Power distribution network risk assessment and avoidance method considering electric vehicle charging load access

Abstract

The invention discloses a power distribution network risk assessment and avoidance method considering electric vehicle charging load access. The method comprises the steps of 1, performing prediction and scene division on an electric vehicle charging load by using a long short-term memory network and a K-means clustering algorithm; step 2, based on a complex network theory and a risk theory, establishing a risk assessment index system including a comprehensive risk index of a node and a comprehensive risk index of a line; step 3, establishing an objective function by adopting the minimum risk assessment index as an optimization objective; and 4, according to the established objective function, carrying out iterative optimization on the charging facility site selection scheme by adopting a binary particle swarm optimization algorithm to obtain a new charging facility site selection scheme, and completing risk avoidance. According to the method, the power distribution network risk assessment indexes are established, the site selection scheme of the charging facilities is optimized, the operation risk of the power distribution network is effectively reduced, and the toughness and reliability of the power distribution network are improved.

Description

Technical Field

[0001] The present invention belongs to the technical field of distribution network, and in particular to a distribution network risk assessment and avoidance method considering the access of electric vehicle charging loads. Background Art

[0002] With the rapid development of the electric vehicle industry, the number of electric vehicles has increased sharply. The charging load of electric vehicles, as a large load connected to the distribution network, will inevitably pose a challenge to the safe and stable operation of the distribution network. The distribution network is located at the end of the line and bears the impact of the load. The spurt-like increase in the grid load may seriously lead to a further increase in the peak-to-valley difference of the distribution network. In addition, the charging behavior of large-scale electric vehicles may cause fluctuations in the grid voltage, and the large number of power electronic devices used in charging facilities may cause harmonic pollution and affect the quality of power energy.

[0003] At present, domestic and foreign scholars have conducted a series of research in the field of risk assessment. The existing assessment technology can comprehensively consider various uncertainties that may be encountered in the power system, assess the probability of an event and its possible impact, and thus accurately capture and fully reflect the comprehensive impact of the fault on the operation of the entire power system. However, there are relatively few studies on the risk assessment of the coordination between the charging load and the distribution network of electric vehicle charging facilities, which have uncertainty and flexibility. Summary of the invention

[0004] In order to solve the above problems, the present invention proposes a distribution network risk assessment and avoidance method considering the access of electric vehicle charging loads, timely identifying the weak links of the distribution network in this case, and effectively reducing the operating risks of the distribution network and improving the resilience and reliability of the power grid by establishing distribution network risk assessment indicators and optimizing the site selection plan of charging facilities.

[0005] To achieve the above object, the technical solution adopted by the present invention is: a distribution network risk assessment and avoidance method considering the access of electric vehicle charging load, comprising the steps of:

[0006] Step 1: Use long short-term memory network and K-means clustering algorithm to predict electric vehicle charging load and divide scenarios;

[0007] Step 2: Based on complex network theory and risk theory, establish a risk assessment index system, including comprehensive risk indicators of nodes and comprehensive risk indicators of lines;

[0008] Step 3: Take the minimum risk assessment index as the optimization goal and establish the objective function;

[0009] Step 4: According to the established objective function, the binary particle swarm optimization algorithm is used to iteratively optimize the site selection plan of the charging facilities to obtain a new site selection plan for the charging facilities and complete risk avoidance.

[0010] Furthermore, the risk assessment index system combines complex network theory and risk theory to comprehensively evaluate the vulnerability of the distribution network from two dimensions: structure and fault risk. Specific indicators include node importance indicators and line importance indicators based on complex network theory, as well as node voltage over-limit risk indicators and line power over-limit risk indicators based on risk theory, thereby obtaining comprehensive risk indicators of nodes and comprehensive risk indicators of lines to build a risk assessment index system.

[0011] Furthermore, the node importance index considers the weights of node degree, node betweenness and node injection power ratio;

[0012] The node importance formula is:

[0013] I Ni =ω1D Ni +ω2B Ni +ω3N Pi ;

[0014] Among them, I Ni is the importance of node i, D Ni is the node degree of node i, B Ni is the node betweenness of node i, N Pi is the node injection power ratio of node i, ω1, ω2 and ω3 are D Ni , B Ni and N Pi , and ω1+ω2+ω3=1.

[0015] Furthermore, the route importance index considers the weights of route degree and route betweenness;

[0016] The line importance formula is:

[0017] I Li =θ1D Li +θ2B Li ;

[0018] Among them, I Li is the importance of line i, D Li is the line degree of line i, B Li is the line betweenness of line i, θ1 and θ2 are D Li , B Li , and θ1+θ2=1.

[0019] Furthermore, the node's comprehensive risk index is obtained by multiplying the node's importance and the node's voltage over-limit risk index;

[0020] The formula for the comprehensive risk index of the node is:

[0021] S Ni =I Ni ×E Ni ;

[0022] Among them, E Ni is the voltage over-limit risk index value of node i, I Ni is the importance of node i.

[0023] Furthermore, the comprehensive risk index of the line is obtained by multiplying the line importance and the line power over-limit risk index;

[0024] The formula for the comprehensive risk index of the line is:

[0025] S Li =I Li ×E Li ;

[0026] Among them, E Li is the power over-limit risk index value of line i, I Li is the importance of line i.

[0027] Furthermore, the objective function is:

[0028]

[0029] Among them, S Ni is the comprehensive risk index of node i, S Li is the comprehensive risk index of line i, n is the number of nodes in the distribution network model, and l is the number of lines in the distribution network model.

[0030] The beneficial effects of adopting this technical solution are:

[0031] The present invention provides a risk assessment and avoidance for the access of electric vehicle charging load to the distribution network. The method provides a new perspective and method for the risk assessment of the distribution network, provides a new idea and tool for the risk avoidance of the distribution network, and helps to more accurately identify the weak links in the distribution network. The present invention effectively reduces the operating risk of the distribution network, improves the resilience and reliability of the power grid, and provides important decision-making support for urban planning and power grid planning by predicting the distribution characteristics of the charging load, establishing a distribution network risk assessment system, and optimizing the site selection and capacity determination plan for electric vehicle charging facilities.

[0032] The present invention can reduce the risk of distribution network operation: The present invention aims to minimize the risk of distribution network operation caused by the access of electric vehicle charging loads, and uses a binary particle swarm algorithm to optimize the charging facility site selection plan. By comparing the comprehensive risk indicators of the distribution network before and after optimization, it is proved that the optimization plan can effectively reduce the risk of distribution network operation and improve safety.

[0033] The charging facility site selection scheme that can be optimized by the present invention: The charging facility site selection optimization method based on the binary particle swarm algorithm adopted by the present invention solves the discrete problem of charging facility site selection. Compared with before optimization, the voltage of the nodes far from the root node of the distribution network under the optimized charging station site selection scheme has increased significantly, especially the voltage of the nodes at the end of the line has increased significantly. The power of the lines that had the risk of exceeding the power limit before optimization has also decreased to a certain extent after optimization.

[0034] The present invention verifies the correctness of the established distribution network risk assessment index system through comprehensive simulation analysis, verifies the effectiveness of the binary particle swarm algorithm in the site selection of charging facilities, and ensures the reliability and practicality of the design. BRIEF DESCRIPTION OF THE DRAWINGS

[0035] Figure 1 A flow chart of a risk assessment and avoidance method for connecting electric vehicle charging loads to a distribution network;

[0036] Figure 2 is a comparison chart of comprehensive risk indicators before and after optimization; 2(a) is a comparison chart of comprehensive risk indicators of nodes before and after optimization; 2(b) is a comparison chart of comprehensive risk indicators of lines before and after optimization;

[0037] Figure 3 The charging load forecast chart for a day predicted by the trained LSTM model;

[0038] Figure 4 This is the K-means clustering result diagram;

[0039] Figure 5 This is the IEEE33 node distribution network model diagram;

[0040] Figure 6 6(a) is the calculation result diagram of the comprehensive risk index of the node in the initial state; 6(b) is the calculation result diagram of the comprehensive risk index of the line in the initial state;

[0041] Figure 7 Figure 7(a) is a comparison of the comprehensive risk indicators of nodes in various scenarios after charging load access and in the initial state; Figure 7(b) is a comparison of the comprehensive risk indicators of lines in various scenarios after charging load access. DETAILED DESCRIPTION

[0042] In order to make the purpose, technical solutions and advantages of the present invention more clear, the present invention is further described below with reference to the accompanying drawings.

[0043] In this embodiment, a charging facility site selection optimization method based on a binary particle swarm optimization algorithm (BPSO) is proposed. The method aims to minimize the risk of the distribution network, proposes a set of risk assessment index systems to measure the operating risk status of the distribution network, identifies the weak links of the distribution network after the load is connected, and obtains the objective function for optimizing the charging facility site selection plan. The optimized charging facility site selection plan is found through iterative optimization to reduce the operating risk of the distribution network.

[0044] See also Figure 1 As shown, the present invention proposes a distribution network risk assessment and avoidance method considering the access of electric vehicle charging loads, which includes three parts: first, load forecasting is performed to provide data support for subsequent risk assessment; then, risk assessment is performed to establish a risk assessment index, and the minimum risk assessment index is used as the optimization objective function; finally, according to the results of the risk assessment, risk avoidance needs to be performed and the site selection plan for charging facilities is optimized.

[0045] Includes steps:

[0046] Step 1: Use long short-term memory network and K-means clustering algorithm to predict electric vehicle charging load and divide scenarios;

[0047] Step 2: Based on complex network theory and risk theory, establish a risk assessment index system, including comprehensive risk indicators of nodes and comprehensive risk indicators of lines;

[0048] Step 3: Take the minimum risk assessment index as the optimization goal and establish the objective function;

[0049] Step 4: According to the established objective function, the binary particle swarm optimization algorithm is used to iteratively optimize the site selection plan of the charging facilities to obtain a new site selection plan for the charging facilities and complete risk avoidance.

[0050] like Figure 2 As shown, the comparison of comprehensive risk indicators before and after risk assessment and avoidance optimization using the present invention is shown in detail. The figure shows a charging facility site selection optimization method based on the binary particle swarm optimization algorithm (BPSO), which clearly depicts the comprehensive risk indicators of nodes and lines in the initial state, original plan, and optimized plan.

[0051] The present invention is based on the prediction of electric vehicle charging load characteristics, such as Figure 3 As shown in the figure, it can be clearly seen that the trained LSTM model is used to predict the charging load within a day, combined with the scene division of the K-Means clustering algorithm, as shown in the figure. Figure 4As shown in the figure, the charging load during the day can be divided into three scenarios of low, medium and high load by the clustering algorithm. The range of each scenario is: 35.93-92.43kW, 104.0-187.45kW, 205.25-310.32kW, and the cluster centers are: 65.54kW, 134.46kW, and 250.38kW respectively.

[0052] like Figure 6 As shown in Figure 1, the risks of the distribution network mainly come from the risk of the node voltage exceeding the limit due to excessive voltage drop at the node far from the root node of the distribution network, and the risk of the branch power exceeding the limit at the branch close to the root node. Figure 2 As shown in the figure, compared with before optimization, the voltage of nodes far from the root node in the distribution network under the optimized charging station site selection scheme has increased significantly, especially the voltage of nodes at the end of the line has increased significantly. After optimization, the power of lines 1, 2, and 3, which had the risk of power exceeding the limit before optimization, has also decreased to a certain extent.

[0053] As an optimization scheme for the above-mentioned embodiment, the risk assessment index system combines complex network theory and risk theory, and comprehensively evaluates the vulnerability of the distribution network from the two dimensions of structure and fault risk. Specific indicators include node importance indicators and line importance indicators based on complex network theory, as well as node voltage over-limit risk indicators and line power over-limit risk indicators based on risk theory, so as to obtain comprehensive risk indicators of nodes and comprehensive risk indicators of lines to build a risk assessment index system.

[0054] The node importance index considers the weights of node degree, node betweenness and node injection power ratio;

[0055] The node importance formula is:

[0056] I Ni =ω1D Ni +ω2B Ni +ω3N Pi ;

[0057] Among them, I Ni is the importance of node i, D Ni is the node degree of node i, B Ni is the node betweenness of node i, N Pi is the node injection power ratio of node i, w1, w2 and ω3 are D Ni , B Ni and N Pi , and ω1+ω2+ω3=1.

[0058] Node betweenness refers to the ratio of the number of paths that pass through a certain node among all the shortest paths in the network. When calculating specifically, for each pair of nodes in the network, the shortest path between them is calculated, and the number of times these shortest paths pass through a specific node is counted. The ratio of this number to the number of all shortest paths in the network is the betweenness of the node. Node betweenness can be regarded as an indicator of network traffic. In scenarios such as network information dissemination or material flow, node betweenness reflects the size of the "traffic" carried by the node in the network. The larger the betweenness of a node, the more significant the "bridge" role it plays in the network, which means that more shortest paths will pass through the node.

[0059] The definition of node betweenness is:

[0060]

[0061] Among them, B Nk is the betweenness of node k; n ij represents the number of all shortest paths between any two nodes i and j in the network; n ij (k) represents the number of shortest paths passing through node k among all shortest paths; V represents the set of all nodes in the network.

[0062] Compared with the node betweenness defined from the perspective of electrical state, although the analysis from the structural perspective using the traditional definition of node betweenness has certain shortcomings, the calculation is greatly simplified and the time required is shorter, and it can still better identify the important and weak links in the network.

[0063] The size of the node injection power reflects the power exchange capability of the node in the system. The larger the injection power of a node, the more important its role in power transmission and distribution in the system. When a node with high injection power fails, its impact on the system is usually greater than that of a node with low injection power. This is because high injection power nodes are usually directly or indirectly connected to more other nodes. Once they have problems, they may trigger a chain reaction and affect a wider range of the system. Therefore, the size of the node injection power is a key basis for evaluating the importance of a node.

[0064] The calculation formula for the node injection power ratio is:

[0065]

[0066] Among them, N Pi is the node injection power ratio of node i, P i is the injected power of node i, S b It is the baseline capacity of the system.

[0067] The route importance index considers the weights of route degree and route betweenness;

[0068] The line importance formula is:

[0069] I Li =θ1D Li +θ2B Li ;

[0070] Among them, I Li is the importance of line i, D Li is the line degree of line i, B Li is the line betweenness of line i, θ1 and θ2 are D Li , B Li , and θ1+θ2=1.

[0071] In complex network theory, degree is a key indicator to measure the importance and influence of a node in a network. Nodes with high degrees usually play the role of hubs or key connection points in the network, and have an important impact on the connectivity and information dissemination of the network. Degree in complex network theory generally refers to node degree, and there is no clear concept of line degree. Obviously, the more important the nodes at both ends of the line are, the more important the line is. In addition, considering the impact of line length on line importance, the line reactance value is used to replace the role of line length.

[0072] In order to measure the importance of a line in the network, the concept of line degree is introduced:

[0073]

[0074] Among them, D Li is the line degree of line i, Z Li is the reactance value of line i, is the average degree of system nodes, D Ni1 , D Ni2 is the node degree value of the first and last nodes of line i.

[0075] The betweenness of a line refers to the number of times a line in a network is passed by the shortest path between all nodes. It can reflect the importance of the line to the network. If a line is passed by many shortest paths, then its betweenness will be high. Lines with high betweenness play the role of bridges in the network and are crucial to the connectivity and information dissemination of the network.

[0076] Similar to node betweenness, the definition of line betweenness is:

[0077]

[0078] Among them, B Lk is the network betweenness of line k; n ijrepresents the number of all shortest paths between any two nodes i and j in the network; n ij (k) represents the number of shortest paths passing through line k among all shortest paths; V represents the set of all nodes in the network.

[0079] Since the dimensions and magnitudes of different indicators are different, in order to eliminate the influence, each indicator needs to be normalized so that different data can be compared on the same scale. The maximum value of each indicator is taken as the reference value, and the data is normalized, so that the value of each indicator is converted to between [0,1].

[0080] The node's comprehensive risk index is obtained by multiplying the node's importance and the node's voltage over-limit risk index.

[0081] The formula for the comprehensive risk index of the node is:

[0082] S Ni =I Ni ×E Ni ;

[0083] Among them, E Ni is the voltage over-limit risk index value of node i, I Ni is the importance of node i.

[0084] The comprehensive risk index of the line is obtained by multiplying the line importance and the line power over-limit risk index;

[0085] The formula for the comprehensive risk index of the line is:

[0086] S Li =I Li ×E Li ;

[0087] Among them, E Li is the power over-limit risk index value of line i, I Li is the importance of line i.

[0088] From the general form of the risk preference utility function and ignoring the coefficient, the voltage over-limit risk index E of node i can be obtained: Ni The formula is:

[0089]

[0090] Among them, V i is the voltage at node i, V B is the reference voltage. This formula shows that when V i When the node voltage is within the range of 0.93-1.07, the node voltage over-limit risk index is 0, which is considered to be normal operation; when V i <0.93 or Vi >1.07, as V i As the deviation in size increases, the value of the node voltage over-limit risk index increases, and the growth rate becomes faster and faster.

[0091] Similar to the method of determining the node voltage over-limit risk index, the line power over-limit risk index E of line i is Li The formula is:

[0092]

[0093] Among them, P Li is the ratio of the actual active power of line i to the rated active power. It means that when the active load of the line is less than or equal to 80%, the risk of line power over-limit is considered to be 0. When the active load exceeds 80%, the line risk increases with the increase of active load, and the risk increases faster and faster.

[0094] As an optimization solution of the above embodiment, the objective function is:

[0095]

[0096] Among them, S Ni is the comprehensive risk index of node i, S Li is the comprehensive risk index of line i, n is the number of nodes in the distribution network model, and l is the number of lines in the distribution network model.

[0097] The present invention takes the charging facility access point as a variable, the sum of the comprehensive risk index of each node in the distribution network model and the comprehensive risk index of each line as the objective function, and uses the binary particle swarm algorithm to find the best solution to the optimization problem in discrete space and find the appropriate charging facility access point, thereby achieving the purpose of reducing risks.

[0098] The present invention uses the charging facility access point as a variable, the sum of the comprehensive risk index of each node in the distribution network model and the comprehensive risk index of each line as the objective function, and uses the binary particle swarm algorithm to find the optimal solution to the discrete space optimization problem and find the appropriate charging facility access point, thereby achieving the purpose of reducing risks. Figure 5 As shown in the figure, the IEEE33-node distribution network model is used for simulation analysis, the high-load scenario with the highest risk under the original scheme is selected for optimization, the BPSO algorithm is used to generate a new site selection scheme, and the comprehensive risk indicators of the distribution networks of the two schemes are calculated respectively. Figure 2 , proving the effectiveness of the new scheme in reducing risks. According to the structure of the IEEE33 node model, the distribution network is a radial structure. The nodes closer to the root node have a smaller voltage drop, while the nodes farther from the root node have a significant voltage drop, so the risk of voltage over-limit is greater and is downward over-limit.

[0099] The present invention takes into account that in actual situations, the spatial distribution of charging stations should be relatively uniform, so it is assumed that there are 6 nodes or more between every two charging stations, for a total of 4 charging stations. And the charging load is only connected to the PQ nodes, that is, nodes 2 to 33. Therefore, the constraints are:

[0100] (1) Access point range constraints:

[0101]

[0102] (2) Access point spacing constraints:

[0103]

[0104] Among them, a k Indicates the number of the kth access point.

[0105] like Figure 7 As shown in the figure, by comparing different scenarios, it is found that the greater the load, the more obvious the voltage drop of each node and the more obvious the power increase of each line, that is, the risk of the distribution network increases further with the increase of load. Moreover, the nodes farther away from the root node of the radial distribution network and the lines closer to the root node are more sensitive to the increase of load. The access of charging load will make the originally risky nodes and branches have greater over-limit risks and comprehensive risks, and as the charging load increases, the over-limit risks and comprehensive risks of nodes and branches will also increase. In addition, the increase of charging load will also put some nodes and lines that were not originally at risk at risk.

[0106] The above shows and describes the basic principles and main features of the present invention and the advantages of the present invention. It should be understood by those skilled in the art that the present invention is not limited to the above embodiments. The above embodiments and descriptions are only for explaining the principles of the present invention. Without departing from the spirit and scope of the present invention, the present invention may have various changes and improvements, which fall within the scope of the present invention to be protected. The scope of protection of the present invention is defined by the attached claims and their equivalents.

Claims

1. A distribution network risk assessment and avoidance method considering the access of electric vehicle charging loads, characterized in that: Includes steps: Step 1: Use long short-term memory network and K-means clustering algorithm to predict electric vehicle charging load and divide scenarios; Step 2: Based on complex network theory and risk theory, establish a risk assessment index system, including comprehensive risk indicators of nodes and comprehensive risk indicators of lines; Step 3: Take the minimum risk assessment index as the optimization goal and establish the objective function; Step 4: According to the established objective function, the binary particle swarm optimization algorithm is used to iteratively optimize the site selection plan of the charging facilities to obtain a new site selection plan for the charging facilities and complete risk avoidance.

2. The method for risk assessment and avoidance of distribution network considering access of electric vehicle charging load according to claim 1 is characterized in that: The risk assessment index system combines complex network theory and risk theory, and comprehensively evaluates the vulnerability of the distribution network from the two dimensions of structure and fault risk. Specific indicators include node importance indicators and line importance indicators based on complex network theory, as well as node voltage over-limit risk indicators and line power over-limit risk indicators based on risk theory, thereby obtaining comprehensive risk indicators of nodes and comprehensive risk indicators of lines to build a risk assessment index system.

3. The distribution network risk assessment and avoidance method considering the access of electric vehicle charging load according to claim 2 is characterized in that: The node importance index considers the weights of node degree, node betweenness and node injection power ratio; The node importance formula is: I Ni =ω1D Ni +ω2B Ni +ω3N Pi ; Among them, I Ni is the importance of node i, D Ni is the node degree of node i, B Ni is the node betweenness of node i, N Pi is the node injection power ratio of node i, ω1, w2 and v3 are D Ni , B Ni and N Pi , and ω1+ω2+ω3=1.

4. The distribution network risk assessment and avoidance method considering the access of electric vehicle charging load according to claim 2 is characterized in that: The route importance index considers the weights of route degree and route betweenness; The line importance formula is: I Li =θ1D Li +θ2B Li ; Among them, I Li is the importance of line i, D Li is the line degree of line i, B Li is the line betweenness of line i, θ1 and θ2 are D Li , B Li , and θ1+θ2=1.

5. The distribution network risk assessment and avoidance method considering the access of electric vehicle charging load according to claim 2 is characterized in that: The node's comprehensive risk index is obtained by multiplying the node's importance and the node's voltage over-limit risk index. The formula for the comprehensive risk index of the node is: S Ni =I Ni ×E Ni ; Among them, E Ni is the voltage over-limit risk index value of node i, I Ni is the importance of node i.

6. The distribution network risk assessment and avoidance method considering the access of electric vehicle charging load according to claim 2 is characterized in that: The comprehensive risk index of the line is obtained by multiplying the line importance and the line power over-limit risk index; The formula for the comprehensive risk index of the line is: S Li =I Li ×E Li ; Among them, E Li is the power over-limit risk index value of line i, I Li is the importance of line i.

7. The method for risk assessment and avoidance of distribution network considering access of electric vehicle charging load according to claim 1 is characterized in that: The objective function is: Among them, S Ni is the comprehensive risk index of node i, S Li is the comprehensive risk index of line i, n is the number of nodes in the distribution network model, and l is the number of lines in the distribution network model.

Images