Dynamic sensing method and system for digital twins of power distribution network

Abstract

The invention discloses a power distribution network digital twin dynamic sensing method and system, and the method comprises the steps: obtaining the real-time data of the operation of a power distribution network, and monitoring the node change and line state change of the power distribution network based on the real-time data; a power distribution network topological structure diagram in the established power distribution network digital twin model is updated in real time based on node changes and line state changes; based on the real-time data and the historical data, analyzing the operation parameters of the power distribution network through a digital twin model of the power distribution network to obtain an analysis result; and detecting and positioning operation faults of the power distribution network through a power distribution network digital twinborn model based on an analysis result. The invention provides a power distribution network digital twinborn dynamic sensing method and system based on multi-element network topology and parameter change, which can capture and reflect variable network structure and parameter change in a power distribution network in real time and improve the dynamic sensing capability of the digital twinborn so as to support fault early warning and rapid recovery of the power distribution network.

Description

Technical Field

[0001] The present invention relates to the technical field of dynamic perception of digital twins of distribution networks, and more specifically, to a dynamic perception method and system for digital twins of distribution networks. Background Art

[0002] With the increasing complexity and intelligence of power systems, traditional means of monitoring and managing distribution networks can no longer meet the needs of modern power grid operation. As an emerging digital technology, digital twin technology is gradually being applied to power systems to improve the efficiency and reliability of power grid operation.

[0003] Digital twin technology refers to the establishment of a digital model of a physical entity to achieve synchronous updating and interaction between the physical entity and the digital model. Specifically, a digital twin is a real-time, dynamic, high-precision digital model that can reflect the state, behavior, and operating environment of a physical entity. The core of digital twin technology is data-driven, collecting data from physical entities through sensors, IoT devices, etc., and inputting these data into the digital model in real time, thereby achieving accurate simulation and prediction of physical entities.

[0004] A typical digital twin system architecture generally includes the following parts:

[0005] Data acquisition layer: responsible for obtaining real-time data from physical entities, including sensor data, historical data, environmental data, etc. The data sources of this layer are extensive, covering various equipment and operating environments in the power system.

[0006] Data transmission and processing layer: The collected data is transmitted to the data processing center through the network for data cleaning, filtering and preprocessing. The key technologies of this layer include edge computing, distributed computing and big data processing technologies to ensure the timeliness and accuracy of data processing.

[0007] Digital twin model layer: Based on the preprocessed data, a digital twin model corresponding to the physical entity is established. The core parts of the model include physical modeling, data-driven modeling, and simulation algorithms. Physical modeling is mainly based on the physical laws of the power system, while data-driven modeling uses machine learning and artificial intelligence technologies to predict and analyze historical data and real-time data.

[0008] Application and service layer: This layer provides users with a visualization interface, analysis tools, and decision support system for digital twins. Users can monitor, simulate, and control digital twins through this interface to optimize the operation of the power system.

[0009] In the field of distribution networks, digital twin technology has been widely used in equipment monitoring, fault diagnosis, operation optimization and intelligent maintenance. For example, digital twins can monitor the operating status of distribution transformers in real time and warn of possible faults in advance through predictive analysis, thereby realizing intelligent fault diagnosis and maintenance.

[0010] However, as the scale and complexity of distribution networks increase, the network topology and operating parameters in distribution networks are becoming more diverse and dynamic. This places higher demands on digital twin technology, mainly in the following aspects:

[0011] Real-time requirements: The network topology and operating parameters in the distribution network may change frequently, and the traditional static digital twin model is difficult to adapt to such dynamic changes. Therefore, how to achieve real-time update and dynamic perception of the digital twin model becomes a key issue.

[0012] Multi-source data fusion: Data in the distribution network comes from various sources, including sensor data, historical operation data, environmental data, etc. These data usually have different formats and time scales. How to effectively integrate these multi-source data into the digital twin is a major challenge in the application of digital twin technology in the distribution network.

[0013] Large-scale parallel computing: With the increase of distribution network nodes, the amount of data and computing required by digital twins is also increasing dramatically. Traditional single-node computing methods are difficult to cope with the real-time simulation and analysis needs of large-scale distribution networks, and parallel computing and distributed computing technologies are urgently needed to improve computing efficiency.

[0014] However, in the existing technology, although the application of digital twin technology in distribution networks has made certain progress, there are still significant deficiencies in dealing with changes in multiple network topologies and parameters:

[0015] Slow response speed: When faced with rapid changes in the distribution network topology, existing digital twin systems usually take a long time to update the model, resulting in slow response speed and inability to meet the requirements of real-time monitoring and control.

[0016] Insufficient perception accuracy: Since multi-source data fusion technology is still immature, existing systems often have data errors and delays when sensing changes in distribution network operating parameters, affecting the accuracy and reliability of digital twins.

[0017] Low computing efficiency: In large-scale distribution network scenarios, the computing burden of digital twins is extremely heavy. Existing systems are unable to efficiently process massive amounts of data, resulting in low computing efficiency and affecting the overall system operation.

[0018] Therefore, a technology is needed to realize the dynamic perception technology of distribution network digital twin based on multiple network topologies and parameter changes. Summary of the invention

[0019] The technical solution of the present invention provides a distribution network digital twin dynamic perception method and system to solve the problem of how to achieve distribution network digital twin dynamic perception based on multiple network topologies and parameter changes.

[0020] In order to solve the above problems, the present invention provides a distribution network digital twin dynamic perception method, the method comprising:

[0021] Acquire real-time data of the operation of the distribution network, and monitor node changes and line status changes of the distribution network based on the real-time data;

[0022] Based on the node changes and the line state changes, a distribution network topology diagram in the established distribution network digital twin model is updated in real time;

[0023] Based on the real-time data and historical data, the operating parameters of the distribution network are analyzed by the distribution network digital twin model to obtain analysis results;

[0024] Based on the analysis results, the distribution network digital twin model is used to detect and locate operational faults in the distribution network.

[0025] Preferably, the method further comprises preprocessing the real-time data, including: removing abnormal values ​​in the real-time data, and removing noise of the real-time data by a low-pass filter.

[0026] Preferably, the real-time updating of the distribution network topology diagram in the established distribution network digital twin model based on the node change and the line state change includes:

[0027] Traversing the distribution network topology diagram based on a dynamic update algorithm, removing failed edges and adding new edges;

[0028] The shortest path of each node in the updated distribution network topology diagram is calculated.

[0029] Preferably, the analyzing the operating parameters of the distribution network through the distribution network digital twin model based on the real-time data and the historical data includes:

[0030] The operating parameters are predicted by a long short-term memory network.

[0031] Preferably, the method further comprises: training the long short-term memory network by a loss function, wherein the loss function is:

[0032]

[0033] Where L is the loss function, which indicates the difference between the predicted value and the true value. The LSTM model is trained by minimizing the loss function to improve the accuracy of the prediction. t is the time step index, which indicates the tth moment in the data sequence. T is the total length of the time series, which indicates the number of time steps contained in the data set. t+1 is the true value at time step t+1, that is, the actual measured value of the distribution network operating parameters (such as voltage, current, power factor, etc.) at a future moment. is the predicted value at time step t+1, which is the value of the distribution network operation parameter at a future moment predicted by the LSTM model based on the previous input data.

[0034] Preferably, the detecting and locating the operation fault of the distribution network through the distribution network digital twin model based on the analysis result includes:

[0035] The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault:

[0036] P(F|T,P)∝P(T|F)·P(P|F)·P(F)

[0037] Among them, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P. This is the result we want to obtain through Bayesian network fusion analysis, which is used to infer the possibility of fault occurrence under the current network topology and parameter state. P(T|F) is the probability of topology change data under the condition of fault occurrence. It represents the probability of a specific change in the topology structure of the distribution network when a fault occurs, which can be obtained based on the statistical relationship between historical fault data and topology changes. P(P|F) is the probability of parameter change data under the condition of fault occurrence. That is, when a fault occurs, the probability of a specific change in the operating parameters of the distribution network can also be learned from historical data. P(F) is the prior probability of fault occurrence, that is, the probability of fault occurrence without considering any topology or parameter change data, which is usually estimated based on historical fault statistics.

[0038] Preferably, the method further comprises updating the Bayesian network:

[0039]

[0040] Among them, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X) describes the possibility of the current observation data under specific parameter settings, which is based on the assumption of the probability model on the data generation process. 1：t) is the observation data X from time step 1 to t. 1：t The prior probability distribution of the parameter θ under the condition of . This is the probability estimate of the parameter before the new data arrives, reflecting previous knowledge or experience. t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 This is a normalization constant that ensures that the sum of the posterior probability distribution is 1. It is used to adjust the proportional relationship of the probability in the calculation so that the posterior probability distribution satisfies the probability axiom. 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 Under the condition of , the posterior probability distribution of the parameters is obtained. This is the probability distribution of the parameters updated with new data and is used to more accurately estimate the state of the distribution network.

[0041] Based on another aspect of the present invention, the present invention provides a distribution network digital twin dynamic perception system, the system comprising:

[0042] A network topology perception unit, used to obtain real-time data of the distribution network operation, monitor node changes and line state changes of the distribution network based on the real-time data; and update the distribution network topology structure diagram in the established distribution network digital twin model in real time based on the node changes and line state changes;

[0043] A parameter change sensing unit, used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on the real-time data and historical data, and obtain analysis results;

[0044] A fusion analysis unit is used to detect and locate operation faults of the distribution network through the distribution network digital twin model based on the analysis results.

[0045] Preferably, the network topology sensing unit is also used to pre-process the real-time data, including: clearing abnormal values ​​in the real-time data, and removing noise from the real-time data by a low-pass filter.

[0046] Preferably, the network topology perception unit is used to update the distribution network topology structure diagram in the established distribution network digital twin model in real time based on the node change and the line state change, and is also used to:

[0047] Traversing the distribution network topology diagram based on a dynamic update algorithm, removing failed edges and adding new edges;

[0048] The shortest path of each node in the updated distribution network topology diagram is calculated.

[0049] Preferably, the parameter change sensing unit is used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on the real-time data and the historical data, and is also used to:

[0050] The operating parameters are predicted by a long short-term memory network.

[0051] Preferably, the parameter change sensing unit is further used to train the long short-term memory network through a loss function, and the loss function is:

[0052]

[0053] Where L is the loss function, which indicates the difference between the predicted value and the true value. The LSTM model is trained by minimizing the loss function to improve the accuracy of the prediction. t is the time step index, which indicates the tth moment in the data sequence. T is the total length of the time series, which indicates the number of time steps contained in the data set. t+1 is the true value at time step t+1, that is, the actual measured value of the distribution network operating parameters (such as voltage, current, power factor, etc.) at a future moment. is the predicted value at time step t+1, which is the value of the distribution network operation parameter at a future moment predicted by the LSTM model based on the previous input data.

[0054] Preferably, the fusion analysis unit is used to detect and locate operation faults of the distribution network through the distribution network digital twin model based on the analysis result, and is also used to:

[0055] The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault:

[0056] P(F|T,P)∝P(T|F)·P(P|F)·P(F)

[0057] Among them, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P. This is the result we want to obtain through Bayesian network fusion analysis, which is used to infer the possibility of fault occurrence under the current network topology and parameter state. P(T|F) is the probability of topology change data under the condition of fault occurrence. It represents the probability of a specific change in the topology structure of the distribution network when a fault occurs, which can be obtained based on the statistical relationship between historical fault data and topology changes. P(P|F) is the probability of parameter change data under the condition of fault occurrence. That is, when a fault occurs, the probability of a specific change in the operating parameters of the distribution network can also be learned from historical data. P(F) is the prior probability of fault occurrence, that is, the probability of fault occurrence without considering any topology or parameter change data, which is usually estimated based on historical fault statistics.

[0058] Preferably, the fusion analysis unit is further used to update the Bayesian network:

[0059]

[0060] Among them, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X) describes the possibility of the current observation data under specific parameter settings, which is based on the assumption of the probability model on the data generation process. 1：t ) is the observation data X from time step 1 to t. 1：t The prior probability distribution of the parameter θ under the condition of . This is the probability estimate of the parameter before the new data arrives, reflecting previous knowledge or experience. t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 This is a normalization constant that ensures that the sum of the posterior probability distribution is 1. It is used to adjust the proportional relationship of the probability in the calculation so that the posterior probability distribution satisfies the probability axiom. 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 Under the condition of , the posterior probability distribution of the parameters is obtained. This is the probability distribution of the parameters updated with new data and is used to more accurately estimate the state of the distribution network.

[0061] Based on another aspect of the present invention, the present invention provides a computer-readable storage medium, wherein the computer-readable storage medium stores a computer program, and the computer program is used to execute a dynamic perception method of a distribution network digital twin.

[0062] Based on another aspect of the present invention, the present invention provides an electronic device, the electronic device comprising: a processor and a memory; wherein,

[0063] The memory is a memory for storing instructions executable by the processor;

[0064] The processor is used to read the executable instructions from the memory and execute the instructions to implement a dynamic perception method of a distribution network digital twin.

[0065] The technical solution of the present invention provides a method and system for dynamic perception of a digital twin of a distribution network, wherein the method includes: obtaining real-time data of the operation of the distribution network, and monitoring node changes and line status changes of the distribution network based on the real-time data; updating the distribution network topology diagram in the established digital twin model of the distribution network in real time based on the node changes and line status changes; analyzing the operating parameters of the distribution network through the digital twin model of the distribution network based on real-time data and historical data to obtain analysis results; and detecting and locating operating faults of the distribution network through the digital twin model of the distribution network based on the analysis results. The technical solution of the present invention provides a method and system for dynamic perception of a digital twin of a distribution network based on multi-network topology and parameter changes, which can capture and reflect the changeable network structure and parameter changes in the distribution network in real time, improve the dynamic perception ability of the digital twin, and thus support real-time monitoring, fault warning and rapid recovery of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0066] A more complete understanding of exemplary embodiments of the present invention may be obtained by referring to the following drawings:

[0067] Figure 1 A flow chart of a dynamic perception method of a distribution network digital twin according to a preferred embodiment of the present invention;

[0068] Figure 2 A flow chart of a distribution network digital twin dynamic perception method for multiple network topologies and parameter changes according to a preferred embodiment of the present invention; and

[0069] Figure 3 The figure is a structural diagram of a distribution network digital twin dynamic perception system according to a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0070] Now, exemplary embodiments of the present invention are described with reference to the accompanying drawings. However, the present invention can be implemented in many different forms and is not limited to the embodiments described herein. These embodiments are provided to disclose the present invention in detail and completely and to fully convey the scope of the present invention to those skilled in the art. The terms used in the exemplary embodiments shown in the accompanying drawings are not intended to limit the present invention. In the accompanying drawings, the same units / elements are marked with the same reference numerals.

[0071] Unless otherwise specified, the terms (including technical terms) used herein have the commonly understood meanings to those skilled in the art. In addition, it is understood that the terms defined in commonly used dictionaries should be understood to have the same meanings as those in the context of the relevant fields, and should not be understood as idealized or overly formal meanings.

[0072] Figure 1 The present invention is a flow chart of a dynamic perception method of a distribution network digital twin according to a preferred embodiment of the present invention.

[0073] The present invention proposes a dynamic perception method of distribution network digital twins for multi-network topology and parameter changes, aiming to overcome the limitations of existing technologies and enhance the real-time dynamic perception capability of distribution network digital twins, so as to better serve the intelligent management and optimization of power grids.

[0074] The present invention provides a dynamic perception method of a distribution network digital twin based on multiple network topologies and parameter changes, which can capture and reflect the changeable network structure and parameter changes in the distribution network in real time, enhance the dynamic perception capability of the digital twin, and thus support real-time monitoring, fault warning and rapid recovery of the distribution network.

[0075] The core of the present invention is to propose a dynamic perception method of distribution network digital twins for multi-network topology and parameter changes. By comprehensively applying advanced graph theory, semantic network representation, machine learning and data fusion technology, the present invention can effectively cope with the complex and changeable network structure and real-time parameter changes in the distribution network. The following content will elaborate on the various component modules of the present invention and their implementation methods, and introduce mathematical models and algorithm descriptions to enhance the accuracy and operability of the technology.

[0076] like Figure 1 As shown, the present invention provides a distribution network digital twin dynamic perception method, the method comprising:

[0077] Step 101: Acquire real-time data of the distribution network operation, and monitor node changes and line status changes of the distribution network based on the real-time data;

[0078] Preferably, the method further comprises preprocessing the real-time data, including: removing abnormal values ​​in the real-time data, and removing noise of the real-time data by a low-pass filter.

[0079] Step 102: updating the distribution network topology diagram in the established distribution network digital twin model in real time based on node changes and line status changes;

[0080] Preferably, the distribution network topology diagram in the established distribution network digital twin model is updated in real time based on node changes and line state changes, including:

[0081] Based on the dynamic update algorithm, the topology diagram of the distribution network is traversed to remove the failed edges and add new edges;

[0082] Calculate the shortest path for each node in the updated distribution network topology graph.

[0083] The present invention comprises a network topology sensing unit:

[0084] 1) Functional description: This unit monitors the status changes of nodes and lines in the distribution network in real time, including the addition of new nodes, node failures, and line connection status. Based on graph theory, the network topology perception module builds and dynamically updates the topological model of the distribution network to ensure that the model can reflect the actual structure of the physical network in real time.

[0085] 2) Technical implementation: Using graph databases (such as Neo4j) and parallel computing technology, this unit can efficiently handle the topology update of large-scale distribution networks. The core algorithm of topology update is based on the shortest path algorithm and the maximum flow algorithm. Assume that the topology structure of the distribution network is a graph G(V,E), where V represents the node set and E represents the edge set. The dynamic update algorithm is described based on the following pseudo code:

[0086]

[0087] Step 103: Based on the real-time data and historical data, the operation parameters of the distribution network are analyzed through the distribution network digital twin model to obtain analysis results;

[0088] Preferably, based on real-time data and historical data, the operation parameters of the distribution network are analyzed by the distribution network digital twin model, including:

[0089] The operating parameters are predicted through long short-term memory network.

[0090] Preferably, the method further comprises: training the long short-term memory network by a loss function, wherein the loss function is:

[0091]

[0092] Where L is the loss function, which indicates the difference between the predicted value and the true value. The LSTM model is trained by minimizing the loss function to improve the accuracy of the prediction. t is the time step index, which indicates the tth moment in the data sequence. T is the total length of the time series, which indicates the number of time steps contained in the data set. t+1 is the true value at time step t+1, that is, the actual measured value of the distribution network operating parameters (such as voltage, current, power factor, etc.) at a future moment. is the predicted value at time step t+1, which is the value of the distribution network operation parameter at a future moment predicted by the LSTM model based on the previous input data.

[0093] The present invention includes a parameter change perception module:

[0094] 1) Functional description: The parameter change perception module is responsible for real-time monitoring of the operating parameters of the distribution network, such as voltage, current, power factor, etc. This module improves the accuracy of data through adaptive filtering and data cleaning technology, and uses machine learning technology to predict parameters and detect anomalies.

[0095] 2) Technical implementation: Combine Kalman filtering and LSTM (Long Short-Term Memory Network) model for parameter prediction and anomaly detection. Assume that the input data sequence is X = {x1, x2, ..., x t}, the output of the LSTM model is the predicted value The prediction model is trained by minimizing the following loss function:

[0096]

[0097] Step 104: Based on the analysis results, the distribution network digital twin model is used to detect and locate the operation faults of the distribution network.

[0098] Preferably, based on the analysis results, the operation faults of the distribution network are detected and located by using the distribution network digital twin model, including:

[0099] The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault:

[0100] P(F|T,P)∝P(T|F)·P(P|F)·P(F)

[0101] Among them, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P. This is the result we want to obtain through Bayesian network fusion analysis, which is used to infer the possibility of fault occurrence under the current network topology and parameter state. P(T|F) is the probability of topology change data under the condition of fault occurrence. It represents the probability of a specific change in the topology structure of the distribution network when a fault occurs, which can be obtained based on the statistical relationship between historical fault data and topology changes. P(P|F) is the probability of parameter change data under the condition of fault occurrence. That is, when a fault occurs, the probability of a specific change in the operating parameters of the distribution network can also be learned from historical data. P(F) is the prior probability of fault occurrence, that is, the probability of fault occurrence without considering any topology or parameter change data, which is usually estimated based on historical fault statistics.

[0102] Preferably, the method further comprises updating the Bayesian network:

[0103]

[0104] Among them, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X) describes the possibility of the current observation data under specific parameter settings, which is based on the assumption of the probability model on the data generation process. 1：t ) is the observation data X from time step 1 to t. 1：t The prior probability distribution of the parameter θ under the condition of . This is the probability estimate of the parameter before the new data arrives, reflecting previous knowledge or experience. t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 This is a normalization constant that ensures that the sum of the posterior probability distribution is 1. It is used to adjust the proportional relationship of the probability in the calculation so that the posterior probability distribution satisfies the probability axiom. 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 Under the condition of , the posterior probability distribution of the parameters is obtained. This is the probability distribution of the parameters updated with new data and is used to more accurately estimate the state of the distribution network.

[0105] The present invention includes a fusion analysis module:

[0106] 1) Functional description: The fusion analysis module comprehensively analyzes the outputs from the network topology and parameter change perception modules, and provides fault warnings, performance evaluations, and optimization suggestions. The core of this module is the multivariate data fusion algorithm, which integrates data from different sources and outputs system-level status evaluations.

[0107] 2) Technical implementation: Bayesian network is used for multi-source data fusion and fault inference. Assuming the topology change data is T and the parameter change data is P, the fused fault probability distribution can be expressed as:

[0108] P(F|T,P)∝P(T|F)·P(P|F)·P(F)

[0109] The model detects and locates faults by maximizing the posterior probability.

[0110] The distribution network digital twin dynamic perception method based on multi-network topology and parameter changes provided by the present invention specifically includes the following steps:

[0111] 1. Data collection and preprocessing:

[0112] 1) Description: The sensor network of the distribution network collects parameter data such as voltage, current, and temperature in real time. After being pre-processed by the edge computing node, the data is transmitted to the central processing unit for further processing.

[0113] 2) Technical details: The Z-score method is used to remove outliers, a low-pass filter is used to remove noise, and data sharding and batch processing techniques are used to optimize transmission and computing efficiency.

[0114] 2 Real-time topology update algorithm:

[0115] 1) Description: The real-time topology update algorithm is used to dynamically adjust the topology of the digital twin to ensure that the model can reflect changes in the physical network in a timely manner.

[0116] 2) Technical details: Based on incremental graph algorithms, such as the incremental Dijkstra algorithm, it handles added or removed edges and nodes. For each topology update, the shortest paths of important nodes are recalculated and the model is updated. The specific pseudo code is described as follows:

[0117]

[0118]

[0119] 3 Dynamic parameter monitoring and analysis:

[0120] 1) Description: Utilize historical data and real-time monitoring data to dynamically predict changes in power grid parameters and detect possible anomalies in advance.

[0121] 2) Technical details: A prediction model based on Bayesian update is used for state estimation and anomaly detection. Every time new data arrives, the model updates the parameter distribution through the Bayesian rule to improve the prediction accuracy:

[0122]

[0123] 4System optimization and feedback:

[0124] 1) Description: Based on the real-time monitoring and analysis results, optimization strategies are provided and fed back to the actual operation of the distribution network. Through the closed-loop control mechanism, system performance is continuously optimized.

[0125] 2) Technical details: The cloud computing platform performs complex optimization calculations and uses reinforcement learning algorithms (such as Q-learning) to optimize the operating parameters of the power grid. The system dynamically adjusts strategies based on feedback data to maximize the stability and efficiency of power grid operation.

[0126] Through the above design and implementation methods, the present invention provides a complete, accurate and dynamic distribution network digital twin perception and management solution, which can effectively cope with the complexity and dynamics of the distribution network.

[0127] The method provided by the present invention enables the digital twin of the distribution network to perceive the dynamic changes of network topology and parameters more efficiently and accurately. The present invention not only improves the reliability and stability of the distribution network operation, but also reduces the risk of failures caused by delays or misjudgments, thereby ensuring the continuity and safety of power supply services.

[0128] The present invention is verified through a practical application example of a city distribution network. The implementation results show that the present invention can update the digital twin model in real time when the network topology changes and parameters fluctuate, and provide an optimized operation strategy to ensure the efficient and stable operation of the distribution network.

[0129] The present invention provides a distribution network digital twin dynamic perception method based on multi-network topology and parameter changes, which realizes real-time dynamic perception and accurate modeling of the distribution network through a network topology perception module, a parameter change perception module and a fusion analysis module.

[0130] 1) In the network topology perception module, for the monitoring of the distribution network topology structure, when facing the situation of adding new nodes, node failure or line connection status change, this algorithm executes a dynamic update algorithm based on the shortest path algorithm (such as Dijkstra algorithm) and the maximum flow algorithm. The algorithm first efficiently traverses the topology graph, accurately identifies and removes the failed edges to ensure the accuracy of the network structure; then adds new edges to reflect the changes in the actual network topology; finally, recalculates the shortest path of each node to provide key path information for subsequent network analysis.

[0131] 2) In the parameter change perception module, in the parameter prediction and anomaly detection stages, the LSTM (Long Short-Term Memory Network) model in machine learning technology is used to minimize a specific loss function during the training process. The model parameters are continuously adjusted to achieve accurate prediction and timely anomaly detection of distribution network operating parameters (such as voltage, current, power factor, etc.).

[0132] 3) In the fusion analysis module, this method calculates the fused fault probability distribution (P(F|T, P)∝P(T|F)·P(P|F)·P(F)) based on the Bayesian network principle. In the fault inference process, the maximum a posteriori probability (MAP) algorithm is used for fault detection and location. In the specific operation, all possible fault states are traversed, the a posteriori probability under each state is calculated, and the state with the largest probability is selected as the inference result, so as to accurately infer the possibility and location of the fault.

[0133] The modules in the present invention work together to ensure that the distribution network digital twin can accurately reflect the dynamic changes of the actual power grid and improve the safety and operating efficiency of the distribution network.

[0134] Figure 3 The figure is a structural diagram of a distribution network digital twin dynamic perception system according to a preferred embodiment of the present invention.

[0135] like Figure 3 As shown, the present invention provides a distribution network digital twin dynamic perception system, the system comprising:

[0136] The network topology sensing unit 301 is used to obtain real-time data of the distribution network operation, monitor the node changes and line status changes of the distribution network based on the real-time data; and update the distribution network topology structure diagram in the established distribution network digital twin model in real time based on the node changes and line status changes;

[0137] Preferably, the network topology sensing unit 301 is also used to pre-process the real-time data, including: clearing abnormal values ​​in the real-time data, and removing noise from the real-time data by using a low-pass filter.

[0138] Preferably, the network topology perception unit 301 is used to update the distribution network topology diagram in the established distribution network digital twin model in real time based on node changes and line state changes, and is also used to:

[0139] Based on the dynamic update algorithm, the topology diagram of the distribution network is traversed to remove the failed edges and add new edges;

[0140] Calculate the shortest path for each node in the updated distribution network topology graph.

[0141] The parameter change sensing unit 302 is used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on real-time data and historical data to obtain analysis results;

[0142] Preferably, the parameter change sensing unit 302 is used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on real-time data and historical data, and is also used to:

[0143] The operating parameters are predicted through long short-term memory network.

[0144] Preferably, the parameter change sensing unit 302 is further used to train the long short-term memory network through a loss function, and the loss function is:

[0145]

[0146] Where L is the loss function, which indicates the difference between the predicted value and the true value. The LSTM model is trained by minimizing the loss function to improve the accuracy of the prediction. t is the time step index, which indicates the tth moment in the data sequence. T is the total length of the time series, which indicates the number of time steps contained in the data set. t+1 is the true value at time step t+1, that is, the actual measured value of the distribution network operating parameters (such as voltage, current, power factor, etc.) at a future moment. is the predicted value at time step t+1, which is the value of the distribution network operation parameter at a future moment predicted by the LSTM model based on the previous input data.

[0147] The fusion analysis unit 303 is used to detect and locate operation faults of the distribution network through the distribution network digital twin model based on the analysis results.

[0148] Preferably, the fusion analysis unit 303 is used to detect and locate the operation fault of the distribution network through the distribution network digital twin model based on the analysis result, and is also used to:

[0149] The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault:

[0150] P(F|T,P)∝P(T|F)·P(P|F)·P(F)

[0151] Among them, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P. This is the result we want to obtain through Bayesian network fusion analysis, which is used to infer the possibility of fault occurrence under the current network topology and parameter state. P(T|F) is the probability of topology change data under the condition of fault occurrence. It represents the probability of a specific change in the topology structure of the distribution network when a fault occurs, which can be obtained based on the statistical relationship between historical fault data and topology changes. P(P|F) is the probability of parameter change data under the condition of fault occurrence. That is, when a fault occurs, the probability of a specific change in the operating parameters of the distribution network can also be learned from historical data. P(F) is the prior probability of fault occurrence, that is, the probability of fault occurrence without considering any topology or parameter change data, which is usually estimated based on historical fault statistics.

[0152] Preferably, the fusion analysis unit 303 is also used to update the Bayesian network:

[0153]

[0154] Among them, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X) describes the possibility of the current observation data under specific parameter settings, which is based on the assumption of the probability model on the data generation process. 1：t ) is the observation data X from time step 1 to t. 1：t The prior probability distribution of the parameter θ under the condition of . This is the probability estimate of the parameter before the new data arrives, reflecting previous knowledge or experience. t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 This is a normalization constant that ensures that the sum of the posterior probability distribution is 1. It is used to adjust the proportional relationship of the probability in the calculation so that the posterior probability distribution satisfies the probability axiom. 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 Under the condition of , the posterior probability distribution of the parameters is obtained. This is the probability distribution of the parameters updated with new data and is used to more accurately estimate the state of the distribution network.

[0155] The present invention provides a computer-readable storage medium, which stores a computer program, and the computer program is used to execute a dynamic perception method of a distribution network digital twin.

[0156] The present invention provides an electronic device, the electronic device comprising: a processor and a memory; wherein:

[0157] Memory, a memory for storing processor executable instructions;

[0158] A processor is used to read executable instructions from a memory and execute the instructions to implement a dynamic perception method of a distribution network digital twin.

[0159] A distribution network digital twin dynamic perception system according to a preferred embodiment of the present invention corresponds to a distribution network digital twin dynamic perception method according to another preferred embodiment of the present invention, and will not be described in detail here.

[0160] It will be appreciated by those skilled in the art that embodiments of the present invention may be provided as methods, systems, or computer program products. Therefore, the present invention may take the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware. Moreover, the present invention may take the form of a computer program product implemented on one or more computer-usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer-usable program codes. The schemes in the embodiments of the present invention may be implemented in various computer languages, for example, object-oriented programming language Java and literal scripting language JavaScript, etc.

[0161] The present invention is described with reference to flowcharts and / or block diagrams of methods, devices (systems), and computer program products according to embodiments of the present invention. It should be understood that each process and / or block in the flowchart and / or block diagram, as well as the combination of processes and / or blocks in the flowchart and / or block diagram, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, a special-purpose computer, an embedded processor, or other programmable data processing device to produce a machine, so that the instructions executed by the processor of the computer or other programmable data processing device generate instructions for implementing the processes in the flowchart and / or block diagram. Figure 1 A process or multiple processes and / or boxes Figure 1 A device that provides the functions specified in a block or multiple blocks.

[0162] These computer program instructions may also be stored in a computer-readable memory capable of directing a computer or other programmable data processing device to operate in a specific manner, so that the instructions stored in the computer-readable memory produce an article of manufacture comprising an instruction device, which implements the process Figure 1 A process or multiple processes and / or boxes Figure 1 A function specified in one or more boxes.

[0163] These computer program instructions can also be loaded onto a computer or other programmable data processing device so that a series of operating steps are executed on the computer or other programmable device to produce a computer-implemented process, thereby providing instructions for implementing the process. Figure 1 A process or multiple processes and / or boxes Figure 1 The steps for the functions specified in one or more boxes.

[0164] Although the preferred embodiments of the present invention have been described, those skilled in the art may make other changes and modifications to these embodiments once they have learned the basic creative concept. Therefore, the appended claims are intended to be interpreted as including the preferred embodiments and all changes and modifications that fall within the scope of the present invention.

[0165] Obviously, those skilled in the art can make various changes and modifications to the present invention without departing from the spirit and scope of the present invention. Thus, if these modifications and variations of the present invention fall within the scope of the claims of the present invention and their equivalents, the present invention is also intended to include these modifications and variations.

[0166] The invention has been described above with reference to a few embodiments. However, it is readily apparent to a person skilled in the art that other embodiments than the ones disclosed above are equally within the scope of the invention, as defined by the appended patent claims.

[0167] Generally, all terms used in the claims are to be interpreted according to their ordinary meaning in the technical field, unless explicitly defined otherwise herein. All references to "a / said / the [means, components, etc.]" are to be openly interpreted as at least one instance of the means, components, etc., unless explicitly stated otherwise. The steps of any method disclosed herein do not necessarily have to be performed in the exact order disclosed, unless explicitly stated otherwise.

Claims

1. A distribution network digital twin dynamic perception method, the method comprising: Acquire real-time data of the operation of the distribution network, and monitor node changes and line status changes of the distribution network based on the real-time data; Based on the node changes and the line state changes, a distribution network topology diagram in the established distribution network digital twin model is updated in real time; Based on the real-time data and historical data, the operating parameters of the distribution network are analyzed by the distribution network digital twin model to obtain analysis results; Based on the analysis results, the distribution network digital twin model is used to detect and locate operational faults in the distribution network.

2. The method according to claim 1, further comprising preprocessing the real-time data, comprising: The outliers in the real-time data are removed, and the noise of the real-time data is removed by a low-pass filter.

3. The method according to claim 1, wherein the real-time updating of the distribution network topology diagram in the established distribution network digital twin model based on the node change and the line state change comprises: Traversing the distribution network topology diagram based on a dynamic update algorithm, removing failed edges and adding new edges; The shortest path of each node in the updated distribution network topology diagram is calculated.

4. The method according to claim 1, wherein the operating parameters of the distribution network are analyzed by the distribution network digital twin model based on the real-time data and historical data, comprising: The operating parameters are predicted by a long short-term memory network.

5. The method according to claim 4, further comprising: The long short-term memory network is trained by a loss function, and the loss function is: Among them, L is the loss function, t is the time step index, which means the tth moment in the data sequence, T is the total length of the time series, and x t+1 is the true value at time step t+1, X t+1 is the predicted value at time step t+1.

6. The method according to claim 5, wherein the detecting and locating of the operation fault of the distribution network through the distribution network digital twin model based on the analysis result comprises: The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault: P(F|T,P)∝P(T|F)·P(P|F)·P(F) Among them, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P; P(T|F) is the probability of topology change data occurrence under the condition of fault occurrence; P(P|F) is the probability of parameter change data occurrence under the condition of fault occurrence; P(F) is the prior probability of fault occurrence, which is the probability of fault occurrence without considering any topology or parameter change data.

7. The method according to claim 6, further comprising updating the Bayesian network: in, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of , the prior probability distribution of parameter θ; P(X t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 The probability of P(θ|X 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 The posterior probability distribution of the parameter under the condition of .

8. A distribution network digital twin dynamic perception system, the system comprising: A network topology sensing unit, used to obtain real-time data of the operation of the distribution network, and monitor node changes and line status changes of the distribution network based on the real-time data; Based on the node changes and the line state changes, a distribution network topology diagram in the established distribution network digital twin model is updated in real time; A parameter change sensing unit, used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on the real-time data and historical data, and obtain analysis results; A fusion analysis unit is used to detect and locate operation faults of the distribution network through the distribution network digital twin model based on the analysis results.

9. The system according to claim 8, wherein the network topology sensing unit is further used to pre-process the real-time data, including: The outliers in the real-time data are removed, and the noise of the real-time data is removed by a low-pass filter.

10. The system according to claim 8, wherein the network topology perception unit is used to update the distribution network topology structure diagram in the established distribution network digital twin model in real time based on the node changes and the line state changes, and is also used to: Traversing the distribution network topology diagram based on a dynamic update algorithm, removing failed edges and adding new edges; The shortest path of each node in the updated distribution network topology diagram is calculated.

11. The system according to claim 8, wherein the parameter change sensing unit is used to analyze the operating parameters of the distribution network through the distribution network digital twin model based on the real-time data and historical data, and is also used to: The operating parameters are predicted by a long short-term memory network.

12. The system according to claim 11, wherein the parameter change sensing unit is further used to train the long short-term memory network through a loss function, wherein the loss function is: in, L is the loss function, t is the time step index, which means the tth moment in the data sequence, T is the total length of the time series, x t+1 is the true value at time step t+1, is the predicted value at time step t+1.

13. The system according to claim 12, wherein the fusion analysis unit is used to detect and locate the operation fault of the distribution network through the distribution network digital twin model based on the analysis result, and is also used to: The topology change data T of the distribution network topology diagram and the operation parameter change data P are integrated through the Bayesian network to detect and locate the operation fault: P(F|T,P)∝P(T|F)·P(P|F)·P(F) in, P(F|T, P) is the probability distribution of fault occurrence under the given topology change data T and parameter change data P; P(T|F) is the probability of topology change data occurrence under the condition of fault occurrence; P(P|F) is the probability of parameter change data occurrence under the condition of fault occurrence; P(F) is the prior probability of fault occurrence, which is the probability of fault occurrence without considering any topology or parameter change data.

14. The system according to claim 13, wherein the fusion analysis unit is further configured to update the Bayesian network: in, P(X t+1 |θ) is the data X observed at time step t+1 under the condition of given parameter θ t+1 The probability of P(θ|X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of , the prior probability distribution of parameter θ; P(X t+1 |X 1：t ) is the observation data X from time step 1 to t. 1：t Under the condition of t+1 The probability of P(θ|X 1：t+1 ) is the observation data X from time step 1 to t+1 1：t+1 The posterior probability distribution of the parameter under the condition of .

15. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is used to execute the method according to any one of claims 1 to 7.

16. An electronic device, characterized in that: The electronic device comprises: a processor and a memory; wherein, The memory is a memory for storing instructions executable by the processor; The processor is configured to read the executable instructions from the memory and execute the instructions to implement the method according to any one of claims 1 to 7.

Images