Power distribution network overhead line risk refined early warning method and system

Abstract

The invention discloses a power distribution network overhead line risk refined early warning method and system. The method comprises the following steps: collecting meteorological time sequence data and power distribution network topology and line parameter data; and preprocessing the data, and constructing a prediction data set. And establishing a fault probability prediction model based on the meteorological data and the line parameters. Calculating the fault probability by using the mathematical model, correcting the model, training in combination with static topology and dynamic meteorological data, and constructing a fault probability network model. And analyzing the importance of the nodes and lines of the power distribution network, and evaluating the load loss proportion of the nodes. And carrying out normalization processing on the importance indexes. And comprehensively calculating the predicted fault probability, the node loss load and the importance index to obtain a branch risk parameter. Fine early warning is carried out according to a set early warning threshold value, and dynamic adjustment is carried out to adapt to the actual operation condition. The method has the advantages that comprehensive risk assessment is provided, and the accuracy of fault prediction and the timeliness of risk early warning are improved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system risk early warning, and in particular to a method and system for fine-grained early warning of risks of overhead lines in a distribution network taking into account a topological grid structure. Background Art

[0002] In modern distribution network systems, with the increasing demand for electricity and the frequent occurrence of extreme weather events caused by climate change, the reliability and stability of distribution networks face huge challenges. Under extreme weather conditions, such as strong winds, heavy rains, ice and snow, direct physical damage will be caused to the overhead lines in the distribution network, resulting in large-scale power outages. Therefore, how to accurately warn and effectively respond to these risks has become an urgent problem to be solved in the current power system.

[0003] Traditional distribution network risk assessment and early warning methods mainly rely on single historical fault data and static risk models, which often fail to reflect the impact of dynamic meteorological conditions on distribution network reliability in real time. In addition, existing methods generally lack a refined analysis of the internal structure and node differences of the distribution network, and can only provide a rough global early warning, making it difficult for resource scheduling to efficiently respond to specific fault areas.

[0004] At present, in the risk assessment of distribution networks, static topology and simple weighted batch methods are usually used to calculate the importance of lines and nodes, which easily ignores the load level, geographical location and key position of different nodes in the overall network. In actual operation, due to the ever-changing meteorological conditions, it is difficult to accurately locate potential risk points by static analysis alone, which greatly reduces the accuracy and practicality of early warning.

[0005] In addition, there is a lack of quantitative calculation methods for the importance of different nodes and lines within the network, which makes it impossible for the early warning system to effectively identify and prioritize the protection of key nodes and important lines in extreme weather, resulting in frequent power outages across the entire network. Therefore, there is an urgent need for a new method that can comprehensively consider meteorological factors, the topological structure of the distribution network, and the importance of nodes and lines, so as to achieve more refined and real-time risk warnings, so that the power dispatching center can take targeted measures in a timely manner to reduce the scope of power outages and economic losses. Summary of the invention

[0006] In view of the defects of the prior art, the present invention provides a method and system for fine-grained early warning of overhead line risks in a distribution network. By considering the importance of edges and nodes in the distribution network and combining with actual needs, the distribution network branches are divided into early warning levels.

[0007] In order to achieve the above invention object, the technical solution adopted by the present invention is as follows:

[0008] A distribution network overhead line risk refined early warning method comprises the following steps:

[0009] S1. Collect meteorological data of different typical areas of the distribution network, including but not limited to: wind speed, maximum wind speed, temperature, humidity, wind direction and rainfall; collect topological overhead line parameter data of the corresponding area, pre-process the data, including: null value interpolation, repeated observation averaging and deletion of repeated data, and establish a data set;

[0010] S2. Construct a distribution network overhead line fault probability prediction model based on time series data. Calculate the fault probability through a mathematical probability model based on overhead line parameter data and meteorological observation data. And modify the mathematical fault probability model through the fault data of the actual power grid. Construct an overhead line fault probability network model by using a method of fusing static topology with dynamic meteorological data. Based on the data set constructed in step S1, train the overhead line fault probability network model to obtain a fault probability prediction model. The probability of each overhead line failing under specific meteorological conditions and topological structures calculated by the fault probability prediction model.

[0011] S3. Model the node importance and line importance of the distribution network according to the network topology of the distribution network. Construct an undirected graph between different nodes and different lines in the distribution network, consider the key role of node type and overhead line in the topology, and calculate the line importance and node importance; the node importance includes: node type, load level and load loss ratio in the topology when the node fails;

[0012] The importance of a line considers the connectivity of the topological network and the types of nodes connected to it. The node types include load, main power source, and distributed power source. The importance of the node is finally superimposed on the importance of the edge by aggregating adjacent nodes.

[0013] S4. The failure probability of each line predicted in S2 and the importance and load loss ratio of the overhead line obtained in S3 are combined to calculate the risk parameter index of the overhead line. The risk index of the overhead line is divided into reasonable sections according to the set warning threshold, and the warning strategy is specified according to the segmentation of the risk index.

[0014] Furthermore, the S1 includes:

[0015] S11: Collect meteorological time series data sets of different typical regions in the distribution network, including time, region, wind speed (v), temperature (t), wind direction angle (θ), humidity, precipitation, and maximum wind speed. Collect the distribution network topology data and line parameter data of the region, and map these data into the IEEE 33-node topology map.

[0016] S12: Clean and preprocess the meteorological data. First, delete duplicate data and replace the observations in the same location and time step with the average value. Use bilinear interpolation to fill in the empty values. Group the data in the same area according to the continuity of the time series to construct the data set required for step S2.

[0017] Further, the S2 includes:

[0018] S21: Construct a mathematical model of the probability of failure of overhead lines in the distribution network, and calculate the probability of failure of overhead lines under different meteorological data. Based on the actual situation and mechanical properties of the conductors in the overhead transmission lines, in order to simplify the processing of line components, the conductors are regarded as hinges, and the resistance changes caused by the changes in the length of the conductors are ignored. It is assumed that the load of the conductors is evenly distributed along the length, so the oblique parabola is used to calculate the relevant parameters. Under the action of horizontal wind v, the wind load along the line direction is ignored, and the horizontal wind load W1 perpendicular to the conductor direction on each conductor is:

[0019] W1=0.625α1μ1N′LDv 2 sin 2 θ×10 -3

[0020] Where: α1 is the conductor wind pressure unevenness coefficient; μ1 is the conductor body coefficient; N′ is the number of phase conductor splits; L is the length of the conductor under wind load; D is the outer diameter of the conductor; θ is the angle between the wind direction and the conductor direction. In the study of plain areas or large-scale terrain environments, only W1 is generally considered. However, on hillsides, the vertical wind load F on the conductor should be considered. v for:

[0021]

[0022] Where: v v It is the wind speed in the vertical direction.

[0023] G=N′m0gL

[0024] Where: m0 is the mass of the wire per unit length; g is the acceleration due to gravity.

[0025] The total load Q and total specific load γ of the conductor under wind action are:

[0026]

[0027]

[0028] Assume that the original length of the conductor at the manufacturing temperature t0 is L0, and the length of the conductor under wind load is L.

[0029]

[0030] Where: av is the average stress of the conductor; E is the elastic modulus of the conductor; α t is the temperature expansion coefficient of the wire; t m is the temperature under wind load.

[0031] According to the principle that the original length of the conductor is equal under no wind and wind load conditions, the horizontal stress σ0 of the conductor under wind load conditions is obtained. For the conductor in the continuous section, the influence of the offset of the insulator on the height difference angle between the suspension points is ignored. The horizontal distance l from the lowest point of the conductor to the higher tower m for:

[0032]

[0033] Where l is the line span, and β is the height difference angle between the suspension points of the line.

[0034] The maximum stress of the conductor σ m

[0035]

[0036] The exponential function is used to fit the transmission line wind load overload disconnection failure rate P i and the maximum stress of the conductor σ m Relationship between typhoon wind speed v:

[0037]

[0038] Where P i is the probability of failure of a utility pole, σ s is the ultimate stress that the conductor can withstand. The horizontal stress σ0 of the conductor under wind load is related to the design stress of the conductor and the conductor safety factor μ2. K1 and T1 are constants related to line parameters, and C1 is a constant related to wind speed parameters.

[0039] S22: Construct a deep neural network model for fault probability prediction. This network model comprehensively considers overhead line parameter data and time series meteorological data. Feature extraction method: Time series data is extracted through a bidirectional LSTM network, overhead line parameter data is processed by embedding, and the last two feature vectors are concatenated into one feature vector. The fault probability in S21 is used as the true value for network model training. A fault probability prediction model is trained.

[0040] Furthermore, the S3 includes indicators of node importance and line importance of the distribution network according to the static topology.

[0041] S31: Analyze the node type weights, including power source, primary load, secondary load, tertiary load and important factor vector.

[0042] S32: Analyze the importance indicators of nodes in the topology, including degree centrality, closeness centrality, betweenness centrality and eigenvector centrality.

[0043] Degree Centrality:

[0044]

[0045] Where deg(v) is the degree of node v, that is, the number of edges connected to node v, and n is the total number of nodes in the network.

[0046] Closeness Centrality:

[0047]

[0048] Among them, d(v,u) is the shortest path length from node v to node u, and the summation is performed on all nodes u that are different from v.

[0049] Betweenness Centrality:

[0050]

[0051] where σ st is the number of shortest paths from node s to node t. st (v) is the number of shortest paths that pass through node v. The summation is performed over all pairs of nodes (s, t) where s and t are not equal to v to avoid self-loops.

[0052] Eigenvector centrality.

[0053] Let A be the adjacency matrix of the network and w be the eigenvector, where w i represents the eigenvector centrality score of node i. The eigenvector centrality is calculated as follows:

[0054] w=λA w

[0055] Among them, A is the adjacency matrix of the network, w is the eigenvector, where w i represents the eigenvector centrality score of node i, and λ is the largest eigenvalue of matrix A.

[0056] S33: Considering the importance of lines, line types include lines connected to power sources, important shadow vectors, and line lengths. For edge importance indicators in static topology, the analysis includes edge betweenness centrality, edge closeness centrality, and eigenvector centrality.

[0057] Edge betweenness centrality:

[0058]

[0059] where σ stis the total number of shortest paths from node s to node t. st (u,v) is the number of shortest paths from s to t that go through edge (u,v). The sum is calculated over all pairs of nodes (s,t) where s and t are not equal to v to exclude self-looping duplicate edges.

[0060] Edge closeness centrality:

[0061]

[0062] Among them, C C (u) and C C (v) are the closeness centrality of the edge’s two end nodes u and u respectively.

[0063] Eigenvector centrality: Calculate the eigenvector corresponding to the maximum eigenvalue, and finally normalize it so that its sum is 1. The normalization method is Min-Max normalization, which maps the data to the [0,1] interval.

[0064]

[0065] Where x is the original data, Min is the minimum value of the data, Max is the maximum value of the data, and x′ is the transformed data.

[0066] Furthermore, the S3 further includes: constructing an undirected graph of different nodes and different lines in the distribution network, analyzing the node type weights, and then calculating the importance and relevance of the lines and nodes:

[0067] The failure probability predicted by meteorological data is denoted as P, which itself ranges from 0 to 1.

[0068] Analyze a branch in the distribution network and calculate its edge importance and node importance for the branch

[0069] Node importance C1: It is composed of the weighted sum of degree centrality, closeness centrality, betweenness centrality and eigenvector centrality.

[0070] The calculated C D (v),C C (v),C B (v) and w are normalized, and then summed to get the node importance C1, the formula is as follows:

[0071] C1=C D (v)+C C (v)+C B (v)+w

[0072] Edge importance C2: It is composed of the weighted sum of edge betweenness centrality, edge closeness centrality, and eigenvector centrality.

[0073] The calculated C B (u,v),C E (u,v),w1 are normalized, and then summed to get the edge importance C2, the formula is as follows:

[0074] C2=C B (u,v)+C E (u,v)+w1

[0075] Node loss load C3: indicates the proportion of the load lost in the circuit to the total load of the distribution network when a node fails. The formula is as follows:

[0076]

[0077] Where L i Refers to the load occupied by node i, L s Refers to the total load of the distribution network.

[0078] Therefore, the above weight indicators are normalized and summed to obtain the importance C of the branches and nodes:

[0079] C=P+C1′+C2′+C3′

[0080] Based on the above importance indicators, refined threshold warnings can be carried out.

[0081] Further, in said S4:

[0082] S41: Modify the importance C, the formula is as follows:

[0083] C=P+ω1C1′+ω2C2′+ω3C3′+C0

[0084] Among them, ω1, ω2, ω3 are weight coefficients, and C0 is the subjective weight.

[0085] S42: Design a hierarchical warning mechanism, set different levels of warning signals according to the risk level, and take corresponding response measures. In the case of a high-level warning, the system can automatically notify relevant emergency personnel and recommend starting the backup power supply or taking emergency measures.

[0086] S43: Use historical fault data to train and optimize the model to improve the model's ability to respond to emergencies. Monitor the operation of the distribution network in real time and feed new data back to the model for iterative optimization to continuously improve model performance.

[0087] S44: Dynamically adjust the weight of each node in the distribution network according to real-time data. At the same time, consider the special needs of special events and holiday peaks and dynamically adjust the early warning level of the distribution network.

[0088] The present invention also discloses a distribution network overhead line risk refined early warning system, which can be used to implement the above distribution network overhead line risk refined early warning method, specifically including:

[0089] Data collection module: collects meteorological data and electrical parameters of the distribution network and records the geographical location of the distribution network lines to obtain complete data input.

[0090] Data preprocessing module: cleans and preprocesses the collected data, including removing duplicate data, interpolating null values, and standardizing data.

[0091] Topology analysis module: Based on the geographical information and line parameters of the distribution network, the static topology structure of the distribution network overhead lines is constructed to generate a static topology map of the distribution network.

[0092] Model building module: Based on distribution network parameters and meteorological time series data, a risk-refined early warning model is built. This includes establishing a fault probability prediction model and a node and edge importance analysis model to assess the risk of the distribution network.

[0093] Calculation and analysis module: According to the fault prediction probability, node load loss, edge importance and node importance, weight normalization is performed. The importance of the branch is obtained through weighted summation, and risk refinement warning is performed according to the set threshold.

[0094] Risk assessment module: Evaluate the risk level of the distribution network based on the calculation results, set the warning threshold, and classify the risk level to identify potential risk areas and issue warnings.

[0095] Early warning notification module: Issue early warning notifications based on risk levels and recommend appropriate emergency measures. The system automatically notifies relevant emergency personnel to respond to risk situations in a timely manner.

[0096] Real-time monitoring and optimization module: monitor the operation status of the distribution network in real time, iterate and optimize the model according to new data, and dynamically adjust the weights of nodes and edges to adapt to weather changes and special events.

[0097] Historical data management module: manage and use historical fault data for model training and optimization to improve the generalization ability and prediction accuracy of the model. Store and analyze historical data to help improve the system's ability to respond to emergencies.

[0098] The present invention also discloses a computer device, including a memory, a processor, and a computer program stored in the memory and executable on the processor, wherein the processor implements the above-mentioned distribution network overhead line risk refined early warning method when executing the program.

[0099] The present invention also discloses a computer-readable storage medium on which a computer program is stored. When the program is executed by a processor, the above-mentioned distribution network overhead line risk refined early warning method is implemented.

[0100] Compared with the prior art, the advantages of the present invention are:

[0101] The present invention collects meteorological data from different typical areas of the distribution network and builds a fault probability prediction model based on time series data to achieve accurate prediction of the probability of distribution network failure. This method overcomes the limitation of traditional prediction models that are insensitive to changes in complex meteorological conditions and improves the accuracy and timeliness of predictions. At the same time, combined with the static topology structure, the overhead line fault index is corrected, making the distribution network model closer to the actual operation situation, providing a solid foundation for risk warning.

[0102] The present invention fully considers the complexity of the distribution network topology and its impact on fault propagation. This innovation makes the risk warning not only limited to the failure probability of a single line or node, but can comprehensively evaluate the risk status of the entire network, thereby formulating a more scientific and reasonable warning strategy.

[0103] The present invention comprehensively considers multi-dimensional information such as distribution network structure parameters, failure probability predicted by meteorological data, line level and load level, and finally obtains the fault warning level of the distribution network by normalizing the weights of different magnitudes. This method effectively avoids the one-sidedness of single factor evaluation and improves the comprehensiveness and accuracy of risk assessment. At the same time, rationality evaluation and early warning strategy formulation are carried out according to the set early warning threshold, making the early warning more timely and accurate, which helps to dispatch adaptive resources in advance and reduce failure losses.

[0104] The present invention uses node importance analysis technology to perform differentiated processing on different nodes in the distribution network, taking into full account the criticality and impact range of the nodes in the network and their importance in real life. This innovation makes the developed fault warning strategy more targeted and better able to cope with the risks brought by uncertain factors. When the warning strategy is implemented, the power supply stability of important nodes can be prioritized to reduce the chain reaction and amplified losses caused by key node failures. BRIEF DESCRIPTION OF THE DRAWINGS

[0105] Figure 1 It is a flow chart of a method for refined early warning of overhead line risks in a distribution network according to an embodiment of the present invention.

[0106] Figure 2 It is a topological diagram of the distribution network structure according to an embodiment of the present invention. DETAILED DESCRIPTION

[0107] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below with reference to the accompanying drawings and examples.

[0108] like Figure 1 As shown, this embodiment uses the improved IEEE 33-node distribution network system as a test example. The specific structure is as follows Figure 2 As shown. The system includes 33 nodes, 32 segment switches and 5 contact switches. Distributed power sources are configured at nodes 10, 18, 21, 24 and 31. Distributed power sources 10, 20 and 24 are photovoltaic power generation groups and are equipped with energy storage devices. Node 18 is directly equipped with energy storage devices, and node 31 is equipped with a diesel generator. Using the method and system for fine-grained early warning of overhead line risks of distribution networks considering topological grid structures described in the present invention, the specific steps are as follows:

[0109] S1. Collect meteorological data of different typical areas of the distribution network (including but not limited to wind speed, maximum wind speed, temperature, humidity, wind direction, rainfall, etc.), collect topological overhead line parameter data of the corresponding area (such as line length, material, etc.), and pre-process the data, including null value interpolation, repeated observation averaging, and deletion of duplicate data;

[0110] The sub-steps of S1 are as follows:

[0111] S11: Collect meteorological time series data sets of different typical regions in the distribution network, including time, region, wind speed v, temperature t, wind direction angle θ, humidity, precipitation, maximum wind speed, etc. And collect the topological structure data and line parameter data of the region. And migrate these data to the topological map of IEEE33 nodes.

[0112] S12: Clean and preprocess the meteorological data. First, delete duplicate data and replace the observations at the same location and time step with the average value. Fill the empty values ​​with bilinear interpolation. Group the data in the same area according to the continuity of the time series. Construct the data set required for S2.

[0113] S2. Construct a distribution network overhead line fault probability prediction model based on time series data. Calculate the fault probability through a mathematical probability model based on overhead line parameter data and meteorological observation data. And modify the mathematical fault probability model through actual power grid fault data. Construct an overhead line fault probability network model by using a method that integrates static topology and dynamic meteorological data. Based on the data set constructed in step S1, train the overhead line fault probability network model to obtain a fault probability prediction model;

[0114] The sub-steps of S2 are as follows:

[0115] S21: Based on time series data, a distribution network overhead line fault probability prediction model is constructed. The total load Q and total load ratio γ of a certain conductor under wind action are:

[0116]

[0117]

[0118] According to the principle that the original length of the conductor is equal under no wind and wind load conditions, the horizontal stress σ0 of the conductor under wind load conditions can be obtained. For the conductor in the continuous section, the influence of the offset of the insulator on the height difference angle between the suspension points is ignored. The horizontal distance l from the lowest point of the conductor to the higher tower m for:

[0119]

[0120] Where l is the line span, and β is the height difference angle between the suspension points of the line.

[0121] The maximum stress of the conductor σ m

[0122]

[0123] The exponential function is used to fit the transmission line wind load overload disconnection failure rate P i and the maximum stress of the conductor σ m Relationship between typhoon wind speed v:

[0124]

[0125] S22: Construct a deep neural network model for fault probability prediction. This network model comprehensively considers overhead line parameter data and time series meteorological data. Feature extraction method: Time series data is extracted through a bidirectional LSTM network, overhead line parameter data is processed by embedding, and the last two feature vectors are concatenated into one feature vector. The fault probability in S21 is used as the true value for network model training. A fault probability prediction model is trained.

[0126] S3. Model the node importance and line importance of the distribution network according to the network topology of the distribution network. Construct an undirected graph between different nodes and different lines in the distribution network, fully consider the key role of node type and overhead line in the topology, and then calculate the line importance and node importance; the node importance includes node type, load level, and load loss ratio in the topology when the node fails. The importance of the edge mainly considers the connectivity of the topological network and the type of nodes connected to it. The node types include load, main power supply, and distributed power supply. The node importance is finally superimposed on the importance of the edge by aggregating adjacent nodes.

[0127] S3 specifically includes:

[0128] Indicators of node importance and line importance for distribution networks based on static topology.

[0129] First, the node type weights are analyzed, including power source, primary load, secondary load, tertiary load and important factor vector. The node importance index analysis in the topology includes degree centrality, closeness centrality, betweenness centrality and eigenvector centrality. Calculate degree centrality C D (v) (for undirected graphs):

[0130]

[0131] Where deg(v) is the degree of node v, that is, the number of edges connected to node v.

[0132] Closeness Centrality is an indicator used to measure the central position of a node in network analysis. It reflects the average shortest path length from a node to all other nodes in the network. The higher the closeness centrality of a node, the shorter its path to other nodes, and the more central its position in the network.

[0133]

[0134] Among them, d(v,u) is the shortest path length from node v to node u, and the summation is performed on all nodes u that are different from v.

[0135] Betweenness centrality is a metric in network analysis that measures the importance of a node in the shortest path between all pairs of nodes in the network.

[0136]

[0137] where σ st is the number of shortest paths from node s to node t. st (v) is the number of shortest paths that pass through node v. The summation is performed over all pairs of nodes (s, t) where s and t are not equal to v to avoid self-loops.

[0138] Eigenvector centrality is a measure of node importance based on node connectivity. It not only considers the number of connections (degree) of a node but also the importance of the node's neighbors. It is calculated through an iterative process that propagates the node score to its neighbors.

[0139] Let A be the adjacency matrix of the network and w be the eigenvector, where w i represents the eigenvector centrality score of node i. The eigenvector centrality is calculated as follows:

[0140] w=λA w

[0141] where λ is the largest eigenvalue of the matrix A.

[0142] Secondly, the importance of the line is considered. The line types include: important shadow vectors of the line connected to the power supply, and the line length. The importance index analysis of the edge in the static topology includes: edge betweenness centrality, edge closeness centrality, and eigenvector centrality.

[0143] The calculation of edge betweenness centrality is based on the number of times an edge appears in the shortest paths between all pairs of nodes. For an edge (u,v) in the network, its edge betweenness centrality is

[0144]

[0145] where σ st is the total number of shortest paths from node s to node t. st (u,v) is the number of shortest paths from s to t that go through edge (u,v). The sum is calculated over all pairs of nodes (s,t) where s and t are not equal to v to exclude self-looping duplicate edges.

[0146] Edge closeness centrality is defined based on the closeness centrality of the nodes at both ends of the edge. It takes into account the shortest path length from the nodes at both ends of the edge to all other nodes and is defined as:

[0147]

[0148] Eigenvector centrality can be obtained by constructing the adjacency matrix of the network, calculating the eigenvector corresponding to its maximum eigenvalue, and finally normalizing it so that its sum is 1.

[0149] In S3, an undirected graph of different nodes and different lines in the distribution network is constructed, the node type weight is analyzed, and then the importance and correlation of the lines and nodes are calculated:

[0150] First, we denote the failure probability predicted by meteorological data as P, which itself ranges from 0 to 1. We analyze one of the branches of the distribution network and calculate its edge importance and node importance for this branch. Its node importance is denoted as C1, and its degree centrality is denoted as C2. D (v), closeness centrality C C (v), betweenness centrality C B (v) and the weighted sum of the eigenvector centrality w, we can first calculate C D (v),C C (v),C B (v) and w are normalized, and then summed to obtain the node importance C1.

[0151] C1=CD (v)+C C (v)+C B (v)+w

[0152] Similarly, for the importance of the edge C2, it is recorded as the edge betweenness centrality C B (u,v), edge closeness centrality C E (u,v), the weighted sum of the eigenvector centrality w1, can be calculated by first B (u,v),C E (u,v),w1 are normalized, and then summed to get the edge importance C2.

[0153] C2=C B (u,v)+C E (u,v)+w1

[0154] At the same time, we need to consider the node's lost load C3, that is, when this node fails, the proportion of the load lost in the circuit to the total load of the distribution network.

[0155]

[0156] Where L i Refers to the load occupied by node i, L s Refers to the total load of the distribution network.

[0157] Therefore, the above weight indicators are normalized and summed to obtain the importance C of the branches and nodes:

[0158] C=P+ω1C1′+ω2C2′+ω3C3′

[0159] Where C1′ is the normalized node importance, C2′ is the normalized edge importance, and C3′ is the normalized node load loss. Based on this importance, a refined threshold warning can be performed.

[0160] S4: The model fully considers the practical significance of the distribution network in real life. For example, for the distribution network of hospitals and research institutes, no matter how much electricity is consumed, its importance level is relatively high. In order to prevent emergency needs in actual production and life, the reality of the area where the distribution network branch is located is taken into account in the model, and it also acts as a subjective weight C0 to affect the fault warning level of the distribution network. That is, the importance C is modified:

[0161] C=P+ω1C1′+ω2C2′+ω3C3′+C0

[0162] The distribution network structure parameters, the predicted probability obtained from meteorological data, the load loss of each node, the node level and the load level are taken into account to improve the generalization ability and robustness of the model.

[0163] In another embodiment of the present invention, a distribution network overhead line risk refined early warning system is provided, which can be used to implement the above distribution network overhead line risk refined early warning method, specifically including:

[0164] Data collection module: collects meteorological data (including wind speed, wind direction, temperature, humidity, rainfall) and electrical parameters (including line impedance, admittance parameters, etc.) of the distribution network, and records the geographical location of the distribution network lines to obtain complete data input.

[0165] Data preprocessing module: Cleans and preprocesses the collected data, including removing duplicate data, interpolating null values, and standardizing data to ensure data quality and consistency, providing a reliable basis for subsequent analysis.

[0166] Topology analysis module: Based on the geographical information and line parameters of the distribution network, the static topology structure of the distribution network overhead lines is constructed, and the static topology map of the distribution network is generated to provide a topological basis for model construction.

[0167] Model building module: Based on distribution network parameters and meteorological time series data, a risk-refined early warning model is built. This includes establishing a fault probability prediction model and a node and edge importance analysis model to assess the risk of the distribution network.

[0168] Calculation and analysis module: According to the fault prediction probability, node load loss, edge importance and node importance, weight normalization is performed. The importance of the branch is obtained through weighted summation, and risk refinement warning is performed according to the set threshold.

[0169] Risk assessment module: Evaluate the risk level of the distribution network based on the calculation results, set the warning threshold, and classify the risk level (such as red, orange, and yellow warnings) to identify potential risk areas and issue warnings.

[0170] Early warning notification module: Issue early warning notifications based on risk levels and recommend appropriate emergency measures (such as starting backup power supplies). The system automatically notifies relevant emergency personnel to respond to risk situations in a timely manner.

[0171] Real-time monitoring and optimization module: monitor the operation status of the distribution network in real time and iterate and optimize the model according to new data. Dynamically adjust the weights of nodes and edges to adapt to weather changes and special events, and improve the responsiveness and accuracy of the model.

[0172] Historical data management module: manage and use historical fault data for model training and optimization to improve the generalization ability and prediction accuracy of the model. Store and analyze historical data to help improve the system's ability to respond to emergencies.

[0173] In another embodiment of the present invention, a terminal device is provided, the terminal device includes a processor and a memory, the memory is used to store a computer program, the computer program includes program instructions, and the processor is used to execute the program instructions stored in the computer storage medium. The processor may be a central processing unit (CPU), or other general-purpose processors, digital signal processors (DSP), application-specific integrated circuits (ASIC), field-programmable gate arrays (FPGA) or other programmable logic devices, discrete gates or transistor logic devices, discrete hardware components, etc., which are the computing core and control core of the terminal, which are suitable for implementing one or more instructions, and are specifically suitable for loading and executing one or more instructions to implement the corresponding method flow or corresponding functions; the processor described in the embodiment of the present invention can be used for the operation of the distribution network overhead line risk refinement early warning method.

[0174] In another embodiment of the present invention, the present invention further provides a storage medium, specifically a computer-readable storage medium (Memory), which is a memory device in a terminal device for storing programs and data. It can be understood that the computer-readable storage medium here can include both the built-in storage medium in the terminal device and the extended storage medium supported by the terminal device. The computer-readable storage medium provides a storage space, which stores the operating system of the terminal. In addition, one or more instructions suitable for being loaded and executed by the processor are also stored in the storage space, and these instructions can be one or more computer programs (including program codes). It should be noted that the computer-readable storage medium here can be a high-speed RAM memory, or a non-volatile memory (non-volatile memory), such as at least one disk storage.

[0175] One or more instructions stored in a computer-readable storage medium can be loaded and executed by a processor to implement the corresponding steps of the method for refined early warning of risks of overhead lines in a distribution network in the above-mentioned embodiment; one or more instructions in a computer-readable storage medium are loaded and executed by a processor.

[0176] Those skilled in the art will appreciate 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 code.

[0177] 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.

[0178] 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.

[0179] 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.

[0180] Those skilled in the art will appreciate that the embodiments described herein are intended to help readers understand the implementation methods of the present invention, and should be understood that the protection scope of the present invention is not limited to such special statements and embodiments. Those skilled in the art can make various other specific variations and combinations that do not deviate from the essence of the present invention based on the technical revelations disclosed by the present invention, and these variations and combinations are still within the protection scope of the present invention.

Claims

1. A distribution network overhead line risk refinement early warning method, characterized in that: The following steps are involved: S1. Collect meteorological data of different typical areas of the distribution network, including but not limited to: wind speed, maximum wind speed, temperature, humidity, wind direction and rainfall; Collect the topological overhead line parameter data of the corresponding area, pre-process the data, including: null value interpolation, repeated observation averaging and deletion of duplicate data, and establish a data set; S2. Construct a distribution network overhead line fault probability prediction model based on time series data; calculate the fault probability through a mathematical probability model according to overhead line parameter data and meteorological observation data; and modify the mathematical fault probability model through actual power grid fault data; construct an overhead line fault probability network model by adopting a method of fusing static topology with dynamic meteorological data; train the overhead line fault probability network model based on the data set constructed in step S1 to obtain a fault probability prediction model; calculate the probability of each overhead line failure under specific meteorological conditions and topological structures through the fault probability prediction model; S3. Model the node importance and line importance of the distribution network according to the network topology of the distribution network; construct an undirected graph between different nodes and different lines in the distribution network, consider the key role of node type and overhead line in the topology, and calculate the line importance and node importance; the node importance includes: node type, load level and load loss ratio in the topology when a node fails; The importance of a line considers the connectivity of the topological network and the types of nodes connected to it; the node types include: load, main power source and distributed power source; the importance of a node is finally superimposed on the importance of the edge by aggregating adjacent nodes; S4. The failure probability of each line predicted in S2 and the importance and load loss ratio of the overhead line obtained in S3 are combined to calculate the overhead line risk parameter index; the overhead line risk index is divided into reasonable sections according to the set warning threshold, and the warning strategy is specified according to the segmentation of the risk index.

2. The method for fine-grained early warning of overhead line risks in a distribution network according to claim 1 is characterized in that: The S1 specifically includes: S11: Collect meteorological time series data sets of different typical regions in the distribution network, including time, region, wind speed v, temperature t, wind direction angle θ, humidity, precipitation and maximum wind speed; collect the distribution network topology data and line parameter data of the region, and map these data into the IEEE 33-node topology map; S12: Clean and preprocess the meteorological data; first delete duplicate data and replace the observations at the same location and time step with the average value; use bilinear interpolation to fill in the empty values; group the data in the same area according to the continuity of the time series to construct the data set required for step S2.

3. The method for fine-grained early warning of overhead line risks in a distribution network according to claim 1 is characterized in that: The S2 includes: S21: Construct a mathematical model of the probability of failure of overhead lines in the distribution network, and calculate the probability of failure of overhead lines under different meteorological data; based on the actual situation and mechanical properties of the conductors in the overhead transmission lines, for the sake of simplifying the line components, the conductors are regarded as hinges, and the resistance changes caused by the changes in the length of the conductors are ignored. It is assumed that the load of the conductors is evenly distributed along the length, so the oblique parabola is used to calculate the relevant parameters; under the action of horizontal wind v, the wind load along the line direction is ignored, and the horizontal wind load W1 perpendicular to the conductor direction on each conductor is: W1=0.625α1μ1N ′ LDv 2 sin 2 θ×10 -3 Where: α1 is the conductor wind pressure unevenness coefficient; μ1 is the conductor body coefficient; N ′ is the number of phase conductor splits; L is the length of the conductor under wind load; D is the outer diameter of the conductor; θ is the angle between the wind direction and the conductor direction; in the study of plain areas or large-scale terrain environments, generally only W1 is considered; but on hillsides, the vertical wind load F on the conductor should be considered v for: Where: v v is the wind speed upward along the vertical line; G=N ′ m0gL Where: m0 is the mass of the wire per unit length; g is the acceleration due to gravity; The total load Q and total specific load γ of the conductor under wind action are: Assume that the original length of the conductor at the manufacturing temperature t0 is L0, and the length of the conductor under wind load is L. Where: av is the average stress of the conductor; E is the elastic modulus of the conductor; α t is the temperature expansion coefficient of the wire; t m is the temperature under wind load; According to the principle that the original length of the conductor is equal under no wind and wind load conditions, the horizontal stress σ0 of the conductor under wind load conditions is obtained; for the conductor in the continuous section, the influence of the offset of the insulator on the height difference angle between the suspension points is ignored; the horizontal distance l from the lowest point of the conductor to the higher tower m for: Where l is the line span, and β is the height difference angle between the suspension points of the line. The maximum stress of the conductor σ m The exponential function is used to fit the transmission line wind load overload disconnection failure rate P i and the maximum stress of the conductor σ m Relationship between typhoon wind speed v: Where P i is the probability of failure of a utility pole, σ s is the limit stress that the conductor can withstand. The horizontal stress σ0 of the conductor under wind load is related to the design stress of the conductor and the safety factor μ2 of the conductor. K1 and T1 are constants related to line parameters, and C1 is a constant related to wind speed parameters. S22: Construct a deep neural network model for fault probability prediction, which comprehensively considers overhead line parameter data and time-series meteorological data; feature extraction method: time series data is extracted through a bidirectional LSTM network, and overhead line parameter data is processed by embedding, and finally the two feature vectors are concatenated into one feature vector; the fault probability in S21 is used as the true value for network model training; and a fault probability prediction model is trained.

4. The method for fine-grained early warning of overhead line risks in a distribution network according to claim 1 is characterized in that: Indicators of node importance and line importance of the distribution network according to the static topology in S3; S31: Analyze the node type weights, including power source, primary load, secondary load, tertiary load and important factor vector; S32: Analyze the importance indicators of nodes in the topology, including degree centrality, closeness centrality, betweenness centrality and eigenvector centrality; Degree Centrality: Where deg(v) is the degree of node v, that is, the number of edges connected to node v, and n is the total number of nodes in the network; Closeness Centrality: Where d(v,u) is the shortest path length from node v to node u, and the sum is performed on all nodes u that are different from v; Betweenness Centrality: where σ st is the number of shortest paths from node s to node t; σ st (v) is the number of shortest paths passing through node v; the sum is performed over all pairs of nodes (s, t) where s and t are not equal to v to avoid self-loops; Eigenvector centrality; Let A be the adjacency matrix of the network and w be the eigenvector, where w i represents the eigenvector centrality score of node i; the eigenvector centrality is calculated as follows: w=λA w Among them, A is the adjacency matrix of the network, w is the eigenvector, where w i represents the eigenvector centrality score of node i, and λ is the largest eigenvalue of matrix A; S33: Considering the importance of lines, line types include lines connected to power sources, important shadow vectors, and line lengths; for edge importance indicators in static topology, the analysis includes edge betweenness centrality, edge closeness centrality, and eigenvector centrality; Edge betweenness centrality: where σ st is the total number of shortest paths from node s to node t; σ st (u,v) is the number of shortest paths from s to t via edge (u,v); the sum is performed over all pairs of nodes (s,t) where s and t are not equal to v to exclude self-looping duplicate edges; Edge closeness centrality: Among them, C C (u) and C C (v) are the closeness centrality of the two end nodes u and u of the edge respectively; Eigenvector centrality: Calculate the eigenvector corresponding to the maximum eigenvalue, and finally normalize it so that its sum is 1; the normalization method is Min-Max normalization, mapping the data to the [0,1] interval; Where x is the original data, Min is the minimum value of the data, Max is the maximum value of the data, and x′ is the transformed data.

5. The method for fine-grained early warning of overhead line risks in a distribution network according to claim 4 is characterized in that: The S3 further includes: constructing an undirected graph of different nodes and different lines in the distribution network, analyzing the node type weights, and then calculating the importance and relevance of the lines and nodes: The failure probability predicted by meteorological data is recorded as P, which itself ranges from 0 to 1; Analyze a branch in the distribution network and calculate its edge importance and node importance for the branch Node importance C1: It is composed of the weighted sum of degree centrality, closeness centrality, betweenness centrality and eigenvector centrality; The calculated C D (v),C C (v),C B (v) and w are normalized, and then summed to get the node importance C1, the formula is as follows: C1=C D (v)+C C (v)+C B (v)+w Edge importance C2: It is composed of the weighted sum of edge betweenness centrality, edge closeness centrality, and eigenvector centrality; The calculated C B (u,v),C E (u,v),w1 are normalized, and then summed to get the edge importance C2, the formula is as follows: C2=C B (u,v)+C E (u,v)+w1 Node loss load C3: indicates the proportion of the load lost in the circuit to the total load of the distribution network when a node fails. The formula is as follows: Where L i Refers to the load occupied by node i, L s Refers to the total load of the distribution network; Therefore, the above weight indicators are normalized and summed to obtain the importance C of the branches and nodes: C=P+C1 ′ +C2 ′ +C3 ′ Based on the above importance indicators, refined threshold warnings can be carried out.

6. The method for fine-grained early warning of overhead line risks in a distribution network according to claim 5 is characterized in that: In S4: S41: Modify the importance C, the formula is as follows: C=P+ω1C1 ′ +ω2C2 ′ +ω3C3 ′ +C0 Among them, ω1, ω2, ω3 are weight coefficients, and C0 is the subjective weight; S42: Design a hierarchical warning mechanism, set different levels of warning signals according to the risk level, and take corresponding response measures; in the case of high-level warnings, the system can automatically notify relevant emergency personnel and suggest starting the backup power supply or taking emergency measures; S43: Use historical fault data to train and optimize the model to improve the model's ability to respond to emergencies; monitor the distribution network operation status in real time, and feed new data back to the model for iterative optimization to continuously improve model performance; S44: Dynamically adjust the weight of each node in the distribution network based on real-time data; at the same time, dynamically adjust the warning level of the distribution network considering the special needs of special events and holiday peak periods.

7. A distribution network overhead line risk refinement early warning system, characterized by: The system can be used to implement the distribution network overhead line risk refinement early warning method according to any one of claims 1 to 6, specifically including: Data collection module: collects meteorological data and electrical parameters of the distribution network and records the geographical location of the distribution network lines to obtain complete data input; Data preprocessing module: cleans and preprocesses the collected data, including removing duplicate data, interpolating null values, and standardizing data; Topology analysis module: Based on the geographical information and line parameters of the distribution network, the static topology structure of the overhead lines of the distribution network is constructed, and the static topology diagram of the distribution network is generated; Model building module: Based on distribution network parameters and meteorological time series data, a risk-refined early warning model is built; this includes establishing a fault probability prediction model, as well as a node and edge importance analysis model to assess the risk of the distribution network; Calculation and analysis module: Perform weight normalization based on the fault prediction probability, node load loss, edge importance, and node importance; obtain the importance of the branch through weighted summation, and perform risk refinement warning based on the set threshold; Risk assessment module: Evaluate the risk level of the distribution network based on the calculation results, set the warning threshold, and classify the risk level to identify potential risk areas and issue warnings; Early warning notification module: issues early warning notifications based on risk levels and recommends appropriate emergency measures; the system automatically notifies relevant emergency personnel to respond to risk situations in a timely manner; Real-time monitoring and optimization module: monitors the operation status of the distribution network in real time, iterates and optimizes the model based on new data; dynamically adjusts the weights of nodes and edges to adapt to weather changes and special events; Historical data management module: manage and use historical fault data for model training and optimization to improve the generalization ability and prediction accuracy of the model; store and analyze historical data to help improve the system's response capabilities to emergencies.

8. A computer device, characterized in that: The invention comprises a memory, a processor and a computer program stored in the memory and executable on the processor, wherein when the processor executes the program, the method for refined early warning of the risk of overhead lines in a distribution network as claimed in one of claims 1 to 6 is implemented.

9. A computer-readable storage medium, characterized in that: A computer program is stored thereon, and when the program is executed by a processor, the method for refined early warning of risks of overhead lines in a distribution network as described in one of claims 1 to 6 is implemented.

Images