Resilient power grid risk assessment method and system based on line flow direction optimization

By extracting and optimizing bidirectional lines in the power grid, using sensitivity analysis method and probability model, the problems of reliability and efficiency in the grid risk assessment process are solved, and more accurate risk identification and evaluation are achieved.

CN113762748BActive Publication Date: 2024-11-05TSINGHUA UNIVERSITY
View PDF 0 Cites 3 Cited by

Patent Information

Application Number
CN202110990858.7
Authority / Receiving Office
CN · China
Patent Type
Patents(China)
Current Assignee / Owner
Filing Date
2021-08-26
Publication Date
2024-11-05
Estimated Expiration
2041-08-26

Smart Images

  • Figure CN113762748B_ABST
    Figure CN113762748B_ABST
Patent Text Reader

Abstract

The present invention provides a resilient power grid risk assessment method and system based on line flow direction optimization, the method comprising: extracting bidirectional lines in the resilient power grid and constructing a bidirectional line set; extracting key lines from the bidirectional line set using a sensitivity analysis method; using the direction probability of the key lines as decision variables, the upper and lower bounds of the decision variables as constraints, and the risk of load outage loss as an objective function to construct a line flow direction optimization model; according to the line flow direction optimization model, performing probability optimization on the flow direction of the key lines to obtain a risk estimation interval of the resilient power grid. The resilient power grid risk assessment method and system based on line flow direction optimization provided by the present invention can not only improve the reliability of the risk assessment results by extracting key lines and optimizing the line directions that have a greater impact on the power grid risk through the line flow direction optimization model, but also estimate the risk interval more quickly, thereby improving the efficiency of risk assessment.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The present invention relates to the technical field of power grid risk assessment, and in particular to a resilient power grid risk assessment method and system based on line flow direction optimization. Background Art

[0002] The distribution network is a power distribution system from the main grid to the user end, and is widely distributed in the power system. Since most of its equipment is medium and low voltage equipment, the disaster prevention level is not high, and the losses are often greater when it is affected by natural disasters. Therefore, it is necessary to conduct pre-disaster accident evolution simulation and pre-disaster early warning preparation before the disaster occurs, explore the occurrence mechanism of extreme weather events and the characteristics of their impact on the power grid, especially the analysis of the weak links in the system. A resilient distribution network refers to a power grid that can comprehensively, quickly and accurately perceive the operation status of the power grid, coordinate internal and external resources of the power grid, actively predict and actively prepare for various disturbances, actively defend, and quickly restore important power loads. The pre-disaster analysis of a resilient distribution network is usually divided into several parts, such as equipment failure modeling, outage loss analysis, and disaster scenario generation.

[0003] In outage loss analysis, it is generally optimistically assumed that all lines have the ability to supply power in both directions. However, in the actual operation of the distribution network, the power flow direction of any line is always determined, either by the topological structure or by the electrical parameter relationship, resulting in the ideal line power flow direction being inconsistent with the actual operation of the system. For example, the power supply can only be supplied from the power supply to the load, and the load nodes at the end that are not connected to the power supply can only obtain power from the remaining nodes. This will lead to optimistic risk estimation results and reduce the reliability of the risk assessment process. At the same time, since the risk assessment process needs to consider all lines, the assessment efficiency is difficult to guarantee.

[0004] Therefore, there is an urgent need for an accurate and efficient resilient power grid risk assessment method based on line flow direction optimization to solve the above problems. Summary of the invention

[0005] The present invention provides a resilient power grid risk assessment method and system based on line flow direction optimization, which are used to solve the defects of low reliability of risk assessment process results and low assessment efficiency in the prior art.

[0006] In a first aspect, the present invention provides a resilient power grid risk assessment method based on line flow direction optimization, comprising:

[0007] Extract bidirectional lines in the resilient power grid and construct a bidirectional line set;

[0008] A sensitivity analysis method is used to extract key lines from the bidirectional lines;

[0009] Taking the direction probability of the key line as the decision variable, the upper and lower bounds of the decision variable as the constraint conditions, and the risk of load outage loss as the objective function, a line flow direction optimization model is constructed;

[0010] According to the line power flow direction optimization model, the power flow direction of the key line is probabilistically optimized to obtain a risk estimation interval of the resilient power grid.

[0011] According to a resilient power grid risk assessment method based on line power flow direction optimization provided by the present invention, a sensitivity analysis method is used to extract key lines from the bidirectional lines, including:

[0012] The normalized probability of each bidirectional line in the bidirectional line set supplying power in both directions is set to 0.5, and the current initial load outage loss risk is calculated;

[0013] sequentially adjusting one of the bidirectional lines in the bidirectional line set to unidirectional power supply, and calculating the load outage loss risk after the adjustment respectively;

[0014] According to the adjusted load outage loss risk and the initial load outage loss risk, the relative change of risk corresponding to each bidirectional line is obtained;

[0015] The relative risk change corresponding to each bidirectional line is compared with the preset risk change threshold, and the bidirectional lines with relative risk change greater than the risk change threshold are extracted as key lines.

[0016] According to a resilient power grid risk assessment method based on line power flow direction optimization provided by the present invention, the calculation process of the load outage loss risk includes:

[0017] An extended probability graph model is constructed, using known data previously obtained from the grid topology file and the grid line fault probability table of the resilient distribution network, and the probability of power supply at the load node at each moment in the target period is calculated through the node power supply probability equation group in the extended probability graph model;

[0018] The probability of having power on at the load node at each moment in the target period is converted into the probability of losing power at the load node at each moment in the target period, and the load outage loss risk is calculated based on the probability of losing power at the load node at each moment in the target period.

[0019] According to a resilient power grid risk assessment method based on line power flow direction optimization provided by the present invention, the node power probability equation group is:

[0020]

[0021] In the formula, P(e i ) represents the probability of node i being powered, ΩS represents the power node set, Ω i represents the set of adjacent nodes of node i, Ω l represents the load node set, P(e j→i ) represents the probability of power supply in the direction of line j→i;

[0022] The calculation formula for the probability of power supply in the direction of line j→i is:

[0023] P(e j→i )=P(e j-i )δ ji

[0024] In the formula, P(e j-i ) represents the probability of power supply in the inter-gird line, δ ji It represents the normalized probability that line ji supplies power in the direction j→i.

[0025] According to a resilient power grid risk assessment method based on line power flow direction optimization provided by the present invention, the calculation formula for the load outage loss risk is:

[0026]

[0027] Among them, R represents the risk of loss caused by outage of the entire network load, P i Represents the load power of load node i, W i represents the value loss of the unit power load outage at load node i, represents the power failure probability of the load node at time t, Ω l Represents a set of load nodes.

[0028] According to a resilient power grid risk assessment method based on line flow direction optimization provided by the present invention, the line flow direction optimization model is:

[0029]

[0030] st0≤δ≤1

[0031] P(e i→j )=P(e i-j )δ ij

[0032] Among them, δ represents the directional probability of the critical line, R(x) represents the risk expression of load outage loss, f(x, p′) represents the directional node power probability equation group, P(e i→j ) represents the probability of power supply in the line i→j direction, that is, the probability that the line from node i to j is fault-free and the power supply direction is from i to j, P(e i-j ) represents the probability of power supply in the line between ij, δ ijIt represents the normalized probability that line ij supplies power in the direction i→j.

[0033] In a second aspect, the present invention further provides a resilient power grid risk assessment system based on line flow direction optimization, comprising:

[0034] A bidirectional line extraction module is used to extract bidirectional lines in the resilient power grid and construct a bidirectional line set;

[0035] A key circuit extraction module, used for extracting key circuits from the bidirectional circuit set by using a sensitivity analysis method;

[0036] An optimization model building module is used to construct a line flow direction optimization model using the direction probability of the key line as a decision variable, the upper and lower bounds of the decision variable as constraints, and the risk of load outage loss as an objective function;

[0037] The risk estimation module is used to perform probability optimization on the power flow direction of the key line according to the line power flow direction optimization model to obtain the risk estimation interval of the resilient power grid.

[0038] In a third aspect, the present invention also provides an electronic device comprising 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 steps of the resilient power grid risk assessment method based on line flow direction optimization as described in any one of the above are implemented.

[0039] The present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, the steps of the resilient power grid risk assessment method based on line flow direction optimization as described in any of the above are implemented.

[0040] The present invention also provides a computer program product, including a computer program, which, when executed by a processor, implements the steps of any of the above-mentioned resilient power grid risk assessment methods based on line flow direction optimization.

[0041] The resilient power grid risk assessment method and system based on line flow direction optimization provided by the present invention can not only improve the reliability of risk assessment results, but also estimate risk intervals more quickly, thereby improving risk assessment efficiency, by extracting key lines and optimizing line directions that have a greater impact on power grid risks through a line flow direction optimization model. BRIEF DESCRIPTION OF THE DRAWINGS

[0042] In order to more clearly illustrate the technical solutions in the present invention or the prior art, the following briefly introduces the drawings required for use in the embodiments or the description of the prior art. Obviously, the drawings described below are some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0043] Figure 1 It is a flow chart of a resilient power grid risk assessment method based on line flow direction optimization provided by the present invention;

[0044] Figure 2 It is a topological diagram of typical power flow direction of a power grid with DG disconnection;

[0045] Figure 3 This is a topology diagram corresponding to an instance with power islands;

[0046] Figure 4 is the risk partial derivative calculation tree diagram;

[0047] Figure 5 It is a structural schematic diagram of a resilient power grid risk assessment system based on line flow direction optimization provided by the present invention;

[0048] Figure 6 It is a structural schematic diagram of the electronic device provided by the present invention. DETAILED DESCRIPTION

[0049] In order to make the purpose, technical solution and advantages of the present invention clearer, the technical solution of the present invention will be clearly and completely described below in conjunction with the drawings of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0050] Figure 1 The embodiment of the present invention provides a resilient power grid risk assessment method based on line flow direction optimization, including:

[0051] S110: extracting bidirectional lines in the resilient power grid and constructing a bidirectional line set.

[0052] In actual distribution networks, most lines have a clear power supply direction, that is, power supply from the power source side to the load side. Lines with bidirectional power supply capabilities must meet one condition, that is, the nodes on both sides of the connection are power sources or can be connected to power sources (including distributed power sources) through other lines.

[0053] Therefore, if a line is disconnected and a power island appears in the power grid, the line can only flow power in one direction. Distinguishing between one-way and two-way line identification is conducive to reducing the number of variables in the optimization problem, greatly improving the speed of solving the optimization problem, and ensuring the effectiveness of risk interval estimation.

[0054] First, in order to facilitate the analysis of resilient power grids, an embodiment of the present invention constructs an extended probabilistic graph model of the power grid, and divides the equipment in the distribution network into two categories: node-type equipment and line-type equipment. Line-type equipment includes power supply lines and transfer lines. In the resilient distribution network, nodes can be divided into two categories: power supply nodes and load nodes according to the type of equipment connected to the bus.

[0055] If the fault of the busbar itself is not considered, it can be considered that the reliability of the power supply in the distribution network is relatively high, and whether it works normally has nothing to do with the surrounding nodes (i.e., no faults). Whether the load node has power depends on the power supply status of the surrounding nodes. If a node has power, at least one of its surrounding nodes must be able to supply power to it through the line between the two nodes.

[0056] In other words, the extended probabilistic graph model believes that the power supply node can reliably supply power with a probability of 1; the load node needs to obtain power from the surrounding nodes and depends on the lines connected to the node.

[0057] Then, the node power probability equation group is established according to the association information of the above nodes and lines to obtain the extended probability graph model. In this embodiment, the node power probability equation group uses the power probability of the connected lines as known variables and the power probability of the target node and its surrounding nodes as unknown variables.

[0058] The extended probabilistic graph model can traverse all lines, set the outage probability of the current line to 1, and solve the node power probability equation group through the Newton iteration method to check whether there are power-free islands (i.e., node sets with a power probability of 0). The problem with this algorithm is that since the adjacency matrix is ​​0-1, the Jacobian matrix may be singular at different initial values, and the iterative method cannot give a numerical solution to the equation at this time. The reason for this problem is that the Jacobian matrix has linear correlation terms, is not full rank, and is extremely ill-conditioned. Referring to the ill-conditioned power flow solution in power flow calculation, commonly used methods include the optimal multiplier method, nonlinear programming method, etc., but these methods are complex to calculate, and the solutions given may not meet the analysis requirements of the graph. Therefore, it is considered to use graph theory to analyze the relationship between edges in abstract graphs.

[0059] The analysis process needs to divide the resilient power grid into two types: single-power distribution network and multi-power distribution network, and perform topological analysis on them respectively. The specific analysis process is as follows:

[0060] For a single-power distribution network, all loads obtain power from a reliable power source. At this time, the network is a connected graph with a connected component number of 1. If the number of connected components in the graph changes after disconnecting an edge, the network is split into two connected parts with power supply and without power supply, and an island without power supply will definitely be formed. For undirected graphs, such edges are called bridges or cut edges. There are many identification algorithms for bridges in undirected graphs, including depth-first search (DFS), Tarjan algorithm, and chain-based decomposition algorithms. The cut edge identification algorithm for undirected graphs is fast and easy to apply to large single-power sparse distribution systems. For dual-ring and multi-ring network systems with redundant power supply, there may be multiple lines between any two nodes. For simplicity, the power flow direction between multiple lines between the same two nodes is considered to be the same, and only single-line network analysis is required.

[0061] In other words, if the system is a single-power distribution network, the undirected graph bridge can be used to screen the bidirectional lines. The undirected graph bridge means that the number of connected components of the graph will change after deleting the edge. The unidirectional lines in the distribution network can also be identified by this method. When the unidirectional line is disconnected, the system will be split into a part with power supply and a part without power supply. According to the topological structure, the unidirectional line set is obtained by the undirected graph bridge algorithm, and the bidirectional line set is obtained by taking the complement set.

[0062] For a distribution network with multiple power sources (including distributed power sources), after a distribution network with multiple power sources forms an island, it is necessary to determine the connection relationship between the island and any power source. If it is assumed that any power source can supply power to the load, the judgment criteria for a single power distribution network will be invalid. At the same time, since distributed power sources are involved, disconnecting a line may result in an operation area containing a power source island. A typical topology diagram containing DG (Distributed Generation) is shown below: Figure 2 shown.

[0063] Depend on Figure 2 It can be seen that in the case of distributed power sources, the judgment method of generating islands without power sources using bridges is not sufficient. According to graph theory knowledge, for undirected graphs where connected components may change, the reachable matrix can be used to describe the island formation of the system. The reachable matrix is ​​defined as a matrix representation of the degree of reachability between nodes of a directed connection graph after a certain length of path. The reachable matrix describes the block connectivity of the graph. If the graph is fully connected, the elements of the reachable matrix are all 1. The row maximal linearly independent group of the reachable matrix describes the block status of the matrix. The basic calculation method of the reachable matrix is:

[0064] P=A∨A 2 ∨…∨A n (1)

[0065] Among them, matrix A is an n-order 0-1 square matrix, A nIt means that matrix A is a power of n and is symmetric. ∨ represents the element-wise logical OR operation of a 0-1 matrix.

[0066] by Figure 2 Take the power island identification as an example, the node number is as follows Figure 3 As shown, the adjacency matrix and reachability matrix are as follows:

[0067]

[0068] In this example, nodes 7 and 8 form a zone, and since node 8 is a distributed power source, the zone does not lose power.

[0069] Finally, according to the block status of the reachable matrix and the number of the power nodes, it can be determined whether the system forms a power-free island.

[0070] It can be seen from this that if the system is a multi-power grid, it is necessary to use the reachable matrix for judgment. Among all the lines, disconnect one line in turn, and calculate the reachable matrix at this time, and judge whether there is a power island. If so, add the current disconnected line to the unidirectional line set. If no power island is formed, add the line to the bidirectional line set.

[0071] S120: The sensitivity analysis method is used to extract the key lines from the bidirectional lines.

[0072] Due to the complex structure of the power grid and the large number of bidirectional lines in actual applications, the scale of the line direction optimization problem is still huge. For large networks, the impact of a single line on system risk is limited, especially the bidirectional lines at the end of the power grid. Therefore, only optimizing the line direction that has a greater impact on the power grid risk can give a faster estimate of the risk range without basically affecting the solution quality.

[0073] In order to distinguish the lines that have a greater impact on risk assessment, this embodiment uses a sensitivity analysis method to extract key lines from the bidirectional line set. The process specifically includes:

[0074] First, under the condition that the remaining lines in the bidirectional line set are bidirectional with a probability of 0.5, a certain line is set to only supply power in one direction (that is, the direction probability is 1), and then the system risk (that is, the risk of load outage loss) is calculated for each situation in turn, and the relative change of risk corresponding to each line is obtained, and the line that causes the system risk change greater than the threshold (for example, 0.1%) is found as the key line.

[0075] Specifically, the calculation process of the load outage loss risk in this embodiment includes:

[0076] First, an extended probability graph model is constructed, wherein the extended probability graph model is used to solve the power-on probability of the target node according to the power-on probability of the surrounding nodes and the connected lines.

[0077] Then, the grid topology file and grid line failure probability table of the resilient distribution network are obtained; wherein, the grid topology file contains load information, node information and line information, and the grid circuit failure probability table contains the cumulative outage probability of each line at each time point; specifically, the grid topology file contains the distribution network node number, node power, node unit power value data, and also contains line number, line head end, end node number, whether it is a transfer line and other information.

[0078] Next, the cumulative outage probability of each line at each time point is converted into the probability of the line being energized at each moment in the target period. The load information, node information, line information and the probability of the line being energized at each moment in the target period are input into the extended probability graph model to solve the probability of the load node being energized at each moment in the target period.

[0079] Finally, the probability of power supply at the load node at each moment in the target period is converted into the probability of power outage at the load node at each moment in the target period. According to the power outage probability of the load node at each moment in the target period, the load outage loss risk is calculated.

[0080] The following is a detailed description of the construction process of the above extended probability graph model, which specifically includes:

[0081] First, the resilient distribution network is abstracted into node-type devices and line-type devices. Node-type devices include power sources and loads, and line-type devices include power supply lines and transfer lines.

[0082] Then, the association information between the node type device and the line type device is set; wherein the association information includes: the probability of the power supply node being powered is 1, and the load node takes power from its surrounding nodes and depends on the lines connected to it;

[0083] Finally, a node power-on probability equation group is established according to the associated information to obtain an extended probabilistic graph model; in which, the node power-on probability equation group takes the power-on probability of the connected lines as known variables and the power-on probability of the target node and its surrounding nodes as unknown variables.

[0084] Specifically, the node power probability equations are:

[0085]

[0086] In the formula, P(e i ) represents the probability of node i being powered, Ω S represents the power node set, Ω i represents the set of adjacent nodes of node i, Ω l represents the load node set, P(e j→i ) represents the probability of power being present in the direction of line j→i.

[0087] Among them, the calculation formula for the probability of power supply in the line j→i direction is:

[0088] P(e j→i )=P(e j-i )δ ji (4)

[0089] In the formula, P(e j-i ) represents the probability of power supply in the inter-gird line, δ ji It represents the normalized probability that line ji supplies power in the direction j→i.

[0090] Since the extended probabilistic graphical model takes the probability of power supply at a node as a variable to be determined, the problem of solving the probability of power supply at a load can be transformed into the problem of solving a nonlinear system of equations. Since the number of unknowns in the system of equations is equal to the number of equations, the Newton iteration method can be used to solve the nonlinear system of equations to obtain the probability of power supply at the load node at each moment in the target period.

[0091] The power-on probability of the load node at each moment in the target period is converted into the power-off probability of the load node at each moment in the target period. The conversion formula used is:

[0092]

[0093] in, represents the power failure probability of the load node at time t, P t (e i ) represents the probability of power being available at the load node at time t.

[0094] The calculation formula for load outage loss risk is:

[0095]

[0096] Among them, R represents the risk of loss caused by outage of the entire network load, P i Represents the load power of load node i, W i represents the value loss of the unit power load outage at load node i, represents the power failure probability of the load node at time t, Ω l Represents a set of load nodes.

[0097] S130: Taking the direction probability of the key line as the decision variable, the upper and lower bounds of the decision variable as constraints, and the load outage loss risk as the objective function, a line flow direction optimization model is constructed.

[0098] For a reliable distribution network of a communication system, the outage probability of a node is related to the power flow direction of the line supplying it. The addition of the power supply direction converts the original undirected graph model into a directed graph. This embodiment uses a solution based on an optimization method to determine the power flow direction, mainly using the normalized probability of the two directions of the line, that is, by adjusting a continuous variable in the [0,1] interval, thereby changing the different probabilities of the two directions. The expression of the entire line power flow direction optimization model is:

[0099]

[0100] st0≤δ≤1

[0101] P(e i→j )=P(e i-j )δ ij (7)

[0102] Among them, δ represents the directional probability of the critical line, R(x) represents the risk expression of load outage loss, f(x, p′) represents the directional node power probability equation group, P(e i→j ) represents the probability of power supply in the line i→j direction, P(e i-j ) represents the probability of power supply in the line between ij, δ ij It represents the normalized probability that line ij supplies power in the direction i→j.

[0103] It is understandable that each feeder should have two power supply direction probabilities when it is working normally, and in any scenario, the direction probabilities meet the following conditions:

[0104] P(e i→j )=P(e i-j )δ ij (8)

[0105] P(e j→i )=P(e i-j )δ ji =P(e i-j )(1-δ ji ) (9)

[0106] If the objective function value is set to a positive number, the direction probability combination with the lowest risk can be found; if the objective function is set to a negative value, the direction probability combination with the highest system risk can be found.

[0107] S140: According to the line power flow direction optimization model, the power flow direction of the key line is probabilistically optimized to obtain the risk estimation range of the resilient power grid.

[0108] In this embodiment, the power supply directions of all bidirectional power supply feeders are used as optimization variables, and the maximum and minimum values ​​of system risk can be given by adjusting the positive and negative values ​​of the objective function.

[0109] Specifically, after determining the optimization model of the line flow direction, the optimization method is used to perform probability optimization on the key line flow direction to find the maximum and minimum system risk. Since the line outage probability is related to time, in order to simplify the optimization problem, the optimization analysis can be performed only on a single time section, and the solution can be applied to the full time section.

[0110] The following are two examples of optimization methods that can be used to achieve probabilistic optimization in actual applications:

[0111] 1. Single time section approximate gradient method

[0112] For a single time section of the system, the system's single-time risk and the probability of downtime at each node can always be obtained by solving a set of equations. ij , which represents the normalized probability of line ij supplying power in the direction i→j, the Jacobian matrix J for this variable can be obtained Δ , according to the chain rule of partial differentials, we have the following relationship:

[0113]

[0114] The above formula includes the calculation of all nodes in the entire network, because P(e i ) or P(e j ) may affect the power supply of other nodes, thus affecting the risk of the entire network, so we can get:

[0115]

[0116] In this way, the probability of node power failure is related to δ ij The relationship is:

[0117]

[0118]

[0119] For non-directly connected nodes, the corresponding partial derivative can be expressed as:

[0120]

[0121] The transfer diagram of partial differentials according to the chain rule is as follows Figure 4 As shown, through Figure 4 The rule shown in the above formula can obtain the system risk for any δ under the result after each function solution is completed. ij The partial derivatives, Jacobian matrix J ΔIt can be used to accelerate the solution of optimization problems. The above rules are only for the gradient approximation of the line flow direction. The exact gradient value requires partial differentiation of the complete path from the indirect connection node to i or j. Due to the sparsity of the distribution network adjacency matrix, the gradient after four-node approximation can be applied to the solution of small-scale optimization models.

[0122] 2. Numerical differentiation method

[0123] The analytical gradient method has low accuracy in calculating the gradient, and it is easy to make the optimization problem converge to a non-optimal solution. The optimization problem in this embodiment has linear constraints (only upper and lower bound constraints), but has strong nonlinearity in the objective function. Common methods usually use numerical methods such as quasi-Newton method to solve, which does not require the user to provide gradient parameters. The method is faster than genetic algorithm, but relatively slower than other iterative methods where the user provides gradients.

[0124] In view of the complexity of the objective function, it is convenient to use numerical differentiation for calculation, but relatively speaking, optimizing the calculation speed of the objective function itself becomes the main requirement. The gradient is given by numerical differentiation, and the variable dimension is n. Therefore, for the risk assessment function, the Jacobian matrix given by numerical differentiation needs to be calculated at least n+1 times. In order to speed up the calculation speed of numerical differentiation, the main method is to reduce the dimension of decision variables. Various methods can also be adopted in other aspects, such as using parallel computing methods to calculate numerical differentiation. The commonly used L-BFGS algorithm can be converted to distributed implementation through MapReduce, which can greatly improve the efficiency of calculation.

[0125] The advantages of the resilient power grid risk assessment method based on line flow direction optimization are explained in detail below through a specific example.

[0126] This embodiment performs calculation and analysis on two substations of an actual power grid, and takes the critical line judgment threshold as 0.1%. The calculation example information is shown in Table 1, with a total of 2237 nodes.

[0127] Table 1 Example information

[0128] Number of study nodes Number of example lines Number of bidirectional lines 706 715 50 2237 2263 231

[0129] According to Table 1 above, the difficulty of analyzing the actual power grid has been greatly reduced through the screening of bidirectional lines, which is also consistent with the judgment of the actual power grid topology. The difficulty of analysis can be further reduced by screening key lines in the subsequent calculation steps.

[0130] This example uses the minimize() function of scipy to solve the risk interval through numerical optimization. The comparison results are shown in Table 2:

[0131] Table 2 Optimization results of the example

[0132]

[0133] Compared with the unoptimized solution (i.e., baseline risk), the optimization effect of the risk interval is more obvious. After reduction, the number of critical lines can also meet the calculation speed requirements.

[0134] The different thresholds of 2237 node examples are analyzed, and the thresholds are 0.05%, 0.1%, 0.2%, and 0.5% respectively. The solutions are shown in Table 3:

[0135] Table 3. Comparative analysis results of example thresholds

[0136] Number of critical lines (threshold) Risk range (no optimization benchmark risk 0.0148) 48(0.05%) [2.7333e-06,0.0358] 38(0.1%) [0.0018,0.0358] 31(0.2%) [0.0021,0.0358] 25(0.5%) [0.0024,0.0358]

[0137] It can be seen that for the three different thresholds, the upper limit of the risk interval is slightly different, while the lower limit is quite different. In practical applications, the threshold should be selected according to the topology and disaster information of the actual case. In severe disaster scenarios, the nodes are significantly interdependent, and the threshold should be appropriately reduced to fully call on the optimizable resources and reduce system risks. In minor disaster scenarios, the failure probability of the line is low, and most nodes only need to obtain power from a certain path. The threshold can be appropriately increased to identify whether there are lines with greater impact and improve computing efficiency.

[0138] The resilient power grid risk assessment system based on line flow direction optimization provided by the present invention is described below. The resilient power grid risk assessment system based on line flow direction optimization described below and the resilient power grid risk assessment method based on line flow direction optimization described above can be referenced to each other.

[0139] Figure 5 The embodiment of the present invention provides a resilient power grid risk assessment system based on line flow direction optimization, including:

[0140] A bidirectional line extraction module 510 is used to extract bidirectional lines in the resilient power grid and construct a bidirectional line set;

[0141] A key circuit extraction module 520 is used to extract key circuits from the bidirectional circuit set using a sensitivity analysis method;

[0142] The optimization model building module 530 is used to build a line flow direction optimization model using the direction probability of the key line as a decision variable, the upper and lower bounds of the decision variable as constraints, and the load outage loss risk as an objective function;

[0143] The risk estimation module 540 is used to perform probability optimization on the power flow direction of the key line according to the line power flow direction optimization model to obtain the risk estimation interval of the resilient power grid.

[0144] Specifically, in this embodiment, the key path extraction module 520 uses the sensitivity analysis method to extract key paths from the bidirectional paths, including:

[0145] First, the normalized probability of each bidirectional line in the bidirectional line set supplying power in both directions is set to 0.5, and the current initial load outage loss risk is calculated;

[0146] Then, one of the two-way lines in the two-way line set is adjusted to single-direction power supply in turn, and the load outage loss risk after adjustment is calculated respectively;

[0147] Then, according to the adjusted load outage loss risk and the initial load outage loss risk, the relative change of risk corresponding to each bidirectional line is obtained;

[0148] Finally, the relative risk change corresponding to each bidirectional line is compared with the preset risk change threshold, and the bidirectional lines with relative risk change greater than the risk change threshold are extracted as key lines.

[0149] It is understandable that the key line extraction module 520 of this embodiment involves the calculation of load outage loss risk during the process of extracting the key lines. The calculation process specifically includes:

[0150] Firstly, an extended probability graph model is constructed. The known data obtained in advance from the grid topology file and grid line fault probability table of the resilient distribution network are used. The probability of power supply at the load node at each moment in the target period is calculated through the node power supply probability equation group in the extended probability graph model.

[0151] It should be noted that the extended probabilistic graph model needs to abstract the distribution network into two types of equipment: node type and line type. Node type equipment includes power supply and load, and line type equipment refers to power supply line and transfer line. The model assumes that the power supply node has reliable power supply and the probability of power supply is 1; the load node needs to get power from the surrounding nodes and depends on the line connected to the node. Therefore, the probability of power supply at a certain node can be expressed by the power supply probability of the surrounding nodes and the power supply probability of the line, and then a group of equations for node power supply probability is constructed, and the solution of the load power supply probability is converted into the problem of solving the group of equations.

[0152] Then, the probability of power supply at the load node at each moment in the target period is converted into the probability of power outage at the load node at each moment in the target period. According to the power outage probability of the load node at each moment in the target period, the load outage loss risk is calculated.

[0153] It is not difficult to find that the resilient power grid risk assessment system based on line flow direction optimization provided by the embodiment of the present invention extracts key lines through the key line extraction module, and optimizes the line direction that has a greater impact on power grid risks through the line flow direction optimization model obtained by the optimization model construction module. Therefore, the system can effectively improve the reliability of risk assessment results, and can also estimate risk intervals more quickly, and the risk assessment efficiency is greatly improved.

[0154] Figure 6 An example of a physical structure diagram of an electronic device is shown in FIG. Figure 6 As shown, the electronic device may include: a processor (processor) 610, a communication interface (Communications Interface) 620, a memory (memory) 630 and a communication bus 640, wherein the processor 610, the communication interface 620, and the memory 630 communicate with each other through the communication bus 640. The processor 610 may call the logic instructions in the memory 630 to execute a resilient power grid risk assessment method based on line flow direction optimization, the method comprising: extracting bidirectional lines in the resilient power grid and constructing a bidirectional line set; extracting key lines from the bidirectional line set using a sensitivity analysis method; using the direction probability of the key line as a decision variable, the upper and lower bounds of the decision variable as constraints, and the load outage loss risk as an objective function to construct a line flow direction optimization model; according to the line flow direction optimization model, the flow direction of the key line is probability optimized to obtain a risk estimation interval of the resilient power grid.

[0155] In addition, the logic instructions in the above-mentioned memory 630 can be implemented in the form of a software functional unit and can be stored in a computer-readable storage medium when it is sold or used as an independent product. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art or the part of the technical solution, can be embodied in the form of a software product, and the computer software product is stored in a storage medium, including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk and other media that can store program codes.

[0156] On the other hand, the present invention also provides a computer program product, which includes a computer program. The computer program can be stored on a non-transitory computer-readable storage medium. When the computer program is executed by a processor, the computer can execute the resilient power grid risk assessment method based on line flow direction optimization provided by the above methods, the method including: extracting bidirectional lines in the resilient power grid and constructing a bidirectional line set; using a sensitivity analysis method to extract key lines from the bidirectional line set; using the direction probability of the key lines as decision variables, the upper and lower bounds of the decision variables as constraints, and the load outage loss risk as the objective function to construct a line flow direction optimization model; based on the line flow direction optimization model, performing probability optimization on the flow direction of the key lines to obtain a risk estimation interval for the resilient power grid.

[0157] On the other hand, the present invention also provides a non-transitory computer-readable storage medium having a computer program stored thereon. When the computer program is executed by a processor, it is implemented to execute the resilient power grid risk assessment method based on line flow direction optimization provided by the above-mentioned methods, the method comprising: extracting bidirectional lines in the resilient power grid and constructing a bidirectional line set; using a sensitivity analysis method to extract key lines from the bidirectional line set; using the direction probability of the key lines as decision variables, the upper and lower bounds of the decision variables as constraints, and the load outage loss risk as the objective function to construct a line flow direction optimization model; based on the line flow direction optimization model, performing probability optimization on the flow direction of the key lines to obtain a risk estimation interval for the resilient power grid.

[0158] The device embodiments described above are merely illustrative, wherein the units described as separate components may or may not be physically separated, and the components displayed as units may or may not be physical units, that is, they may be located in one place, or they may be distributed on multiple network units. Some or all of the modules may be selected according to actual needs to achieve the purpose of the scheme of this embodiment. Ordinary technicians in this field can understand and implement it without paying creative labor.

[0159] Through the description of the above implementation methods, those skilled in the art can clearly understand that each implementation method can be implemented by means of software plus a necessary general hardware platform, and of course, it can also be implemented by hardware. Based on this understanding, the above technical solution is essentially or the part that contributes to the prior art can be embodied in the form of a software product, and the computer software product can be stored in a computer-readable storage medium, such as ROM / RAM, a disk, an optical disk, etc., including a number of instructions for a computer device (which can be a personal computer, a server, or a network device, etc.) to execute the methods described in each embodiment or some parts of the embodiments.

[0160] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, rather than to limit it. Although the present invention has been described in detail with reference to the aforementioned embodiments, those skilled in the art should understand that they can still modify the technical solutions described in the aforementioned embodiments, or make equivalent replacements for some of the technical features therein. However, these modifications or replacements do not deviate the essence of the corresponding technical solutions from the spirit and scope of the technical solutions of the embodiments of the present invention.

Claims

1. A resilient power grid risk assessment method based on line flow direction optimization, characterized in that: include: Extract bidirectional lines in the resilient power grid and construct a bidirectional line set; A sensitivity analysis method is used to extract key lines from the bidirectional lines; Taking the direction probability of the key line as the decision variable, the upper and lower bounds of the decision variable as the constraint conditions, and the risk of load outage loss as the objective function, a line flow direction optimization model is constructed; According to the line power flow direction optimization model, probability optimization is performed on the power flow direction of the key line to obtain a risk estimation interval of the resilient power grid; The sensitivity analysis method is used to extract key lines from the bidirectional lines, including: The normalized probability of each bidirectional line in the bidirectional line set supplying power in both directions is set to 0.5, and the current initial load outage loss risk is calculated; sequentially adjusting one of the bidirectional lines in the bidirectional line set to unidirectional power supply, and calculating the load outage loss risk after the adjustment respectively; According to the adjusted load outage loss risk and the initial load outage loss risk, the relative change of risk corresponding to each bidirectional line is obtained; The relative risk changes corresponding to each bidirectional line are compared with the preset risk change threshold, and the bidirectional lines whose relative risk changes are greater than the risk change threshold are extracted as key lines; the calculation process of the load outage loss risk includes: An extended probability graph model is constructed, using known data previously obtained from the grid topology file and the grid line fault probability table of the resilient distribution network, and the probability of power supply at the load node at each moment in the target period is calculated through the node power supply probability equation group in the extended probability graph model; The probability of power on at the load node at each moment in the target period is converted into the probability of power off at the load node at each moment in the target period. According to the power off probability of the load node at each moment in the target period, the load outage loss risk is calculated; the equation group of the node power on probability is: ; In the formula, Representation Node i The probability of having electricity, represents the set of power nodes, Representation Node i The set of adjacent nodes of represents the load node set, Indicates line Direction probability of electricity; Among them, the line j → i The calculation formula of the probability of electric current in the direction is: ; In the formula, express The probability of the line being charged is Indicates line j - i towards j → i The normalized probability of power supply in the direction of the load outage is: ; in, R Indicates the risk of loss due to outage of the entire network load. P i Represents load node i The load power, W i Represents load node i The value loss of unit power load outage, express t The probability of power failure at the load node at the moment, represents the load node set; the line flow direction optimization model is: ; in, represents the directional probability of the critical path, represents the load outage loss risk expression, The nodes with directionality have an electrical probability equation system, Indicates line i → j The probability of electric charge in the direction, express i - j The probability of the line being charged is Indicates line i - j towards i → j Normalized probability of directional power supply.

2. A resilient power grid risk assessment system based on line flow direction optimization, characterized in that: include: A bidirectional line extraction module is used to extract bidirectional lines in the resilient power grid and construct a bidirectional line set; A key circuit extraction module, used for extracting key circuits from the bidirectional circuit set by using a sensitivity analysis method; An optimization model building module is used to construct a line flow direction optimization model using the direction probability of the key line as a decision variable, the upper and lower bounds of the decision variable as constraints, and the risk of load outage loss as an objective function; A risk estimation module, used to perform probability optimization on the power flow direction of the key line according to the line power flow direction optimization model, and obtain a risk estimation interval of the resilient power grid; The key line extraction module is specifically used for: The normalized probability of each bidirectional line in the bidirectional line set supplying power in both directions is set to 0.5, and the current initial load outage loss risk is calculated; sequentially adjusting one of the bidirectional lines in the bidirectional line set to unidirectional power supply, and calculating the load outage loss risk after the adjustment respectively; According to the adjusted load outage loss risk and the initial load outage loss risk, the relative change of risk corresponding to each bidirectional line is obtained; Compare the relative risk changes corresponding to each bidirectional line with the preset risk change threshold, and extract the bidirectional lines whose relative risk changes are greater than the risk change threshold as key lines; The calculation process of the load outage loss risk includes: An extended probability graph model is constructed, using known data previously obtained from the grid topology file and the grid line fault probability table of the resilient distribution network, and the probability of power supply at the load node at each moment in the target period is calculated through the node power supply probability equation group in the extended probability graph model; The probability of power supply at the load node at each moment in the target period is converted into the probability of power failure at the load node at each moment in the target period, and the load outage loss risk is calculated according to the power failure probability of the load node at each moment in the target period; The node has electric probability equations: ; In the formula, Representation Node i The probability of having electricity, represents the set of power nodes, Representation Node i The set of adjacent nodes of represents the load node set, Indicates line j → i Direction probability of electricity; Among them, the line j → i The calculation formula of the probability of electric current in the direction is: ; In the formula, express j - i The probability of the line being charged is Indicates line j - i towards j → i Normalized probability of directional power supply; The calculation formula for the load outage loss risk is: ; in, R Indicates the risk of loss due to outage of the entire network load. P i Represents load node i The load power, W i Represents load node i The value loss of unit power load outage, express t The probability of power failure at the load node at the moment, represents the set of load nodes; The line flow direction optimization model is: ; in, represents the directional probability of the critical path, represents the load outage loss risk expression, The nodes with directionality have an electrical probability equation system, Indicates line i → j The probability of electric charge in the direction, express i - j The probability of the line being charged is Indicates line i - j towards i → j Normalized probability of directional power supply.

3. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that: When the processor executes the program, the steps of the resilient power grid risk assessment method based on line flow direction optimization as described in claim 1 are implemented.

4. A non-transitory computer-readable storage medium having a computer program stored thereon, characterized in that: When the computer program is executed by the processor, the steps of the resilient power grid risk assessment method based on line flow direction optimization as claimed in claim 1 are implemented.

5. A computer program product, comprising a computer program, characterized in that When the computer program is executed by the processor, the steps of the resilient power grid risk assessment method based on line flow direction optimization as claimed in claim 1 are implemented.

Citation Information

Cited By

  • Explanatable analysis method for comprehensive toughness of hybrid AC / DC power system

    CN121881081A

  • Grid resiliency quantification using multi-criteria decision model

    EP4815256A1

  • Grid resiliency quantification using multi-criteria decision model

    WO2026202006A1