Power distribution network dispatching management and control method based on situation awareness technology

Abstract

The invention relates to the technical field of artificial intelligence, and discloses a situation awareness technology-based power distribution network dispatching management and control method, which comprises the following steps of: acquiring multi-source data of a power distribution network; constructing an initial power distribution network model and time sequence features of the power distribution network based on the multi-source data, and generating feature vectors of nodes in the initial power distribution network model; vector fusion is carried out on the feature vectors through an improved adjacency matrix to obtain local fusion features, and the improved adjacency matrix dynamically adjusts the connection weight between nodes through a dynamic influence function and corrects the node degree based on the load difference; generating a dynamic weight of the local fusion feature according to the time sequence feature, predicting a load change trend of the power distribution network after performing feature dimension reduction on the local fusion feature according to the dynamic weight, generating a scheduling strategy of the power distribution network in combination with safety constraint conditions of voltage limitation and current limitation, and performing scheduling management and control on the power distribution network according to the scheduling strategy. The efficiency of power distribution network dispatching management and control can be improved.

Description

Technical Field

[0001] The present invention relates to the field of artificial intelligence technology, and in particular to a distribution network dispatching and control method using situation awareness technology. Background Art

[0002] With the continuous development of the power system, the importance of the distribution network in the entire power supply system has become increasingly prominent. However, the current dispatching and control of the distribution network faces many challenges. On the one hand, the structure of the distribution network is becoming more and more complex, with many nodes and changing connection relationships. Traditional dispatching and control methods are difficult to fully and accurately grasp its operating status; on the other hand, the factors affecting the operation of the distribution network are diverse. In addition to the equipment status and power load changes, the impact of environmental factors (such as temperature, humidity, light, etc.) on the distribution network is also gradually being paid attention to. The existing dispatching and control methods are insufficient in comprehensively considering these multi-source factors.

[0003] At the same time, in terms of data processing, a large amount of structural data, environmental data, real-time operation data and equipment data from different channels are difficult to integrate and analyze efficiently, resulting in the inability to timely and accurately predict load change trends, which in turn affects the scientificity and rationality of the dispatching strategy, making the efficiency of distribution network dispatching and control low and difficult to meet the strict requirements of modern power systems for reliability and stability. Therefore, how to improve the efficiency of distribution network dispatching and control has become an urgent problem to be solved. Summary of the invention

[0004] The present invention provides a distribution network dispatching and control method based on situation awareness technology, the main purpose of which is to solve the problem of low efficiency in distribution network dispatching and control.

[0005] To achieve the above-mentioned purpose, the present invention provides a distribution network dispatching and control method using situation awareness technology, comprising: Collecting multi-source data of the distribution network, wherein the multi-source data includes: structural data, environmental data, real-time operation data and equipment data; Constructing an initial distribution network model of the distribution network based on the structure data and the device data, performing vector transformation on the attribute information corresponding to the nodes in the initial distribution network model to obtain the feature vectors of the nodes; Performing vector fusion on the feature vectors using an improved adjacency matrix to obtain local fusion features of nodes in the distribution network, wherein the improved adjacency matrix dynamically adjusts the connection weights between nodes through a dynamic influence function and corrects the node degrees based on load differences; Generate a time series feature of the distribution network according to the environmental data and the real-time operation data, generate a dynamic weight of the local fusion feature according to the time series feature, and perform feature dimension reduction on the local fusion feature according to the dynamic weight to obtain a dimension reduction feature of a node in the distribution network; The load change trend is predicted according to the dimensionality reduction features, and the dispatching strategy of the distribution network is generated in combination with the safety constraints of voltage limitation and current limitation, and the distribution network is dispatched and controlled according to the dispatching strategy.

[0006] Optionally, the collecting of multi-source data of the power distribution network includes: Collecting data from nodes of the distribution network based on a distributed sensor network to obtain initial data of the distribution network; Performing data cleaning on the initial data to obtain cleaned data of the distribution network; The cleaned data is classified to obtain classified data of the distribution network, and it is determined that the classified data is multi-source data of the distribution network.

[0007] Optionally, constructing an initial distribution network model of the distribution network based on the structure data and the device data includes: Determine nodes and connection relationships of the distribution network according to the structure data and the device data; Generating a topological structure of the distribution network according to the nodes and the connection relationships; An initial distribution network model of the distribution network is generated based on a graph theory method and the topological structure.

[0008] Optionally, performing vector conversion on the attribute information corresponding to the node in the initial distribution network model to obtain the feature vector of the node includes: Extracting attribute information of nodes in the initial distribution network model, wherein the attribute information includes: voltage level, load type and equipment status; Numerical encoding is performed on the attribute information to obtain numerical features of the attribute information; The numerical features are combined to obtain a feature vector of the node.

[0009] Optionally, the step of performing vector fusion on the feature vectors by using an improved adjacency matrix to obtain local fusion features of nodes in the distribution network includes: The feature vector is multiplied by the improved adjacency matrix to obtain the local fusion features of the nodes in the distribution network, wherein the improved adjacency matrix is: ; in, is the improved adjacency matrix, is the degree matrix, is the adjacency matrix, Is a node The degree of Indicates whether there is an edge connecting the nodes and nodes , Is a node and nodes The improved connection weights between Representation and Node The set of all adjacent nodes, Is a node and nodes The dynamic influence function between Is a node and nodes The load change difference between Is a node and nodes The difference in environmental data between Is a node and nodes The similarity of device status between Indicates The node ID of each node, It is The node ID of each node, is a constant coefficient that adjusts the weight of the adjacency matrix, is the total number of nodes, It is a constant coefficient that adjusts the effect of load difference on the degree.

[0010] Optionally, generating the dynamic weight of the local fusion feature according to the temporal feature includes: Determining key influencing factors of the local fusion feature based on the time series feature; A dynamic weight of the local fusion feature is generated based on the key influencing factors.

[0011] Optionally, performing feature dimensionality reduction on the local fusion feature according to the dynamic weight to obtain the dimensionality reduction feature of the node in the distribution network includes: Multiplying the dynamic weight by the local fusion feature element by element to obtain a weighted local fusion feature; The weighted local fusion features are subjected to dimensionality reduction processing based on a principal component analysis algorithm to obtain dimensionality reduction features of the nodes in the distribution network.

[0012] Optionally, predicting the load change trend according to the dimensionality reduction feature includes: Inputting the dimension reduction features into a pre-trained load change trend prediction model; The load change trend prediction model is used to perform pattern recognition and trend inference on the dimensionality reduction features to obtain the load change trend of the distribution network, wherein the load change trend includes: the time of load rise, fall, stable state information, load peak and valley value occurrence and corresponding load values ​​under different time scales.

[0013] Optionally, it is characterized in that the method for generating the dispatching strategy is as follows: determining the load demand and power supply capacity of the distribution network in different time periods according to the load change trend; A dispatching strategy for the distribution network is generated according to the load demand and the power supply capacity in combination with the safety constraints of voltage limitation and current limitation.

[0014] Optionally, the performing dispatching and controlling of the distribution network according to the dispatching strategy includes: Performing task analysis on the dispatching strategy to obtain the dispatching task of the distribution network; Performing instruction conversion on the dispatching task to obtain a control instruction of the distribution network; The control instruction is sent to the control device of the power distribution network based on a preset communication network.

[0015] The present invention constructs an initial distribution network model based on structural data and equipment data, uses graph theory combined with topological structure to accurately simulate the operation characteristics of the distribution network, and at the same time, performs vector conversion on attribute information, extracts attribute information such as voltage level, load type and equipment status, generates feature vectors through numerical coding and feature combination, represents complex attribute information in a concise and computable vector form, introduces an improved adjacency matrix for vector fusion, and constructs the improved adjacency matrix taking into account dynamic factors such as load change differences, environmental data differences and equipment status similarity between nodes. The improved connection weight and degree calculation method enable the matrix to more accurately reflect the actual association between nodes, and the obtained local fusion features can more comprehensively and accurately reflect the state of the node in the distribution network and the relationship with other nodes, thereby enhancing the modeling capability of complex relationships in the distribution network. The dispatching strategy formulated based on load change trend and power supply capacity, combined with safety constraints, can achieve optimal allocation of power resources. Therefore, the present invention proposes a distribution network dispatching and control method based on situation awareness technology, which can solve the problem of low dispatching and control efficiency of the distribution network. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 A schematic diagram of a flow chart of a distribution network dispatching and control method using a situation awareness technology provided by an embodiment of the present invention; The realization of the purpose, functional features and advantages of the present invention will be further explained in conjunction with embodiments and with reference to the accompanying drawings. DETAILED DESCRIPTION

[0017] It should be understood that the specific embodiments described herein are only used to explain the present invention, and are not used to limit the present invention.

[0018] The embodiment of the present application provides a distribution network dispatching and control method using situational awareness technology. The execution subject of the distribution network dispatching and control method using situational awareness technology includes but is not limited to at least one of the electronic devices such as a server and a terminal that can be configured to execute the method provided by the embodiment of the present application. In other words, the distribution network dispatching and control method using situational awareness technology can be executed by software or hardware installed on a terminal device or a server device. The server includes but is not limited to: a single server, a server cluster, a cloud server or a cloud server cluster, etc. The server can be an independent server, or it can be a cloud server that provides basic cloud computing services such as cloud services, cloud databases, cloud computing, cloud functions, cloud storage, network services, cloud communications, middleware services, domain name services, security services, and artificial intelligence platforms.

[0019] Reference Figure 1 As shown, it is a flow chart of a distribution network dispatching and control method of a situation awareness technology provided by an embodiment of the present invention. In this embodiment, the distribution network dispatching and control method of the situation awareness technology includes: S1. Collect multi-source data of the distribution network, wherein the multi-source data includes: structural data, environmental data, real-time operation data and equipment data.

[0020] In an embodiment of the present invention, the collecting of multi-source data of the power distribution network includes: Collecting data from nodes of the distribution network based on a distributed sensor network to obtain initial data of the distribution network; Performing data cleaning on the initial data to obtain cleaned data of the distribution network; The cleaned data is classified to obtain classified data of the distribution network, and it is determined that the classified data is multi-source data of the distribution network.

[0021] In detail, a sensor network distributed at each node of the distribution network is used to collect data. These sensors are responsible for monitoring and recording various information at the nodes, and the collected information is integrated to form an initial data set of the distribution network.

[0022] In detail, situational awareness technology is an advanced technology that integrates multi-source data to achieve comprehensive, real-time monitoring and accurate prediction of the operating status of complex systems, and plays a key role in the field of distribution network dispatching and control.

[0023] In detail, various sensors, such as voltage sensors, current sensors, temperature sensors, humidity sensors, etc., are installed at key node locations such as transformers, pole towers, and distribution boxes in the distribution network. The sensors transmit the collected data to the data collection terminal through wired or wireless communication technologies, such as optical fiber, ZigBee, 4G / 5G, etc., and the terminal summarizes the initial data.

[0024] For example, the voltage sensor collects the voltage value of a node at a specific moment as 225V; the current sensor records the current of the node as 50A at the same moment; and the temperature sensor measures the surface temperature of the device at the node as 30°C.

[0025] Furthermore, the initial data may contain noise, erroneous data, duplicate data and other problems. Data cleaning is to process these problematic data, remove or correct inaccurate, incomplete and duplicate data, and improve data quality.

[0026] In detail, a variety of data cleaning algorithms and techniques are used. Among them, the 3σ criterion based on statistical methods is used to identify and process outliers. If a data point deviates from the mean by more than 3 times the standard deviation, it is determined to be an outlier and is corrected or deleted. For duplicate data, a data comparison algorithm is used to find and delete duplicate records based on the unique identifier or feature combination of the data. Missing values ​​are filled with mean filling, median filling or prediction methods based on machine learning.

[0027] For example, if the current data at a certain moment shows an obvious abnormal value of 1000A (the normal range is 0-100A), it will be identified as an abnormality using the 3σ criterion and may be corrected to a value near the current mean of the period; if there are multiple identical voltage data records, the duplicate records will be deleted.

[0028] Furthermore, according to the properties of the data, such as its nature, source or purpose, the cleaned data is divided into different categories, such as structural data, environmental data, real-time operation data and equipment data, so as to facilitate subsequent targeted processing and analysis.

[0029] In detail, formulate data classification rules, such as classification based on the physical meaning of the data and the monitored object.

[0030] In detail, the cleaned data is automatically classified according to the classification rules by writing program scripts or using data management software.

[0031] For example: data such as node connection relationships, line parameters, lengths of each line in the distribution network, and conductor models are classified as structural data; real-time collected active power and reactive power data are classified as real-time operation data; and equipment model, manufacturing year, and other data are classified as equipment data.

[0032] S2. constructing an initial distribution network model of the distribution network based on the structural data and the equipment data, performing vector transformation on the attribute information corresponding to the nodes in the initial distribution network model, and obtaining the feature vectors of the nodes.

[0033] In an embodiment of the present invention, constructing the initial distribution network model of the distribution network based on the structure data and the device data includes: Determine nodes and connection relationships of the distribution network according to the structure data and the device data; Generating a topological structure of the distribution network according to the nodes and the connection relationships; An initial distribution network model of the distribution network is generated based on a graph theory method and the topological structure.

[0034] Specifically, the structural data contains the topological information of the distribution network, such as line connections, substation locations, etc. The equipment data records the installation locations and connection methods of various types of power equipment (such as transformers, switches, etc.). By analyzing these data, it is possible to identify the key locations in the distribution network as nodes and the electrical connection relationships between nodes.

[0035] In detail, the structural data and equipment data are sorted and analyzed, and data associations are established.

[0036] In detail, the connection of the node is determined by the starting and ending point information of the line; based on the access point information of the device, the node location of the device and the connection with other nodes are clarified, and database query and data matching algorithms can be used to complete this process.

[0037] For example, the structural data records that line L1 connects nodes N1 and N2, and the equipment data shows that transformer T1 is installed on node N1. This determines nodes N1 and N2 and the connection relationship between them, and clarifies the association between transformer T1 and node N1.

[0038] In detail, after determining the nodes and connection relationships, this information is represented in an intuitive way to form the topological structure of the distribution network, where the topological structure describes the interconnection mode and hierarchical relationship between nodes and is an abstract representation of the overall architecture of the distribution network.

[0039] In detail, the topological structure can be represented by a graph, where nodes are represented by vertices and connections are represented by edges.

[0040] Furthermore, relevant algorithms and data structures in graph theory, such as adjacency matrix and adjacency list, are used to store and represent topological structures. These algorithms can be implemented through programming to convert nodes and connection relationships into corresponding graph structures.

[0041] In detail, assuming that there are nodes N1, N2, and N3, and the connection relationships are N1-N2 and N2-N3, when the adjacency matrix is ​​used to represent the topological structure, the elements in the matrix are assigned 1 (connected) or 0 (not connected) according to the connection status between the nodes.

[0042] In detail, graph theory is a mathematical method used to study the properties and applications of graphs. The concepts and algorithms in graph theory are used to model the distribution network in combination with the generated topological structure. Through graph theory, the nodes and connection relationships of the distribution network can be mathematically described and analyzed, thereby constructing an initial model that can reflect the operating characteristics of the distribution network.

[0043] In detail, the physical characteristics of the distribution network (such as voltage, current, power, etc.) are associated with the nodes and edges of the graph to establish mathematical equations and models. Professional modeling software or programming languages ​​(such as Python combined with the NetworkX library) can be used to implement graph theory algorithms and model construction.

[0044] For example, the Dijkstra algorithm is used to calculate the shortest path between two points in the distribution network to optimize the power transmission path. The electrical characteristics of the distribution network are described by establishing equations for node voltage and current.

[0045] In general, the initial distribution network model generated based on graph theory and topological structure can accurately simulate the operating status and characteristics of the distribution network, provide effective model support for subsequent load forecasting, scheduling strategy formulation, etc., and improve the management and operation efficiency of the distribution network.

[0046] In the embodiment of the present invention, the vector conversion of the attribute information corresponding to the node in the initial distribution network model to obtain the feature vector of the node includes: Extracting attribute information of nodes in the initial distribution network model, wherein the attribute information includes: voltage level, load type and equipment status; Numerical encoding is performed on the attribute information to obtain numerical features of the attribute information; The numerical features are combined to obtain a feature vector of the node.

[0047] In detail, the key attribute information related to the node is identified and extracted from the initial distribution network model. The voltage level reflects the voltage level of the node in the distribution network; the load type (such as residential load, industrial load, etc.) reflects the characteristics of the load connected to the node; the equipment status (such as normal operation, fault, etc.) describes the working condition of the equipment on the node.

[0048] In detail, in the initial distribution network model, through data query and screening operations, the corresponding attribute information is extracted according to the identification information of the node, and this task can be completed using a database query statement or a data processing program.

[0049] For example, for node N1, the voltage level extracted from the model is 10 kV, the load type is residential load, and the equipment status is normal operation.

[0050] In detail, since attribute information usually exists in text or classification form, in order to facilitate computer processing and model analysis, it is necessary to convert these attribute information into numerical form. Through numerical coding, different attribute values ​​can be mapped to specific numerical values ​​to form the numerical characteristics of attribute information.

[0051] Specifically, different encoding methods are used for different types of attribute information. For voltage levels, their numerical representation can be used directly; for load types, one-hot encoding can be used to represent each load type with a binary vector, where only the corresponding type position is 1 and the rest are 0; for equipment status, 0 and 1 can be used to represent normal and fault status respectively.

[0052] In detail, it is assumed that there are three types of loads: residential load, industrial load, and commercial load. The residential load is represented by [1, 0, 0], the industrial load is represented by [0, 1, 0], and the commercial load is represented by [0, 0, 1]. If the load type of node N1 is residential load, its encoded numerical feature is [1, 0, 0].

[0053] In detail, the numerical features of the encoded attribute information are combined in a certain order to form a vector, which is the feature vector of the node. The feature vector can comprehensively reflect the various attribute information of the node, which is convenient for unified processing and analysis in subsequent models.

[0054] Specifically, the voltage level values, the load type coded vectors, the device status values, etc. are arranged in a predetermined order to form a vector. For example, the voltage level is arranged first, then the load type coded vectors, and finally the device status values.

[0055] Assume that the voltage level of node N1 is 10kV, the load type is coded as [1, 0, 0], and the equipment status is normal (indicated by 1), then its eigenvector is [10, 1, 0, 0, 1].

[0056] In general, the generated node feature vector can comprehensively and concisely represent the attribute information of the node. S3. Using an improved adjacency matrix, the feature vectors are fused to obtain local fusion features of the nodes in the distribution network.

[0057] In an embodiment of the present invention, the step of performing vector fusion on the feature vectors using an improved adjacency matrix to obtain local fusion features of nodes in the distribution network includes: The characteristic vector is subjected to matrix multiplication operation with the improved adjacency matrix to obtain local fusion features of the nodes in the distribution network.

[0058] In detail, matrix multiplication operation is performed on the eigenvector and the improved adjacency matrix. The construction of the improved adjacency matrix considers various factors between nodes, such as load change difference, environmental data difference and equipment status similarity, so as to more accurately reflect the association between nodes.

[0059] In detail, the improved adjacency matrix is: ; in, is the improved adjacency matrix, is the degree matrix, is the adjacency matrix, Is a node The degree of Indicates whether there is an edge connecting the nodes and nodes , Is a node and nodes The improved connection weights between Representation and Node The set of all adjacent nodes, Is a node and nodes The dynamic influence function between Is a node and nodes The load change difference between Is a node and nodes The difference in environmental data between Is a node and nodes The similarity of device status between Indicates The node ID of each node, It is The node ID of each node, is a constant coefficient that adjusts the weight of the adjacency matrix, is the total number of nodes, It is a constant coefficient that adjusts the effect of load difference on the degree.

[0060] In detail, the degree matrix ,in Is a node The degree matrix is ​​a diagonal matrix with the elements on the diagonal Representation Node The degree of , all non-diagonal elements are 0.

[0061] In detail, the degree reflects how closely a node is connected to other nodes. Degree The calculation formula is ,in, Representation and Node The set of all adjacent nodes. The degree calculation here not only considers the connection weights between nodes , and also introduces the load change difference and the adjustment factor , so that the degree of a node can more dynamically reflect its importance and connectivity in the distribution network.

[0062] In detail, the adjacency matrix Used to represent the connection relationship between nodes.

[0063] In detail, the traditional adjacency matrix elements Usually indicates whether there are edges connecting nodes and nodes (like Indicates that there is a connection. Indicates no connection), and here the improved connection weight is used ,in, is a constant coefficient for adjusting the weight of the adjacency matrix, which is used to adjust the overall weight of the adjacency matrix. Is a node and nodes The dynamic influence function between nodes takes into account and nodes The load change difference between , environmental data differences Similarity to device status , so that the connection weights between nodes can be adjusted according to these dynamic factors, reflecting the actual relationship between nodes more accurately.

[0064] In detail, the improved adjacency matrix dynamically adjusts the connection weights between nodes through a dynamic influence function and corrects the node degrees based on load differences.

[0065] In detail, assume that there are three nodes in the distribution network, namely , the feature vector of the node .

[0066] First determine the adjacency matrix Elements :Assume that node 1 is connected to node 2, node 2 is connected to node 3, and node 1 is not connected to node 3, then the traditional adjacency matrix , .

[0067] Then calculate the improved connection weights : Assumption ,but . Calculate the other The value of , get the adjacency matrix .

[0068] Then calculate the degree matrix : Assumption , for node 1, ,but . The same calculation and ,get .

[0069] Finally, calculate the improved adjacency matrix .

[0070] Furthermore, the feature vector With the improved adjacency matrix Multiply, such as ,but , It is the local fusion feature of the node.

[0071] In detail, the feature vectors are fused using an improved adjacency matrix, taking into account dynamic factors such as load change differences, environmental data differences, and equipment status similarities among nodes, so that the obtained local fusion features can more comprehensively and accurately reflect the actual status of the node in the distribution network and its association with other nodes.

[0072] S4. Generate time series characteristics of the distribution network according to the environmental data and the real-time operation data, generate dynamic weights of the local fusion characteristics according to the time series characteristics, and perform feature dimensionality reduction on the local fusion characteristics according to the dynamic weights to obtain dimensionality reduction characteristics of the nodes in the distribution network.

[0073] In the embodiment of the present invention, the step of generating the dynamic weight of the local fusion feature according to the temporal feature includes: Determining key influencing factors of the local fusion feature based on the time series feature; A dynamic weight of the local fusion feature is generated based on the key influencing factors.

[0074] In detail, environmental data (such as temperature, humidity, light, etc.) and real-time operation data (such as voltage, current, power, etc.) are sequence data that change over time. By processing and analyzing these data, the time series characteristics contained therein are extracted. These characteristics can reflect the operating status and changing trends of the distribution network at different time points.

[0075] In detail, time series analysis methods can be used, such as sliding window method, autoregressive integrated moving average model (ARIMA), etc. For example, using the sliding window method, environmental data and real-time operation data are divided into certain time windows (such as 1 hour, 24 hours, etc.), and the statistical characteristics of the data in each window, such as mean, variance, maximum value, minimum value, etc., are calculated as time series features.

[0076] Specifically, assume that the temperature in the environmental data is recorded once an hour, and the power in the real-time operation data is also recorded once an hour. Use a 24-hour sliding window to calculate the mean temperature and variance of the power in each window. The obtained statistical values ​​are time series features. For example, the mean temperature in the first 24-hour window is 25°C, and the variance of the power is 100 (kW²).

[0077] In detail, after obtaining the time series characteristics of the distribution network, the correlation between these characteristics and the local fusion characteristics is analyzed to find out the factors that have a greater impact on the local fusion characteristics. These key influencing factors can reveal the intrinsic connection between the operating status of the distribution network and the local fusion characteristics, and provide a basis for generating dynamic weights.

[0078] In detail, methods such as correlation analysis and causal analysis can be used. For example, by calculating the Pearson correlation coefficient between the time series feature and the local fusion feature, the time series feature with a larger absolute value of the correlation coefficient is found as the key influencing factor.

[0079] In detail, through correlation analysis, it is found that the correlation coefficient between the time series characteristics of temperature and a component in the local fusion feature is 0.8, the correlation coefficient between the time series characteristics of power and the component is 0.6, and the correlation coefficient between the time series characteristics of humidity and the component is 0.2. It can be considered that temperature and power are the key influencing factors of the local fusion feature.

[0080] In detail, according to the determined key influencing factors, corresponding weights are assigned to each component of the local fusion feature, and these weights are dynamically adjusted as the key influencing factors change to reflect the impact of the operating status of the distribution network at different times on the local fusion feature.

[0081] In detail, linear regression and machine learning algorithms (such as neural networks, decision trees, etc.) can be used to establish the mapping relationship between key influencing factors and dynamic weights. For example, a linear regression model is used, with key influencing factors as input and the weights of local fusion features as output, and the weight calculation formula is obtained by training the model.

[0082] In detail, assuming that the dynamic weights are obtained by linear regression With temperature and power The relationship is , when the temperature ,power When the dynamic weight is calculated .

[0083] In detail, the generated dynamic weights can be dynamically adjusted according to the real-time operating status of the distribution network, so that the local fusion features can more accurately reflect the actual situation of the distribution network in subsequent processing, thereby improving the accuracy of the dispatching and control of the distribution network.

[0084] In an embodiment of the present invention, the step of performing feature dimensionality reduction on the local fusion feature according to the dynamic weight to obtain the dimensionality reduction feature of the node in the distribution network includes: Multiplying the dynamic weight by the local fusion feature element by element to obtain a weighted local fusion feature; The weighted local fusion features are subjected to dimensionality reduction processing based on a principal component analysis algorithm to obtain dimensionality reduction features of the nodes in the distribution network.

[0085] In detail, the generated dynamic weight is multiplied element-by-element with each component of the local fusion feature, and the local fusion feature is weighted, which can highlight the local fusion feature components corresponding to the key influencing factors, suppress unimportant components, and improve the effectiveness of the features.

[0086] In detail, if the local fusion feature is represented as a vector , the dynamic weight is represented as a vector , then the weighted local fusion feature , element-wise multiplication can be achieved using array operations in a programming language such as Python.

[0087] In detail, assuming that the local fusion feature , dynamic weight , then the weighted local fusion feature .

[0088] In detail, principal component analysis is a commonly used data dimensionality reduction method, which projects high-dimensional data into low-dimensional space by finding the principal components of the data, while retaining the main information of the data as much as possible. Performing PCA dimensionality reduction on the weighted local fusion features can reduce the dimension of the features, reduce the computational complexity, and retain the key information of the distribution network operation status.

[0089] Specifically, the weighted local fusion features are first standardized to make their mean 0 and variance 1. Then the covariance matrix of the data is calculated, and the eigenvalue decomposition of the covariance matrix is ​​performed to obtain the eigenvalues ​​and eigenvectors. The top feature with the largest eigenvalue is selected. feature vectors as the principal component, and project the weighted local fusion features onto this The features after dimensionality reduction are obtained on the principal components.

[0090] In detail, assuming that the weighted local fusion feature is a 10-dimensional vector, after PCA dimensionality reduction, the first three principal components are selected, and the 10-dimensional vector is projected onto these three principal components to obtain a 3-dimensional reduced dimensional feature vector.

[0091] S5. Predict the load change trend according to the dimensionality reduction features, generate a dispatching strategy for the distribution network in combination with the safety constraints of voltage limitation and current limitation, and perform dispatching and control on the distribution network according to the dispatching strategy.

[0092] In an embodiment of the present invention, predicting the load change trend according to the dimensionality reduction feature includes: Inputting the dimension reduction features into a pre-trained load change trend prediction model; The load change trend prediction model is used to perform pattern recognition and trend inference on the dimensionality reduction features to obtain the load change trend of the distribution network, wherein the load change trend includes: the time of load rise, fall, stable state information, load peak and valley value occurrence and corresponding load values ​​under different time scales.

[0093] In detail, the reduced-dimensional features obtained through the previous processing contain key information about the operating status of the distribution network. These reduced-dimensional features are provided as input to the pre-trained load change trend prediction model, and the model's learning ability and prediction function are used to analyze the load change of the distribution network.

[0094] In detail, in practical applications, programming languages ​​​​(such as Python) can be used in combination with machine learning or deep learning frameworks (such as TensorFlow, PyTorch, etc.) to implement feature input operations, organize the reduced dimensionality features in the format required by the model (such as matrix or tensor form), and then call the model's prediction interface to input them.

[0095] Specifically, the load change trend prediction model analyzes and processes the input dimensionality reduction features to identify the patterns and rules contained therein. Through these patterns and rules, the load change trends of the distribution network at different time scales are inferred, including the load rise, fall, and stable state information, as well as the time of load peaks and valleys and the corresponding load values.

[0096] In detail, the model can adopt a variety of technologies, such as neural networks, time series analysis algorithms, etc. Taking neural networks as an example, the model learns the mapping relationship between dimensionality reduction features and load change trends during the training process. When the dimensionality reduction features are input, the neural network outputs the predicted load change trend information through forward propagation calculation.

[0097] In detail, the model predicts that in the next 24 hours, the load will be in an increasing state from 9 to 11 in the morning, reaching a load peak of 5000kW at 11; the load will be in a decreasing state from 3 to 5 in the afternoon, reaching a load valley of 3000kW at 5; the load will be relatively stable in the rest of the time periods.

[0098] In an embodiment of the present invention, the method for generating the dispatching strategy is as follows: determining the load demand and power supply capacity of the distribution network in different time periods according to the load change trend; According to the load demand and the power supply capacity, combined with the safety constraints of voltage limit and current limit, the dispatching strategy of the distribution network is generated. In detail, according to the predicted load change trend, the load size of the distribution network in each time period, that is, the load demand, is analyzed. At the same time, combined with the power generation equipment, transmission lines and other resources of the distribution network, the power supply capacity of the distribution network in different time periods is determined.

[0099] In detail, a physical model of the distribution network can be established, taking into account factors such as the rated power of the power generation equipment and the transmission capacity of the transmission line, and combining the load change trend to calculate the power supply capacity. For load demand, the load values ​​of different time periods are directly extracted from the load change trend.

[0100] In detail, according to the load change trend forecast, the load demand is 4500-5000kW between 9-11 am, and the power supply capacity that the power generation equipment and transmission lines of the distribution network can provide is 5500kW.

[0101] In detail, the pre-acquired safety constraints (voltage limit and current limit) are important parameters to ensure the safe operation of the distribution network. Combined with the load demand and power supply capacity, a reasonable scheduling strategy is formulated to achieve the optimal allocation of power resources while meeting the safety constraints.

[0102] In detail, optimization algorithms can be used, such as linear programming, integer programming, etc. Load demand, power supply capacity and safety constraints are taken as constraints, and the minimum power supply cost, load balance, etc. are taken as objective functions, and the scheduling strategy is obtained by solving the optimization problem.

[0103] Specifically, in a certain period of time, the load demand is 4000kW, the power supply capacity is 5000kW, the voltage limit is 10kV±5%, and the current limit is 200A. The optimization algorithm calculates that 2000kW of electricity needs to be dispatched from power station A and 2000kW of electricity needs to be dispatched from power station B to meet the load demand and ensure that the voltage and current are within a safe range.

[0104] In an embodiment of the present invention, the dispatching and controlling the distribution network according to the dispatching strategy includes: Performing task analysis on the dispatching strategy to obtain the dispatching task of the distribution network; Performing instruction conversion on the dispatching task to obtain a control instruction of the distribution network; The control instruction is sent to the control device of the power distribution network based on a preset communication network.

[0105] In detail, the scheduling strategy is usually a comprehensive plan that needs to be broken down into specific, executable scheduling tasks. These tasks specify what operations to perform on which devices at what time to achieve the goals of the scheduling strategy.

[0106] In detail, task parsing rules can be formulated to decompose the content of the dispatching strategy according to dimensions such as equipment, time, and operation type. For example, the dispatching strategy of "adjusting the output power of substation A to 3000kW at 10 am" can be parsed into a specific dispatching task.

[0107] In detail, the scheduling strategy is "In the next 24 hours, adjust the output power of the three power stations according to the load changes". After task analysis, specific scheduling tasks are obtained, such as "Increase the output power of power station 1 by 500kW at 9 am" and "Reduce the output power of power station 2 by 300kW at 3 pm".

[0108] Specifically, the dispatching task is described in natural language or business logic, which needs to be converted into control instructions that can be recognized and executed by the distribution network control equipment. These instructions usually adopt specific communication protocols and formats.

[0109] Specifically, command conversion rules are formulated according to the type and communication protocol of the distribution network control equipment. For example, for control equipment using the Modbus protocol, the scheduling task is converted into a command that conforms to the Modbus protocol format.

[0110] In detail, the scheduling task is "close circuit breaker A". After instruction conversion, a control instruction that complies with the communication protocol is obtained, such as "01050001FF008C3A" (assuming it is a ModbusRTU protocol instruction).

[0111] In detail, the generated control instructions are transmitted to the control equipment of the distribution network (such as circuit breakers, transformer tap controllers, etc.) through a preset communication network (such as fiber optic communication, wireless communication, etc.). After receiving the instructions, the control equipment performs corresponding operations to realize the dispatching and control of the distribution network.

[0112] Specifically, a communication link is established using communication equipment (such as communication gateways, routers, etc.), and control instructions are encapsulated and transmitted according to the requirements of the communication protocol. During the transmission process, the reliability and security of the data need to be considered, such as the use of encryption technology and retransmission mechanisms.

[0113] In detail, the control command is sent to the control device of the remote substation through the optical fiber communication network. After receiving the command, the control device adjusts the tap position of the transformer according to the command requirements.

[0114] It is obvious to those skilled in the art that the present invention is not limited to the details of the above exemplary embodiments, and that the present invention can be implemented in other specific forms without departing from the spirit or essential characteristics of the present invention.

[0115] The embodiments of the present application can acquire and process relevant data based on artificial intelligence technology. Artificial intelligence is the theory, method and technology of using digital computers or machines controlled by digital computers to simulate, extend and expand human intelligence, perceive the environment, acquire knowledge and use knowledge to obtain the best results.

[0116] Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention rather than to limit it. Although the present invention has been described in detail with reference to the preferred embodiments, those skilled in the art should understand that the technical solution of the present invention can be modified or replaced by equivalents without departing from the spirit and scope of the technical solution of the present invention.

Claims

1. A distribution network dispatching and control method based on situation awareness technology, characterized in that: The method comprises: Collecting multi-source data of the distribution network, wherein the multi-source data includes: structural data, environmental data, real-time operation data and equipment data; Constructing an initial distribution network model of the distribution network based on the structure data and the device data, performing vector transformation on the attribute information corresponding to the nodes in the initial distribution network model to obtain the feature vectors of the nodes; Performing vector fusion on the feature vectors using an improved adjacency matrix to obtain local fusion features of nodes in the distribution network, wherein the improved adjacency matrix dynamically adjusts the connection weights between nodes through a dynamic influence function and corrects the node degrees based on load differences; Generate a time series feature of the distribution network according to the environmental data and the real-time operation data, generate a dynamic weight of the local fusion feature according to the time series feature, and perform feature dimension reduction on the local fusion feature according to the dynamic weight to obtain a dimension reduction feature of a node in the distribution network; The load change trend is predicted according to the dimensionality reduction features, and the dispatching strategy of the distribution network is generated in combination with the safety constraints of voltage limitation and current limitation, and the distribution network is dispatched and controlled according to the dispatching strategy.

2. The distribution network dispatching and control method based on the situation awareness technology according to claim 1 is characterized in that: The multi-source data of the distribution network is collected, including: Collecting data from nodes of the distribution network based on a distributed sensor network to obtain initial data of the distribution network; Performing data cleaning on the initial data to obtain cleaned data of the distribution network; The cleaned data is classified to obtain classified data of the distribution network, and it is determined that the classified data is multi-source data of the distribution network.

3. The distribution network dispatching and control method based on the situation awareness technology according to claim 1 is characterized in that: The constructing the initial distribution network model of the distribution network based on the structure data and the equipment data comprises: Determine nodes and connection relationships of the distribution network according to the structure data and the device data; Generating a topological structure of the distribution network according to the nodes and the connection relationships; An initial distribution network model of the distribution network is generated based on a graph theory method and the topological structure.

4. The distribution network dispatching and control method based on the situation awareness technology according to claim 1 is characterized in that: The vector transformation of the attribute information corresponding to the node in the initial distribution network model to obtain the feature vector of the node includes: Extracting attribute information of nodes in the initial distribution network model, wherein the attribute information includes: voltage level, load type and equipment status; Numerical encoding is performed on the attribute information to obtain numerical features of the attribute information; The numerical features are combined to obtain a feature vector of the node.

5. The distribution network dispatching and control method based on the situation awareness technology according to claim 1, characterized in that: The step of using an improved adjacency matrix to perform vector fusion on the feature vectors to obtain local fusion features of the nodes in the distribution network includes: The feature vector is multiplied by the improved adjacency matrix to obtain the local fusion features of the nodes in the distribution network, wherein the improved adjacency matrix is: ； in, is the improved adjacency matrix, is the degree matrix, is the adjacency matrix, Is a node The degree of Indicates whether there is an edge connecting the nodes and nodes , Is a node and nodes The improved connection weights between Representation and Node The set of all adjacent nodes, Is a node and nodes The dynamic influence function between Is a node and nodes The load change difference between Is a node and nodes The difference in environmental data between Is a node and nodes The similarity of device status between Indicates The node ID of each node, It is The node ID of each node, is a constant coefficient that adjusts the weight of the adjacency matrix, is the total number of nodes, It is a constant coefficient that adjusts the effect of load difference on the degree.

6. The distribution network dispatching and control method based on the situation awareness technology according to claim 1, characterized in that: Generating the dynamic weight of the local fusion feature according to the temporal feature includes: Determining key influencing factors of the local fusion feature based on the time series feature; A dynamic weight of the local fusion feature is generated based on the key influencing factors.

7. The distribution network dispatching and control method based on the situation awareness technology according to claim 1, characterized in that: The performing feature dimensionality reduction on the local fusion feature according to the dynamic weight to obtain the dimensionality reduction feature of the node in the distribution network includes: Multiplying the dynamic weight by the local fusion feature element by element to obtain a weighted local fusion feature; The weighted local fusion features are subjected to dimensionality reduction processing based on a principal component analysis algorithm to obtain dimensionality reduction features of the nodes in the distribution network.

8. The distribution network dispatching and control method based on the situation awareness technology according to claim 1, characterized in that: The predicting of load change trend according to the dimension reduction feature includes: Inputting the dimension reduction features into a pre-trained load change trend prediction model; The load change trend prediction model is used to perform pattern recognition and trend inference on the dimensionality reduction features to obtain the load change trend of the distribution network, wherein the load change trend includes: the time of load rise, fall, stable state information, load peak and valley value occurrence and corresponding load values ​​under different time scales.

9. The distribution network dispatching and control method according to any one of claims 1 to 8, characterized in that: The method for generating the dispatching strategy is as follows: determining the load demand and power supply capacity of the distribution network in different time periods according to the load change trend; A dispatching strategy for the distribution network is generated according to the load demand and the power supply capacity in combination with the safety constraints of voltage limitation and current limitation.

10. The distribution network dispatching and control method based on situation awareness technology according to claim 1, characterized in that: The dispatching and controlling the distribution network according to the dispatching strategy includes: Performing task analysis on the dispatching strategy to obtain the dispatching task of the distribution network; Performing instruction conversion on the dispatching task to obtain a control instruction of the distribution network; The control instruction is sent to the control device of the power distribution network based on a preset communication network.

Images