined, not only considering physical connections, but also factors such as electrical characteristics and operation modes, so as to more comprehensively reflect the actual structure and function of the power grid, and determine the transmission paths between each power equipment component and the power equipment component in the target power grid. Then, the further defined power equipment components and transmission paths are mapped into a panoramic three-dimensional space, so as to build a power grid space topology with geographical attributes. Among them, the abstract power grid structure is combined with the actual geographical space by using a geographic information system. In this process, each power equipment component is given precise geographic coordinates and becomes a node in the power grid space topology. At the same time, the transmission path between the equipment components is converted into an edge in the topology structure, where the length of the edge corresponds to the physical distance or electrical distance of the actual transmission path.

[0037] By converting the complex power grid system into a spatial topological structure with clear geographical attributes, not only the functional connection relationship of the power grid is retained, but also the real geographic spatial information is incorporated, providing a more comprehensive and accurate basic model for subsequent multi-level clustering analysis and fault propagation simulation, which helps to more accurately predict and analyze the fault propagation behavior in the power grid.

[0038] Furthermore, the embodiment of the present application also includes:

[0039] Acquire node features of each node in the spatial topological structure of the power grid, wherein the node features include geographical location distance features, electrical similarity features, and connection edge length features;

[0040] Generate a similarity metric for each node according to the geographic location distance feature, electrical similarity feature, and connection edge length feature;

[0041] Node allocation is performed according to the similarity measurement index of each node until all nodes of the power grid spatial topology structure are traversed, and a multi-layer topology network is output, wherein the stability of each layer of the topology network meets a preset threshold.

[0042] In a feasible implementation, first, the node features of each node in the spatial topology of the power grid are obtained, and the node features include geographical location distance features, electrical similarity features, and connection edge length features. Among them, the geographical location distance feature reflects the relative position relationship of the node in the physical space, and is quantified by indicators such as the Euclidean distance or Manhattan distance between nodes; the electrical similarity feature describes the similarity of the nodes in electrical characteristics, including factors such as voltage level, load characteristics, and equipment type; the connection edge length feature represents the length of the power transmission path between nodes. Then, based on the obtained node features, similarity metrics of each node are generated. For example, a multi-dimensional similarity calculation method is adopted, and the geographical location distance features, electrical similarity features, and connection edge length features are comprehensively considered by weighted Euclidean distance, cosine similarity, etc., to form a unified similarity metric standard to reflect the similarity between nodes. Then, based on the generated similarity metric, the nodes are allocated and clustered, and a multi-layer topology network is output. Among them, an iterative method is adopted to traverse all nodes in the spatial topology of the power grid one by one, and the nodes with high similarity are clustered to the same level. At the same time, a preset threshold is introduced, that is, each layer of the topological network needs to meet the preset stability threshold. If the stability of a certain layer of the topological network does not reach the preset threshold, corresponding adjustment strategies are adopted, such as reallocating boundary nodes, adjusting clustering parameters, etc., until the stability requirements are met.

[0043] By converting the complex spatial topological structure of the power grid into a multi-layer topological network with hierarchical characteristics, not only the key information of the original topology is retained, but also the potential functional modules and hierarchical relationships in the power grid are revealed through clustering, providing a network model for subsequent fault propagation analysis, which helps to predict and analyze the fault propagation behavior in the power grid more accurately and efficiently.

[0044] Furthermore, the embodiment of the present application also includes:

[0045] Acquire an intra-layer fault simulation data set, wherein the intra-layer fault simulation data set includes a propagation path and propagation probability of a fault sample between nodes in the layer;

[0046] Initialize the random walk model;

[0047] According to the intra-layer fault simulation data set, an intra-layer propagation probability matrix is ​​determined, wherein the element P in the intra-layer propagation probability matrix ij represents the probability of a fault propagating from node i to adjacent node j;

[0048] Configure the initial state vector and time step for fault propagation within the layer;

[0049] The initialized random walk model is trained with the intra-layer propagation probability matrix, the initial state vector of the intra-layer fault propagation and the time step, and an intra-layer fault propagation model that achieves a steady-state distribution is output.

[0050] In a preferred embodiment, an intra-layer fault propagation model corresponding to each layer of topological network in a multi-layer topological network is constructed. First, an intra-layer fault simulation data set is obtained, which includes the propagation path and propagation probability of fault samples between nodes in the layer. Among them, fault samples refer to power grid fault cases of different types and severity, the propagation path describes how the fault spreads from one node to other nodes, and the propagation probability quantifies the possibility of such diffusion. Then, initialize the random walk model. The random walk model is a mathematical model that describes the random movement of particles on the network, which is used to simulate the random propagation behavior of faults in the power grid. The initialization process involves defining the state space (that is, all nodes in the network), setting the initial transfer rules, etc., to prepare for subsequent model training.

[0051] Next, the intra-layer propagation probability matrix is ​​determined based on the intra-layer fault simulation data set. The element P in the intra-layer propagation probability matrix is ij Represents the probability of a fault propagating from node i to adjacent node j. For example, through maximum likelihood estimation, the acquired data is converted into a mathematical representation that can be directly used by the model, and the complex fault propagation characteristics are quantified into a computable probability value. Subsequently, the initial state vector and time step of the intra-layer fault propagation are configured. Among them, the initial state vector is a vector that describes the state of each node at the beginning of the fault, indicating the location of the fault source; the time step defines the time dimension of the model simulating fault propagation. After that, the initialized random walk model is trained using the determined intra-layer propagation probability matrix, initial state vector and time step. The training process adopts an iterative method to repeatedly apply the propagation probability matrix to the state vector until a steady-state distribution is reached. Among them, the steady-state distribution refers to the state in which the model output no longer changes significantly in a statistical sense, reflecting the long-term trend of fault propagation. After the training is completed, the intra-layer fault propagation model can reliably predict the propagation trend of the fault in the topological network of this layer.

[0052] By constructing a mathematical model that can accurately describe the fault propagation characteristics within a layer, not only the characteristics of the power grid topology are taken into account, but also the randomness and dynamic characteristics of fault propagation are incorporated, providing support for achieving high-precision power grid fault prediction.

[0053] Furthermore, the embodiment of the present application also includes:

[0054] Identifying a cross-layer connection node between layers in the multi-layer topology network, wherein the cross-layer connection node is a critical path for a fault to propagate from one hierarchical network to another hierarchical network;

[0055] Defining a cross-layer adjacency matrix according to cross-layer connection nodes between layers in the multi-layer topology network;

[0056] A transfer simulation is performed according to the cross-layer adjacency matrix to construct a cross-layer fault transfer model between layers in the multi-layer topology network.

[0057] In a feasible implementation, first, the cross-layer connection nodes between layers in the multi-layer topology network are identified. The cross-layer connection node is the path for the fault to propagate from one hierarchical network to another hierarchical network, and is represented by transformers, tie switches and other equipment between different voltage levels in the power grid. The identification process is achieved by analyzing the structural characteristics of the multi-layer topology network. Specifically, all nodes in the network are traversed to check whether each node has a direct connection with nodes at other levels. If such a connection exists, the node is marked as a cross-layer connection node. Then, a cross-layer adjacency matrix is ​​defined according to the identified cross-layer connection nodes, and the cross-layer adjacency matrix is ​​a mathematical representation describing the connection relationship between nodes at different levels. Then, a transfer simulation is performed according to the cross-layer adjacency matrix to construct a cross-layer fault transfer model between layers in the multi-layer topology network. Specifically, the cross-layer adjacency matrix is ​​first normalized to obtain a transfer probability matrix; for each cross-layer connection node, the simulated fault starts from the node and randomly walks to nodes at other levels according to the transfer probability matrix. Through a large number of simulation experiments, the probability distribution of the fault propagating from one level to another is statistically calculated to form a cross-layer fault transfer model.

[0058] By constructing a mathematical model that describes the fault transfer characteristics between layers, the cross-layer fault transfer model not only takes into account the multi-layer structural characteristics of the power grid, but also captures the dynamic process of fault propagation between different levels, providing an important tool for achieving comprehensive and accurate power grid fault prediction. This model, combined with the aforementioned intra-layer fault propagation model, can fully describe the propagation and evolution process of faults in the power grid, thereby improving the accuracy and reliability of fault prediction.

[0059] Furthermore, the embodiment of the present application also includes:

[0060] Acquire a cross-layer fault simulation data set, wherein the cross-layer fault simulation data set includes a transfer path and a transfer probability of a fault sample between layers;

[0061] Determine a cross-layer propagation probability matrix according to the cross-layer fault simulation data set, wherein the elements in the cross-layer propagation probability matrix are represents the probability that a fault propagates from node i in the kth layer of the network to node j in the lth layer of the network;

[0062] Configure the initial state vector and time step for cross-layer failover;

[0063] The initialized random walk model is trained with the cross-layer propagation probability matrix, the initial state vector of the cross-layer fault transfer and the time step, and a cross-layer fault transfer model that achieves a steady-state distribution is output.

[0064] In a preferred embodiment, first, a cross-layer fault simulation data set is obtained, which contains the transfer paths and transfer probability information of fault samples between layers. The transfer path describes how the fault propagates from a node at one level to a node at another level, and the transfer probability quantifies the possibility of such cross-layer propagation. Then, a cross-layer propagation probability matrix is ​​determined based on the cross-layer fault simulation data set. The elements Represents the probability that a fault propagates from node i of the kth layer network to node j of the lth layer network. For example, through maximum likelihood estimation, the obtained cross-layer fault simulation data set is converted into a mathematical representation that can be directly used by the model, and the complex cross-layer fault propagation characteristics are quantified into a computable probability value. Then, the initial state vector and time step of the cross-layer fault transfer are configured. Among them, the initial state vector is a vector that describes the state of each node in each layer at the beginning of the fault, indicating which node in which layer the fault source is located; the time step defines the time dimension of the model simulation of cross-layer fault propagation. After that, the initialized random walk model is trained using the determined cross-layer propagation probability matrix, initial state vector and time step. The training process adopts an iterative method to repeatedly apply the cross-layer propagation probability matrix to the state vector until a steady-state distribution is reached. Among them, the steady-state distribution refers to the state in which the model output no longer changes significantly in a statistical sense, reflecting the long-term trend of cross-layer fault propagation. After the training is completed, the cross-layer fault transfer model can reliably predict the cross-layer propagation trend of faults in a multi-layer topology network.

[0065] By constructing a mathematical model that can accurately describe the cross-layer fault propagation characteristics, not only the characteristics of the multi-layer topology of the power grid are considered, but also the randomness and dynamic characteristics of cross-layer fault propagation are incorporated, laying a solid theoretical foundation for achieving high-precision power grid fault prediction. This model, combined with the aforementioned intra-layer fault propagation model, can fully describe the propagation and evolution process of faults in the power grid, thereby greatly improving the accuracy, comprehensiveness and reliability of fault prediction.

[0066] Furthermore, the embodiment of the present application also includes:

[0067] Performing fault prediction on the real-time power grid status data based on the intra-layer fault propagation model, and outputting a plurality of intra-layer fault prediction indicators corresponding to the multi-layer topology network;

[0068] Performing fault prediction on the real-time power grid status data based on the cross-layer fault transfer model, and outputting a plurality of cross-layer fault prediction indicators;

[0069] Performing conditional probability calculation on the multiple intra-layer fault prediction indicators and the multiple cross-layer fault prediction indicators, and outputting multiple fault prediction indicators;

[0070] Output multiple fault warning information according to the size of the fault prediction index.

[0071] In a feasible implementation, first, the real-time power grid state data is predicted based on the intra-layer fault propagation model, and multiple intra-layer fault prediction indicators corresponding to the multi-layer topology network are output. Specifically, the real-time power grid state data is input into the constructed intra-layer fault propagation model, which uses a random walk algorithm to simulate the propagation process of the fault within each layer of the topology network. Through multiple simulations, the probability of each node failing in a specific time period in the future is calculated, forming an intra-layer fault prediction indicator. This process is performed on each layer in the multi-layer topology network to obtain a set of multiple intra-layer fault prediction indicators covering all levels. Next, the real-time power grid state data is predicted based on the cross-layer fault transfer model, and multiple cross-layer fault prediction indicators are output. Specifically, the real-time data is input into the cross-layer fault transfer model, which considers the propagation characteristics of the fault between different levels and simulates the process of the fault propagating from one level to another. Through calculation, multiple cross-layer fault prediction indicators describing the possibility of cross-layer propagation of the fault are obtained, reflecting the potential diffusion trend of the fault in the entire multi-layer network.

[0072] Subsequently, conditional probability calculations are performed on multiple intra-layer fault prediction indicators and multiple cross-layer fault prediction indicators, and multiple comprehensive fault prediction indicators are output. For example, using Bayes' theorem, the comprehensive probability of failure at each node is calculated under the given intra-layer fault probability and cross-layer propagation probability, taking into account the mutual influence of fault propagation within the layer and across layers, so as to obtain more comprehensive and accurate multiple fault prediction indicators. Afterwards, multiple fault warning information is output based on the size of the fault prediction indicator. Specifically, multiple fault prediction indicators are converted into specific multiple fault warning information, including the location where the fault may occur, the expected time of occurrence, the potential impact range, etc.

[0073] Through comprehensive and accurate prediction of power grid faults, not only the propagation of faults within a single layer is taken into account, but also the impact of cross-layer propagation, thereby providing more comprehensive and accurate fault warning information, improving the accuracy and timeliness of power grid fault warnings, and providing strong support for the safe and stable operation of the power grid.

[0074] In summary, the power grid fault prediction and processing method based on panoramic situation awareness provided by the embodiment of the present application has the following technical effects:

[0075] Through the panoramic situation awareness system, data collection is carried out on the target power grid to obtain multi-source perception data sets to ensure the comprehensiveness and accuracy of the data. According to the multi-source perception data sets, the spatial topology of the power grid is constructed to obtain the connection relationship and spatial distribution between the components of the power grid, laying the foundation for subsequent multi-level analysis. Clustering is performed on each node in the spatial topology of the power grid, and a multi-layer topology network is output, which creates conditions for more refined and targeted fault analysis. The intra-layer fault propagation model corresponding to each layer of the multi-layer topology network and the cross-layer fault transfer model between layers in the multi-layer topology network are constructed to comprehensively characterize the dynamic propagation characteristics of the fault and improve the accuracy of fault prediction. The panoramic situation awareness system is connected to obtain the real-time power grid status data of the target power grid, so that fault prediction can be carried out based on the current actual situation, improving the timeliness and accuracy of the prediction. Fault prediction is carried out on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and multiple fault warning information is output to improve the accuracy and timeliness of power grid fault prediction.

[0076] Embodiment 2 is based on the same inventive concept as the power grid fault prediction and processing method based on panoramic situation awareness in the above embodiment. Figure 2 As shown, the embodiment of the present application provides a power grid fault prediction and processing system based on panoramic situation awareness, the system comprising:

[0077] The data acquisition module 11 is used to collect data from the target power grid through the panoramic situation awareness system to obtain a multi-source perception data set;

[0078] A topology building module 12, used to build a power grid spatial topology structure according to the multi-source sensing data set;

[0079] A multi-layer clustering module 13, used to cluster each node in the power grid spatial topology structure and output a multi-layer topology network;

[0080] A model building module 14 is used to build an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network, and a cross-layer fault transfer model between layers in the multi-layer topology network;

[0081] A real-time data acquisition module 15 is used to connect to the panoramic situation awareness system to acquire real-time grid status data of the target grid;

[0082] The fault prediction module 16 is used to perform fault prediction on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and output multiple fault warning information.

[0083] Furthermore, the topology building module 12 includes the following execution steps:

[0084] Defining each power device component in the target power grid and a transmission path between the power device components according to the multi-source sensing data set;

[0085] The various power equipment components and the transmission paths are mapped into a panoramic three-dimensional space using a geographic information system to construct a power grid space topology structure, wherein the nodes of the power grid space topology structure are power equipment components, and the lengths of the sides of the power grid space topology structure correspond to the sizes of the transmission paths.

[0086] Furthermore, the multi-layer clustering module 13 includes the following execution steps:

[0087] Acquire node features of each node in the spatial topological structure of the power grid, wherein the node features include geographical location distance features, electrical similarity features, and connection edge length features;

[0088] Generate a similarity metric for each node according to the geographic location distance feature, electrical similarity feature, and connection edge length feature;

[0089] Node allocation is performed according to the similarity measurement index of each node until all nodes of the power grid spatial topology structure are traversed, and a multi-layer topology network is output, wherein the stability of each layer of the topology network meets a preset threshold.

[0090] Furthermore, the model building module 14 includes the following execution steps:

[0091] Acquire an intra-layer fault simulation data set, wherein the intra-layer fault simulation data set includes a propagation path and propagation probability of a fault sample between nodes in the layer;

[0092] Initialize the random walk model;

[0093] According to the intra-layer fault simulation data set, an intra-layer propagation probability matrix is ​​determined, wherein the element P in the intra-layer propagation probability matrix ij represents the probability of a fault propagating from node i to adjacent node j;

[0094] Configure the initial state vector and time step for fault propagation within the layer;

[0095] The initialized random walk model is trained with the intra-layer propagation probability matrix, the initial state vector of the intra-layer fault propagation and the time step, and an intra-layer fault propagation model that achieves a steady-state distribution is output.

[0096] Furthermore, the model building module 14 also includes the following execution steps:

[0097] Identifying a cross-layer connection node between layers in the multi-layer topology network, wherein the cross-layer connection node is a critical path for a fault to propagate from one hierarchical network to another hierarchical network;

[0098] Defining a cross-layer adjacency matrix according to cross-layer connection nodes between layers in the multi-layer topology network;

[0099] A transfer simulation is performed according to the cross-layer adjacency matrix to construct a cross-layer fault transfer model between layers in the multi-layer topology network.

[0100] Furthermore, the model building module 14 also includes the following execution steps:

[0101] Acquire a cross-layer fault simulation data set, wherein the cross-layer fault simulation data set includes a transfer path and a transfer probability of a fault sample between layers;

[0102] Determine a cross-layer propagation probability matrix according to the cross-layer fault simulation data set, wherein the elements in the cross-layer propagation probability matrix are represents the probability that a fault propagates from node i in the kth layer of the network to node j in the lth layer of the network;

[0103] Configure the initial state vector and time step for cross-layer failover;

[0104] The initialized random walk model is trained with the cross-layer propagation probability matrix, the initial state vector of the cross-layer fault transfer and the time step, and a cross-layer fault transfer model that achieves a steady-state distribution is output.

[0105] Furthermore, the fault prediction module 16 includes the following execution steps:

[0106] Performing fault prediction on the real-time power grid status data based on the intra-layer fault propagation model, and outputting a plurality of intra-layer fault prediction indicators corresponding to the multi-layer topology network;

[0107] Performing fault prediction on the real-time power grid status data based on the cross-layer fault transfer model, and outputting a plurality of cross-layer fault prediction indicators;

[0108] Performing conditional probability calculation on the multiple intra-layer fault prediction indicators and the multiple cross-layer fault prediction indicators, and outputting multiple fault prediction indicators;

[0109] Output multiple fault warning information according to the size of the fault prediction index.

[0110] Any step of the method described above can be stored as a computer instruction or program in an unlimited computer memory, and can be called and recognized by an unlimited computer processor to implement any method in the embodiments of the present application, without any unnecessary restrictions.

[0111] Obviously, those skilled in the art can make various changes and modifications to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A power grid fault prediction and processing method based on panoramic situation awareness, characterized in that: The method comprises: Collect data from the target power grid through the panoramic situation awareness system to obtain multi-source perception data sets; Building a power grid spatial topology structure according to the multi-source sensing data set; Clustering each node in the power grid spatial topology structure and outputting a multi-layer topology network; Constructing an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network and a cross-layer fault transfer model between layers in the multi-layer topology network; Connecting to the panoramic situation awareness system to obtain real-time grid status data of the target grid; Based on the intra-layer fault propagation model and the cross-layer fault transfer model, fault prediction is performed on the real-time power grid status data, and multiple fault warning information is output.

2. The method according to claim 1, characterized in that According to the multi-source sensing data set, a power grid spatial topology structure is constructed, and the method further includes: Defining each power device component in the target power grid and a transmission path between the power device components according to the multi-source sensing data set; The various power equipment components and the transmission paths are mapped into a panoramic three-dimensional space using a geographic information system to construct a power grid space topology structure, wherein the nodes of the power grid space topology structure are power equipment components, and the lengths of the sides of the power grid space topology structure correspond to the sizes of the transmission paths.

3. The method according to claim 1, characterized in that Clustering each node in the power grid spatial topology structure and outputting a multi-layer topology network, the method includes: Acquire node features of each node in the spatial topological structure of the power grid, wherein the node features include geographical location distance features, electrical similarity features, and connection edge length features; Generate a similarity metric for each node according to the geographic location distance feature, electrical similarity feature, and connection edge length feature; Node allocation is performed according to the similarity measurement index of each node until all nodes of the power grid spatial topology structure are traversed, and a multi-layer topology network is output, wherein the stability of each layer of the topology network meets a preset threshold.

4. The method according to claim 1, characterized in that Constructing an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network, the method comprising: Acquire an intra-layer fault simulation data set, wherein the intra-layer fault simulation data set includes a propagation path and propagation probability of a fault sample between nodes in the layer; Initialize the random walk model; According to the intra-layer fault simulation data set, an intra-layer propagation probability matrix is ​​determined, wherein the element P in the intra-layer propagation probability matrix ij represents the probability of a fault propagating from node i to adjacent node j; Configure the initial state vector and time step for fault propagation within the layer; The initialized random walk model is trained with the intra-layer propagation probability matrix, the initial state vector of the intra-layer fault propagation and the time step, and an intra-layer fault propagation model that achieves a steady-state distribution is output.

5. The method according to claim 1, characterized in that Constructing a cross-layer fault transfer model between layers in the multi-layer topology network, the method comprising: Identifying a cross-layer connection node between layers in the multi-layer topology network, wherein the cross-layer connection node is a critical path for a fault to propagate from one hierarchical network to another hierarchical network; Defining a cross-layer adjacency matrix according to cross-layer connection nodes between layers in the multi-layer topology network; A transfer simulation is performed according to the cross-layer adjacency matrix to construct a cross-layer fault transfer model between layers in the multi-layer topology network.

6. The method according to claim 5, characterized in that Performing transfer simulation according to the cross-layer adjacency matrix to construct a cross-layer fault transfer model between layers in the multi-layer topology network, the method comprising: Acquire a cross-layer fault simulation data set, wherein the cross-layer fault simulation data set includes a transfer path and a transfer probability of a fault sample between layers; Determine a cross-layer propagation probability matrix according to the cross-layer fault simulation data set, wherein the elements in the cross-layer propagation probability matrix are represents the probability that a fault propagates from node i in the kth layer of the network to node j in the lth layer of the network; Configure the initial state vector and time step for cross-layer failover; The initialized random walk model is trained with the cross-layer propagation probability matrix, the initial state vector of the cross-layer fault transfer and the time step, and a cross-layer fault transfer model that achieves a steady-state distribution is output.

7. The method according to claim 1, characterized in that Outputting multiple fault warning information, the method also includes: Performing fault prediction on the real-time power grid status data based on the intra-layer fault propagation model, and outputting a plurality of intra-layer fault prediction indicators corresponding to the multi-layer topology network; Performing fault prediction on the real-time power grid status data based on the cross-layer fault transfer model, and outputting a plurality of cross-layer fault prediction indicators; Performing conditional probability calculation on the multiple intra-layer fault prediction indicators and the multiple cross-layer fault prediction indicators, and outputting multiple fault prediction indicators; Output multiple fault warning information according to the size of the fault prediction index.

8. A power grid fault prediction and processing system based on panoramic situational awareness, characterized in that: A method for predicting and processing power grid faults based on panoramic situational awareness according to any one of claims 1 to 7, the method comprising: A data acquisition module is used to collect data from the target power grid through a panoramic situation awareness system to obtain a multi-source perception data set; A topology building module, used to build a power grid spatial topology structure according to the multi-source sensing data set; A multi-layer clustering module, used to cluster each node in the power grid spatial topology structure and output a multi-layer topology network; A model building module, used to build an intra-layer fault propagation model corresponding to each layer of the multi-layer topology network, and a cross-layer fault transfer model between layers in the multi-layer topology network; A real-time data acquisition module, used to connect to the panoramic situation awareness system to acquire real-time grid status data of the target grid; The fault prediction module is used to perform fault prediction on the real-time power grid status data based on the intra-layer fault propagation model and the cross-layer fault transfer model, and output multiple fault warning information.

Images