A power grid key line identification method based on topology and operating characteristics

Abstract

The application discloses a power grid key line identification method based on topology and operation characteristics and belongs to the technical field of power grid key line identification. In the aspect of operation characteristics, the improved webpage link algorithm is adopted, the product of the initial load rate and the historical failure rate of each line is taken as an initial value of the algorithm, the flow fluctuation caused by the disconnection of each line on other lines is sequentially counted, and a link matrix required by the algorithm is constructed. Through the link matrix, a propagation matrix and a fragile matrix are obtained, the line key propagation value vector and the key fragile value vector of the power grid are calculated by using the initial value and the propagation / fragile matrix, the line key value is obtained by weighted summation of the key propagation value and the key fragile value of each line, the importance of the line in the local topology of the power grid is considered, the comprehensive degree of the nodes at the two ends of the line is constructed, the line key value correction coefficient is formed, and then the line comprehensive key value is constructed. The comprehensive key values of all lines are sorted, and the lines in the front row are identified as the key lines. The application can obtain the key degree of each line of the power grid and has certain effectiveness.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of power grid key line identification, and particularly relates to a power grid key line identification method based on topology and operation characteristics. BACKGROUND

[0002] In recent years, a number of power grid blackouts have occurred worldwide. Research has shown that the cause of the blackout is the transfer of power flow in the power grid after the failure of some key initial lines, the redistribution of power on each line, the overload of some lines, and the subsequent disconnection of lines, which ultimately leads to a system blackout. Due to the increasing demand for electricity, the scale of the power grid is expanding, and the complexity is gradually increasing. Therefore, in the identification of key lines in the power grid, the topology structure is considered to identify the key lines that will cause a blackout. This is of great significance for the prevention and control of cascading failures in the future and the improvement of the reliability of the power grid.

[0003] For power grid key identification, document 1 (Wei Zhenbo, Ju Qi, Yi Gangchun, et al. Power grid cascading failure key line identification method based on improved LeaderRank algorithm [J]. High Voltage Technology, 2021, 47(12): 4265-4273.) proposes a power grid key line identification method combining cascading failure consequence evaluation and improved LeaderRank algorithm. This method is based on N-1 check results, quantifies the relationship between lines, reflects the vulnerability of the line itself, and defines background nodes to approximate the correlation between first-order failures and subsequent failures. The method organically combines line vulnerability analysis and failure consequence analysis. However, this document only considers the mutual influence between lines when considering the influencing factors of line failure, and does not consider the probability of hidden failures caused by natural environmental factors, component aging, etc. in actual situations, and the model is not comprehensive enough.

[0004] Document 2 (Nan B, Dong Shufeng, Xu Chengsi, et al. Identification of key lines in power grid based on improved deviation maximization method[J]. Power System Technology, 2022, 46(10):4076-4084.) aims to address the problem that existing key line identification methods rarely consider the information of nodes at both ends of the line and the insufficient consideration of line contingency faults. By considering the voltage stability, line transmission capacity, line failure rate, and topological importance of nodes, four indicators of voltage stability, capacity margin, real-time failure rate, and topological connection degree are established to identify the importance of lines. The improved deviation maximization method is used to objectively weight the four indicators to obtain the comprehensive importance index of the line. The sorting of the indicators gives the ranking of the key lines. However, this method does not consider the role of lines in the actual cascading failure propagation mechanism during the identification process, and the simple weighted sum of the line criticality indicators from completely different angles is not reasonable.

[0005] Document 3 (Pansong, Li Changcheng, Kang Haipeng, et al. Identification of key lines in power grid considering outage probability entropy and network structure entropy[J]. Power System Automation, 2025, 49(1): 139-149.) proposes a method for identifying key lines in power grid considering outage probability entropy and network structure entropy. The outage probability entropy includes outage probability distribution entropy and outage probability change entropy. The outage probability entropy model is used as an indicator to measure the distribution of power grid outage probability, reflecting the influence of lines in different operating states on the identification results. The network structure entropy can measure the pivotal role of the target line in power transmission and the uniformity of the power grid structure after the outage. The three entropy indicators are objectively weighted using the fuzzy entropy weight method to obtain the comprehensive key line index. This comprehensive index considers the importance of lines in the overall operation of the power grid, the operation of local lines, and the stability of the power grid structure. However, this document does not consider the role of lines in cascading failures and only considers the influence of line disconnection probability without considering the consequences of key line disconnection. The method ignores key lines with low occurrence probability but serious consequences for the power grid, making the method relatively single.

[0006] The existing power grid key line identification methods are summarized from two aspects, the first aspect is based on power system state analysis, the method mainly focuses on the influence of line fault on electrical quantity in the power grid, and the key degree of the line is judged according to the change of the electrical quantity, such as node voltage, power on the line, etc., which focuses on the operation state of the power grid, and the cascading failure evolution mechanism is deduced based on the power flow calculation, and the key degree is evaluated by using the fault data; the second aspect is based on complex network to identify the key line, the topological structure of the network is considered, and indexes such as average path length, node degree and electrical betweenness are proposed to evaluate the key degree, in the existing research methods, although the calculation result of the power system state analysis method can reflect the influence of the key line disconnection on the system, the calculation time cost is high, the efficiency is low, and the real-time requirement of the safety analysis cannot be met; the analysis method based on the complex network can realize the rapid identification of the key line of the power grid, but the dynamic characteristics and physical details of the power grid are ignored, the electrical characteristics of the power grid cannot be well reflected, and the accuracy of the identification is insufficient.

[0007] In order to solve the problem, the application provides a power grid key line identification method based on topological and operation comprehensive characteristics. SUMMARY

[0008] The application aims to provide a power grid key line identification method based on topological and operation characteristics to solve the problems in the background art, so that the structural characteristics of the line in the system and the electrical characteristics of the power grid can be considered in the analysis process.

[0009] In order to achieve the above-mentioned purpose, the application adopts the following technical scheme:

[0010] A power grid key line identification method based on topological and operation characteristics, comprising:

[0011] S1, network modeling: an electric grid is abstracted as a topological network Z, the bus and transformer in the power grid are taken as nodes in the topological network Z, and the transmission line is taken as an edge in the topological network Z, the degree, the number of next adjacent nodes and the number of nodes in the two-step neighborhood of each node in the topological network Z are counted, and the comprehensive degree of each node is calculated;

[0012] S2, fault rate statistics: a large number of cascading failure simulation experiments are carried out by using the Monte Carlo method, the initial fault line is randomly disconnected, the power flow calculation is carried out on the power grid after the fault occurs, the line operation reliability model is used to analyze the probability of line outage, and it is judged whether to continue to occur cascading failure, one accident chain is generated in one cascading failure simulation, and the frequency of each line fault in the simulation experiment is counted;

[0013] S3, initial key value vector construction: based on the power flow calculation result, the initial load rate of each line is calculated, and the product of the initial load rate of each line and its historical failure rate is taken as the initial key value, and then the initial key values of all lines are constructed into the initial key value vector of the entire power grid line;

[0014] S4, link matrix construction: simulate N-1 fault, record the power fluctuation of each line as the link matrix element, construct the link matrix of the entire network line, and process the obtained link matrix to obtain the propagation matrix and the fragile matrix;

[0015] S5, iterative convergence calculation: multiply the initial key value vector obtained in S3 with the propagation / fragile matrix obtained in S4 iteratively until the difference between adjacent iteration results is less than the set error, to obtain the final key propagation value vector and the key fragile value vector;

[0016] S6, key value vector construction: the key propagation value and the key fragile value of each line are weighted and summed to obtain the key value of the line, and the key values of all lines of the power grid are constructed into the key value vector of the line;

[0017] S7, comprehensive key value calculation: considering the local topological importance of the line in the power grid, taking the square root of the product of the comprehensive degrees of the two end nodes of the line as the correction coefficient, correcting the key value of the line, and multiplying the correction coefficient with the key value of the line to obtain the comprehensive key value of the line;

[0018] S8, key line identification: the comprehensive key value obtained in S7 is sorted in descending order, and the lines ranked in the front are identified as key lines.

[0019] Preferably, S1 specifically includes the following contents:

[0020] Considering the local topological connection relationship of the node, the power grid is abstracted into a topological graph with only point and edge connection, the edge in the topological graph represents the transmission line in the original power grid, and the point represents the bus, load node, generator node, and transformer in the original power grid. i The degree D of the node i is calculated according to the following formula:

[0021] (1)

[0022] In the formula, N is the number of nodes in the network; n ij represents the connection relationship between the node i and the node j , if the two nodes are connected, then n ij is 1, if there is no connection relationship, then n ijis 0;

[0023] The importance of connection of each node in the topology graph is measured by a node comprehensive degree, and the node comprehensive degree of each node is calculated according to the following formula: i C Di is:

[0024] (2)

[0025] In the formula, C i is the number of secondary neighbor nodes of the node; i Q i is the number of nodes in the two-step neighborhood of the node. i

[0026] Preferably, the S2 specifically includes the following contents:

[0027] The Monte Carlo method is used to perform a cascading failure simulation experiment, an initial fault line is randomly disconnected, power flow calculation is performed on the power grid after the fault occurs, the power on each line is obtained, whether a line overload occurs is judged according to the power size, and the risk of other lines around the disconnected line occurring a hidden fault is considered, whether the next-order disconnection will occur is judged according to whether a line overload occurs or a hidden fault occurs; if a line is disconnected, the cascading failure continues to occur until the power balance of the power grid is re-established by cutting the load or the power balance of each island after splitting is realized, and at this time, there is no line overload or hidden fault, and the cascading failure ends, as a cascading failure simulation experiment; the cascading failure simulation experiment is performed multiple times, the frequency of disconnection of each line in the simulation experiment is counted, that is, the proportion of the number of cascading failure chains containing a line in the total number of cascading failure chains is taken as the historical failure rate of the line; in the actual power grid, the failure rate of the line can be directly counted according to the historical operation situation;

[0028] The cascading failure simulation experiment specifically includes the following steps:

[0029] 1) Import the initial network data, and randomly select an initial fault line;

[0030] 2) After disconnecting a line, update the network topology structure, and judge whether the network is split to form an island;

[0031] 3) If no island is formed, calculate the probability of disconnection of each line p , as follows:

[0032] (3)

[0033] In the formula, p 0 is the probability of line hidden failure, which is generally set to 0.001;​​​f P is the real-time power flow of the line; F P is the power flow limit of the line in normal operation; F max P is the thermal stability limit of the line; p 1 is the probability of the line being disconnected after exceeding the thermal stability limit; wherein only the lines connected to the line disconnected in the previous stage are defined to have the probability of having a hidden fault;

[0034] If an island is formed, the power balance within the island is first performed for each island, and then the probability of each line being disconnected is calculated according to formula (3);

[0035] 4) After each disconnection, the power balance within the grid or the island is performed to determine whether the power generated by the generators in the grid or the island is balanced with the load at this time; if that is, the total power generated by the generators in the grid or the island is less than the total load size at this time, load shedding needs to be performed on the grid, and the load size of the updated load node d is:

[0036] (4)

[0037] In the formula, d is the load node; L is the set of load nodes; P Ld is the initial load of the load node d ; is the generator node; g is the set of generator nodes; G P Gg is the power generated by the generator node g ; that is, the updated load size of each node is the product of the original load size and the power shortage ratio;

[0038] 5) After balancing the power within the island or the entire grid, it is determined whether the simulation continues to generate cascading disconnection lines this time; if a new line is disconnected due to overload, the overloaded line is disconnected, and step 3) is returned until the cascading failure ends, and the accident chain and loss of load of this cascading failure are counted;

[0039] 6) After the cascading failure simulation, the failure rate of each line in the grid is calculated, and the failure rate k of the line V k is:

[0040] (5)

[0041] In the formula, W is the total number of fault chains; W ​​k The number of fault chains of the line k .

[0042] Preferably, the S3 specifically comprises the following contents:

[0043] Import the initial data of the power grid, including line parameters, node parameters, generator parameters, load parameters, etc., perform power flow calculation to obtain the initial load rate of each line in the power grid, multiply the initial load rate of each line by the historical failure rate of the line to obtain the initial key value required for improving the SALSA algorithm, and then construct the initial key value vector of all lines in the power grid as the initial key value vector of the lines;

[0044] The SALSA algorithm gives the initial key value of each line , and then calculates the initial key value of the line k :

[0045] (6)

[0046] In the formula, f k The real-time power flow of the line k ; F k The power flow limit of the line k when it is in normal operation;

[0047] The initial propagation value vector of each line , the initial fragile value vector are obtained after calculation as follows:

[0048] (7)

[0049] In the formula, l is the total number of lines.

[0050] Preferably, the S4 specifically comprises the following contents:

[0051] Assume that each line of the initial power grid has an N-1 fault, obtain the power fluctuation on other lines in the power grid after the line is disconnected, and take the power fluctuation on other lines as the corresponding row elements in the line linkage matrix of the line in turn; disconnect each line in the power grid in turn to construct the line linkage matrix of the entire network; normalize the rows and columns of the constructed line linkage matrix, multiply the normalized row matrix by the transpose of the normalized column matrix to obtain the propagation matrix, and multiply the transpose of the normalized column matrix by the normalized row matrix to obtain the fragile matrix;

[0052] Calculate the linkage matrix B of the SALSA algorithm, and the element b ij ​The calculation formula is:

[0053] (8)

[0054] In the formula, f ij For the line i Disconnected line j Active power; f j0 For the line j Initial power; S This represents the total load of the power grid; the size of each element in the link matrix indicates the impact of a disconnection on the remaining lines.

[0055] After obtaining the initial link matrix, the propagation matrix H and vulnerability matrix A are calculated from the link matrix to further evaluate the propagation criticality and vulnerability criticality of the route:

[0056] (9)

[0057] In the formula, B r B is the row-normalized matrix of the link matrix; c This is the column normalized matrix of the link matrix.

[0058] Preferably, step S5 specifically includes the following:

[0059] The initial critical value vector of the line is multiplied by the propagation matrix and the vulnerability matrix respectively to obtain the first-order critical propagation value vector and the first-order critical vulnerability value vector. Then, the first-order critical propagation value vector is multiplied by the propagation matrix to obtain the second-order critical propagation value vector, and the first-order critical vulnerability value vector is multiplied by the vulnerability matrix to obtain the second-order critical vulnerability value vector. This method is repeated iteratively until the convergence condition is met, that is, the difference between the critical propagation value vectors after two adjacent iterations is less than the set maximum error value, and the difference between the critical vulnerability value vectors after two adjacent iterations is less than the set specified error value. The final line critical propagation value vector and line critical vulnerability value vector are obtained. The specific calculation method is as follows:

[0060] Calculate the propagation value vector of the path using the defined propagation matrix and vulnerability matrix. π h and fragile value vector π a The calculation formula is as follows:

[0061] (10)

[0062] In the formula, For the first m The propagation value vector of the path during the next iteration is composed of the propagation values ​​of all paths. For the firstm The vulnerability vector of a line during the next iteration is composed of the vulnerability vectors of all lines.

[0063] Repeat equation (10) until the convergence condition is met. ,in The maximum error value is obtained after convergence. π h and π a , which are the propagation value vector and vulnerability value vector of the power grid line, respectively.

[0064] Preferably, the key value calculation in S6 specifically refers to:

[0065] By calculating the propagation value and vulnerability value of each line, the larger the propagation value or vulnerability value of a line, the more likely that line is to act as a propagation line or a vulnerable line. Critical propagation lines are more likely to worsen fault propagation and expand the scale of line faults. Critical vulnerable lines are more susceptible to the impact of other line faults in a cascading failure. However, since a faulty line can act as a vulnerable line for the preceding faulty line and also as a propagation line for the next faulty line during the cascading failure propagation process, the line is defined... k Key Value Indicators π ( k This comprehensively reflects the propagation capability and vulnerability of line deterioration faults:

[0066] (11)

[0067] In the formula, ω 1 is the line k As the weight of the propagation path; ω 2 is the line k As a weight for vulnerable lines.

[0068] Preferably, the method for calculating the comprehensive key value in S7 is as follows:

[0069] Lines based on topology and operational characteristics k Comprehensive key value :

[0070] (12)

[0071] In the formula, C Dk ( p )and C Dk ( q ( ) is the line k Two-end nodes p , q The overall level of comprehensiveness.

[0072] Preferably, the key line ranking method in S8 is:

[0073] After obtaining the comprehensive key values of all lines in the power grid, the comprehensive key values are ranked from large to small to obtain the order of the line comprehensive key values, and the lines ranked in the front have a higher key degree in the power grid.

[0074] Compared with the prior art, the application provides a power grid key line identification method based on topology and operating characteristics, which has the following beneficial effects:

[0075] (1) The application proposes a web link algorithm (SALSA algorithm) to identify the key lines in the power grid, which analogizes the cascading failure propagation in the power grid to the mutual jump between web pages on the Internet, and improves the parameter setting in the method to be suitable for the power grid. In the initial value setting of the algorithm, the historical failure probability and the initial load rate of the line are considered, which conforms to the understanding of the initial key value of the line, that is, the line with high initial load rate and high historical failure probability is more likely to become the disconnected line in the failure; in the setting of the link matrix, the correlation between lines is considered, that is, the influence of the disconnection of a line on the power flow on other lines is defined as the element in the matrix, and the change of the electrical quantity in the power system during the failure propagation process is considered;

[0076] (2) The application proposes to use the SALSA algorithm to identify the key lines, which can not only identify the key propagation lines that will worsen the failure propagation and expand the line failure scale in the cascading failure propagation process, but also identify the key fragile lines that are easily affected by other line failures in the cascading failure process. Although this type of line has less influence on other lines, it may play an important role in promoting the failure in the cascading failure propagation process and have a greater impact on the system. The lines identified by the algorithm are more comprehensive;

[0077] (3) The application proposes a method for correcting the key value of the line by using the comprehensive degree of the nodes at both ends of the line, which considers the topological importance of the line in the power grid. The comprehensive degree of the node not only considers the size of the node itself, but also considers the number of secondary neighbor nodes and nodes in the two-step neighborhood, and comprehensively considers the local topological connection relationship, which helps to identify the lines located in the key transmission channel or the lines with higher voltage level. This type of line is difficult to identify its key due to its large power margin setting and is not easy to disconnect, but once disconnected, it will have extremely serious consequences. By correcting the key value of the line through the node comprehensive degree, the lines located in the key position in the power grid can be more comprehensively identified. BRIEF DESCRIPTION OF DRAWINGS

[0078] Figure 1A flow chart of a power grid key line identification method based on topology and operating characteristics according to the present application;

[0079] Figure 2 An IEEE39 node system diagram in Example 1 of the present application;

[0080] Figure 3 An IEEE39 node network topology diagram in Example 1 of the present application;

[0081] Figure 4 Line key propagation values and key vulnerable values identified in Example 1 of the present application;

[0082] Figure 5 Line key values identified in Example 1 of the present application;

[0083] Figure 6 Line comprehensive key values identified in Example 1 of the present application;

[0084] Figure 7 Efficiency reduction ratios under network transmission in Example 1 of the present application;

[0085] Figure 8 Loss of load in Example 1 of the present application. DETAILED DESCRIPTION

[0086] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, but not all the embodiments.

[0087] Please refer to Figure 1 The present application proposes a power grid key line identification method based on topology and operating characteristics. Firstly, the improved SALSA algorithm considers the influence of line disconnection on other line power flow in the calculation process, and considers the probability of line disconnection in the initial value setting in the process of cascading failure, fully considering the electrical operating characteristics of the line. Secondly, the SALSA algorithm can not only identify the key propagation line in the power grid, i.e. the line that has a greater impact on the power grid after disconnection, but also identify the key vulnerable line that is easily affected by other lines and then fails. This vulnerable line may be an important link that triggers the propagation of power grid cascading failure. Finally, the topology importance of the line in the power grid is considered, and the comprehensive degree of the nodes at both ends of the line is used as a correction coefficient to correct the line key value, which comprehensively reflects the influence of the line on the global and local topology structure of the power grid. The specific steps include the following:

[0088] Step 1:

[0089] The power grid is abstracted as a topology network Z, the bus and transformer in the power grid are taken as nodes in the topology network, and the transmission line is taken as an edge in the topology network. The degree, number of next adjacent nodes, and number of nodes in two-step neighborhood of each node in the topology network are counted, and the comprehensive degree of each node is calculated.

[0090] In order to consider the local topology connection relationship of the node, the power grid is abstracted as a topology graph only connected by points and edges. The edge in the topology graph represents the transmission line in the original power grid, and the point represents the bus, load node, generator node, transformer, etc. in the original power grid. The degree of the node in the topology graph is: i D i

[0091] (1)

[0092] In the formula, N is the number of nodes in the network. N n ij The connection relationship between the node and the node is represented by, if the two nodes are connected, then i j n ij is 1, if there is no connection relationship, then n ij is 0.

[0093] The importance of the connection of each node in the topology graph is measured by the comprehensive degree of the node. The comprehensive degree C of the node is: i Di

[0094] (2)

[0095] In the formula, is the number of next adjacent nodes of the node. C i i Q i is the number of nodes in the two-step neighborhood of the node. i

[0096] Step 2:

[0097] ​​​​​​​​​​Monte Carlo method is used to simulate a large number of chain failure experiments, and the initial fault line is randomly disconnected. The power on each line is obtained by power flow calculation after the fault occurs. Whether the line is overloaded is determined according to the power size, and the risk of hidden failure of other lines around the disconnected line is considered. According to whether the line is overloaded or hidden failure occurs, it is determined whether the next stage of line disconnection will occur. If a line is disconnected, the chain failure continues to occur until the power balance of the grid is restored or the power balance of each island after splitting is realized. At this time, there is no line overload or hidden failure, and the chain failure ends as a chain failure simulation experiment. Multiple chain failure simulation experiments are carried out, and the frequency of each line disconnection in the simulation experiment is counted, that is, the proportion of the number of chain failure chains containing a certain line in the total number of chain failure chains, which is the historical failure rate of the line. In the actual power grid, the failure rate of the line can be directly counted according to the historical operation;

[0098] The steps of simulating chain failure are as follows:

[0099] 1) Import the initial network data, and randomly select the initial fault line;

[0100] 2) After disconnecting a line, update the network topology structure, and judge whether the network is split to form an island;

[0101] 3) If no island is formed, calculate the probability of line disconnection of each line p , as follows:

[0102] (3)

[0103] In the formula, p 0 is the probability of line hidden failure, which is generally set to 0.001; f is the real-time power flow of the line; F is the power flow limit of the line in normal operation; F max is the thermal stability limit of the line; p 1 is the probability of line disconnection after exceeding the thermal stability limit, which is 1; wherein only the probability of hidden failure of the line connected to the line disconnected in the last stage is defined;

[0104] If an island is formed, the power balance of each island is first carried out, and then the probability of line disconnection of each line is calculated according to the above formula;

[0105] 4) After each disconnection, the power balance of the grid or the island is carried out to judge whether the power generated by the generator and the load in the grid or the island are balanced. If , that is, the total generator output in the grid or island is less than the total load size, load shedding is needed for the grid, and the load node is updatedd Load size for:

[0106] (4)

[0107] In the formula, d For load nodes; L For the set of load nodes; P Ld For load nodes d The initial load; g For generator nodes; G For the set of generator nodes; P Gg For generator nodes g The amount of power emitted; that is, the updated load size of each node is the product of the original load size and the power deficit ratio;

[0108] 5) After balancing the power within the island or the entire power grid, determine whether the simulation will continue to generate cascading disconnections. If a new line disconnects due to overload, disconnect the overloaded line and return to step 3) until the cascading fault ends, and statistically analyze the accident chain and load loss of this cascading fault.

[0109] 6) After simulating cascading failures, calculate the failure rate of each line in the power grid. k Failure rate V k for:

[0110] (5)

[0111] In the formula, W This represents the total number of fault chains. W k For including lines k The number of fault chains;

[0112] Step 3:

[0113] Import the initial data of the power grid, including line parameters, node parameters, generator parameters, load parameters, etc., perform power flow calculations to obtain the initial load rate of each line in the power grid. Multiply the initial load rate of each line by the historical failure rate of the line as the initial key value required to improve the SALSA algorithm. Construct the initial key value vector of all lines in the power grid.

[0114] The SALSA algorithm requires initial key values ​​for each path. ,line initial key value :

[0115] (6)

[0116] wherein, f k is the real-time power flow of the line k ; F k is the power flow limit of the line k when it is in normal operation;

[0117] After the calculation, the initial propagation value vector of each line , and the initial vulnerable value vector are obtained as follows:

[0118] (7)

[0119] wherein, l is the total number of lines;

[0120] Step 4:

[0121] For each line of the initial power grid, it is assumed that an N-1 fault occurs in the line, and the power fluctuation on other lines in the power grid after the line is disconnected is obtained. The power fluctuation on the lines other than the disconnected line is sequentially taken as the corresponding row element in the line linkage matrix of the line. Each line in the power grid is sequentially disconnected, and the line linkage matrix of the entire network is constructed. The row and column of the constructed line linkage matrix are normalized, the matrix after the row normalization is multiplied by the transpose of the matrix after the column normalization to obtain the propagation matrix, and the transpose of the matrix after the column normalization is multiplied by the matrix after the row normalization to obtain the vulnerable matrix.

[0122] The linkage matrix B of the SALSA algorithm is calculated, and the element b ij of the linkage matrix B is:

[0123] (8)

[0124] wherein, f ij is the active power on the line i after the line is disconnected; j f j0 is the initial power on the line j ; S is the total load of the power grid; the size of the element in the linkage matrix represents the influence of a disconnected line on the remaining lines;

[0125] After the initial linkage matrix is obtained, the propagation matrix H and the vulnerable matrix A are calculated through the linkage matrix to further evaluate the propagation critical value and the vulnerable critical value of the line:

[0126] (9)

[0127] ​where B r is the row-normalized matrix of the link matrix; B c is the column-normalized matrix of the link matrix;

[0128] Step 5:

[0129] The line initial key value vector is multiplied by the propagation matrix and the vulnerability matrix respectively to obtain a first-order key propagation value vector and a first-order key vulnerability value vector, then the first-order key propagation value vector is multiplied by the propagation matrix to obtain a second-order key propagation value vector, and the first-order key vulnerability value vector is multiplied by the vulnerability matrix to obtain a second-order key vulnerability value vector, and the iteration is repeated in this way until the convergence condition is met, that is, the difference between the key propagation value vectors after adjacent two iterations is less than the set maximum error value, and the difference between the key vulnerability value vectors after adjacent two iterations is less than the set specified error value, to obtain the final line key propagation value vector and the line key vulnerability value vector;

[0130] The propagation value vector and the vulnerability value vector of the line are calculated using the defined propagation matrix and vulnerability matrix and the vulnerability value vector as follows:

[0131] (10)

[0132] wherein, is the propagation value vector of the line in the i-th iteration calculation, which is composed of the propagation values of all lines; m is the vulnerability value vector of the line in the i-th iteration calculation, which is composed of the vulnerability value vectors of all lines; m The above formula is iterated repeatedly until the convergence condition is met

[0133] wherein, is the maximum error value, and the obtained π h and π a is the propagation value vector and the vulnerability value vector of the power grid line;

[0134] Step 6:

[0135] The key propagation value and the key vulnerability value of each line are weighted and summed to obtain the key value of the line, and the key values of all lines of the power grid are composed to obtain the line key value vector;

[0136] ​​The propagation value and the vulnerability value of each line are calculated, and the greater the propagation value or the vulnerability value corresponding to the line, the more likely the line is to be a propagation line or a vulnerable line, and the more likely the key propagation line is to deteriorate the fault propagation and expand the line fault scale; the key vulnerable line is more likely to be affected by other line faults in the cascading failure, but since the fault line can be both a vulnerable line of the previous fault line and a propagation line of the next fault line in the cascading failure propagation process, in order to comprehensively reflect the deterioration of the fault propagation ability and the vulnerability of the line, the line k Key value index π ( k ):

[0137] (11)

[0138] wherein, ω 1 is the weight of the line as a propagation line; k 2 is the weight of the line as a vulnerable line; ω k

[0139] Step 7:

[0140] Considering the local topological importance of the line in the power grid, the comprehensive degrees of the nodes at both ends of the line are calculated, and the square root of the product of them is taken as a correction coefficient, the line key value is corrected by multiplying the correction coefficient and the line key value to obtain the comprehensive key value of the line;

[0141] Comprehensive key value of the line k based on the topological and operational comprehensive characteristics:

[0142] (12)

[0143] wherein, C Dk p ) and C Dk q are the comprehensive degrees of the nodes at both ends of the line; k p q

[0144] Step 8:

[0145] After obtaining the comprehensive key values of all lines in the power grid, the comprehensive key values are sorted from large to small to obtain the order of the line comprehensive key values, and the line with a higher ranking is more critical in the power grid. The lines ranked in the front are identified as key lines.

[0146] ​​​​​​​​Based on the above, the following specific examples and drawings are used to illustrate the method for identifying key lines of power grid based on topology and operating characteristics, and the specific content is as follows.

[0147] Embodiment 1:

[0148] Please refer to Figure 2 In order to verify the effectiveness of the protection method, specific experiments are designed to illustrate the effectiveness of the method for identifying key lines of power grid based on topology and operating characteristics, and the specific content is as follows:

[0149] (1) Import IEEE39 nodes (IEEE39 node system diagram as shown in Figure 2 , IEEE39 node network topology diagram as shown in Figure 3 ), calculate the initial load rate of each line, use Monte Carlo method, and perform 10000 times of cascading failure simulation experiment on the system, and count the line breaking rate as the historical failure rate, the results are shown in Table 1.

[0150] Table 1 Line historical failure rate and initial load rate

[0151]

[0152] Through the initial load rate and failure rate of the line, the initial key value of each line is obtained, and the initial key value array is normalized to obtain the initial key value array, as shown in Table 2.

[0153] Table 2 Line initial key value

[0154]

[0155] (2) N-1 failure is performed on each line to obtain the power flow fluctuation on other lines, and a link matrix B is constructed, the link matrix is normalized in row and column, and the propagation matrix H and the vulnerable matrix A are calculated:

[0156]

[0157] (3) The initial value of each line is multiplied by the propagation matrix and the vulnerable matrix to obtain the first-order line key propagation value and the key vulnerable value (as shown in Figure 4 ), the first-order line key value is multiplied by the propagation matrix and the vulnerable matrix respectively to obtain the second-order, and the iteration is repeated according to this method until the convergence condition is met, in this embodiment, the maximum error value ε =0.000001, the key propagation value vector and the key vulnerable value vector of the line after iteration are shown in Table 3.

[0158] Table 3 Line key propagation value and key vulnerable value

[0159]

[0160] (4) In this embodiment, the line disconnection causes other line overload, i.e. the line propagation, which plays a more important role in causing the chain failure to occur, and the key propagation value weight is set ω h = 0.7, the key vulnerability value weight ω a = 0.3, please refer to Figure 5 The key propagation value and the key vulnerability value are weighted and summed to obtain the line key value vector, as shown in Table 4.

[0161] Table 4 Line key value vector

[0162]

[0163] (5) The degree, the number of second neighbors, and the number of nodes in the two-step neighborhood of all nodes in the network are counted, as shown in Table 5, and the comprehensive degree of each node is calculated using the obtained data, as shown in Table 6.

[0164] Table 5 Degree, number of second neighbors, and number of nodes in two-step neighborhood of each node

[0165]

[0166] Table 6 Comprehensive degree of each node

[0167]

[0168] (6) Please refer to Figure 6 The comprehensive degree of the nodes at both ends of the line is used as a correction coefficient of the line key value, and the key value is corrected to obtain the corrected line comprehensive key value, as shown in Table 7.

[0169] Table 7 Line comprehensive key value

[0170]

[0171] (7) The obtained line comprehensive key value is sorted from large to small to obtain the ranking of the key line, and the top ten lines are selected as the key lines in this embodiment, as shown in Table 8.

[0172] Table 8 Key line ranking

[0173]

[0174] Please refer to Figure 7-8In order to verify the effectiveness and accuracy of the method, the random theory and entropy theory combination method and the LeaderRank method are used to perform the loss of load comparison experiment and the network transmission efficiency comparison experiment. The top ten lines identified by the three methods are compared in terms of the size of the loss of load and the network transmission efficiency reduction ratio:

[0175] Table 9 Comparison experiment

[0176]

[0177] In this example, the initial data of the network is first imported, the load rate of each line in the initial network is calculated, the Monte Carlo method is used to perform 10000 times of cascading failure simulation, the line breakage rate of each line is counted as the historical failure rate of the line, and the product of the load rate and the historical failure rate is taken as the initial value of the SALSA algorithm. By performing N-1 failure on each line, the power fluctuation on other lines is taken as the element in the link matrix of the line. The link matrix is normalized in row and column, and the propagation matrix and the vulnerable matrix are calculated. The critical value vector is obtained by operating the initial value and the matrix, and the final line critical propagation value and the critical vulnerable value are obtained by repeated iteration until the convergence condition is met. The critical value of each line is calculated by weighted summation. The importance of the line in topology is considered, the comprehensive degree of the nodes at both ends of the line is used to construct the correction coefficient of the line critical value, the critical value is corrected to obtain the comprehensive critical value of the line considering the topology and operation comprehensive characteristics. The comprehensive critical value is sorted in descending order, and the lines with higher ranking are more critical.

[0178] In this example, the line 27 is ranked first in the identified critical line. According to the system connection diagram, line 27 is an important power flow transmission channel. If it is disconnected, the entire network will be divided into two parts, and the large power grid will directly lose two generators and one load node, which will have a huge impact on the power grid.

[0179] The second critical line identified is line 14, which is an important generator sending line, and the connected generator is a generator with large active power output in the power grid. If it is disconnected, the power grid will lose a large load.

[0180] Although line 10 alone has little impact on the power grid, as shown in the above results, the critical vulnerable value of line 10 is large, and this line is easily affected by the disconnection of other lines and becomes a subsequent disconnected line, becoming an important link to promote the occurrence of cascading failure.

[0181] It can be seen from the system connection diagram that the lines 13, 23, 25 and 26 are located in an important topological connection area with dense line distribution. Since the lines are located in a dense area, disconnecting them will easily cause the influence on the power flow fluctuation of the surrounding lines and be easily affected by the disconnection of adjacent or next adjacent lines, and the propagation and vulnerability are high. It can be seen from the analysis that the power grid key line identification method has certain accuracy.

[0182] The network transmission efficiency comparison experiment and the load loss comparison experiment of the three methods are performed. According to the data, the network transmission efficiency caused by sequentially disconnecting the key lines identified by the application decreases the most and the fastest, and the load loss is the most, which proves that the lines identified by the application occupy a more critical position in the network, and the identification result of the application has certain reliability.

[0183] The above is only a preferred specific embodiment of the application, but the protection scope of the application is not limited to this. Any person skilled in the art can make equivalent replacement or change according to the technical solution and the inventive concept of the application within the technical range disclosed by the application, which should be covered in the protection scope of the application.

Claims

1. A method for identifying critical power grid lines based on topology and operational characteristics, characterized in that, include: S1. Network Modeling: Abstract a power grid into a topological network Z, with buses and transformers as nodes and transmission lines as edges. Calculate the degree, number of next-nearest neighbors, and number of nodes within two-step neighborhoods for each node in the topological network Z, and then calculate the overall degree of each node. Specifically, this includes the following: Considering the local topological connections of nodes, the power grid is abstracted into a topological graph with only points and edges connected. In this graph, edges represent transmission lines in the original power grid, and points represent buses, load nodes, generator nodes, and transformers. The degree D of node i is... i The calculation formula is: (1) In the formula, N is the number of nodes in the network; n ij This represents the connection relationship between node i and node j. If there is a connection between the two nodes, then n... ij The value is 1 if there is no connection, then n. ij =0; The importance of each node in the topology graph is measured by its comprehensiveness. The comprehensiveness C of node i is... Di for: (2) In the formula, C i Q is the number of second-neighbor nodes of node i; i Let i be the number of nodes in the two-step neighborhood of node i. S2. Failure Rate Statistics: A large number of cascading failure simulation experiments were conducted using the Monte Carlo method. The initial faulty line was randomly disconnected, and power flow calculations were performed on the power grid after the fault occurred. The probability of line outage was analyzed using the line operation reliability model to determine whether cascading failures would continue. One cascading failure simulation generated one accident chain, and the frequency of each line failure in the simulation experiment was statistically analyzed. S3. Initial Key Value Vector Construction: Based on the power flow calculation results, calculate the initial load rate of each line, and use the product of the initial load rate of each line and its historical failure rate as the initial key value. Then, construct the initial key value vector of the entire power grid lines using the initial key values ​​of all lines. This includes the following: Import the initial data of the power grid, including line parameters, node parameters, generator parameters, and load parameters, perform power flow calculations to obtain the initial load rate of each line in the power grid. Multiply the initial load rate of each line by the historical failure rate of the line as the initial key value required to improve the SALSA algorithm. Then, construct the initial key value vector of all lines in the power grid. The SALSA algorithm provides the initial key values ​​for each line. Then, the initial critical value of line k is calculated. : (6) In the formula, f k F represents the real-time power flow of line k; k This represents the power flow limit of line k during normal operation. The initial propagation value vectors of each line are obtained after calculation. Initial Vulnerability Value Vector for: In the formula, l represents the total number of lines; S4. Link Matrix Construction: Simulate N-1 faults, record the power fluctuations of each line as elements of the link matrix, construct the link matrix of the entire network, and process the obtained link matrix to obtain the propagation matrix and vulnerability matrix. S5. Iterative convergence calculation: Iteratively multiply the initial key value vector obtained in S3 with the propagation / vulnerability matrix obtained in S4 until the difference between adjacent iteration results is less than the set error, and obtain the final key propagation value vector and key vulnerability value vector. S6. Construction of critical value vector: The critical propagation value and critical vulnerability value of each line are weighted and summed to obtain the critical value of the line. The critical values ​​of all lines in the power grid are used to construct the critical value vector of the line. S7. Calculation of line comprehensive critical value: Considering the local topological importance of the line in the power grid, the square root of the product of the comprehensive degree of the nodes at both ends of the line is used as the correction coefficient to correct the line critical value. The line comprehensive critical value is obtained by multiplying the correction coefficient by the line critical value. S8. Critical Path Identification: Sort the comprehensive critical values ​​of each path obtained in S7 in descending order, and identify the path with the highest ranking as the critical path.

2. The method according to claim 1, characterized in that, S2 specifically includes the following: The Monte Carlo method was used to conduct cascading failure simulation experiments. The initial faulty line was randomly disconnected, and power flow calculations were performed on the power grid after the fault occurred to obtain the power on each line. Based on the power magnitude, it was determined whether any line was overloaded, and the risk of latent faults in other lines around the disconnected line was considered. Based on the presence or absence of line overload or latent faults, it was determined whether a next-order line disconnection would occur. If any line was disconnected, the cascading failure continued until the power grid regained power balance through load shedding or power balance was achieved within each isolated island after disconnection, and at this point, no line overload or latent faults occurred. This was then considered as one cascading failure simulation experiment. Multiple cascading failure simulation experiments were conducted, and the frequency of line disconnection in each line during the simulation experiments was statistically analyzed. The proportion of cascading failure chains involving a certain line to the total number of cascading failure chains was used as the historical failure rate of that line. The cascading failure simulation experiment specifically includes the following steps: 1) Import initial network data and randomly select the initial faulty line; 2) After disconnecting a line, update the network topology and determine whether the network has been split into isolated islands; 3) If no isolated island is formed, calculate the probability p of each line breaking, as follows: (3) In the formula, p0 is the probability of a latent fault occurring on the line, set to 0.001; f is the real-time power flow of the line. F represents the power flow limit during normal operation of the line; F max p1 represents the thermal stability limit of the line; p1 is the probability of the line disconnecting after exceeding the thermal stability limit, which is 1. Only lines that are disconnected from the previous stage are defined as having a probability of latent faults. If an island is formed, power balancing is performed within each island first, and then the probability of each line breaking is calculated according to equation (3). 4) After each power outage, perform power balancing within the power grid or island to determine if the power generated by the generators within the power grid or island is balanced with the load. If... If the total generator output in the power grid or island is less than the total load, then load shedding is required, and the load value of load node d is updated. for: (4) In the formula, d represents the load node; L represents the set of load nodes; P Ld Let g be the initial load of load node d; g be the generator node; G be the set of generator nodes; P ... generator node set. Gg The power output of generator node g; that is, the updated load of each node is the product of the original load and the power deficit ratio; 5) After balancing the power within the island or the entire power grid, determine whether the simulation will continue to generate cascading disconnected lines. If a new line disconnects due to overload, disconnect the overloaded line and return to step 3) until the cascading fault ends, and statistically analyze the accident chain and load loss situation of this cascading fault. 6) After simulating cascading failures, calculate the failure rate of each line in the power grid, and the failure rate V of line k. k for: (5) In the formula, W represents the total number of fault chains; W k This represents the number of fault chains containing line k.

3. The method according to claim 1, characterized in that, S4 specifically includes the following: For each line of the initial power grid, assume that it has an N-1 fault, obtain the power fluctuation on other lines in the power grid after the line is disconnected, and take the power fluctuation on other lines except the disconnected line as the corresponding row element in the link matrix of the line. Disconnect each line in the power grid in sequence to construct the line link matrix of the entire network; Then, the constructed line link matrix is ​​normalized in both rows and columns. The propagation matrix is ​​obtained by multiplying the row-normalized matrix with the transpose of the column-normalized matrix. The vulnerability matrix is ​​obtained by multiplying the transpose of the column-normalized matrix with the row-normalized matrix. Calculate the link matrix B of the SALSA algorithm, and its elements b ij The calculation formula is: (8) In the formula, f ij f represents the active power on line j after line i is disconnected; j0 Let J be the initial power on line j; S represents the total load of the power grid; the size of each element in the link matrix represents the impact of a disconnection of one line on the remaining lines. After obtaining the initial link matrix, the propagation matrix H and vulnerability matrix A are calculated from the link matrix to further evaluate the propagation criticality and vulnerability criticality of the route: (9) In the formula, B r B is the row-normalized matrix of the link matrix; c This is the column normalized matrix of the link matrix.

4. The method according to claim 1, characterized in that, S5 specifically includes the following: The initial critical value vector of the line is multiplied by the propagation matrix and the vulnerability matrix respectively to obtain the first-order critical propagation value vector and the first-order critical vulnerability value vector. Then, the first-order critical propagation value vector is multiplied by the propagation matrix to obtain the second-order critical propagation value vector, and the first-order critical vulnerability value vector is multiplied by the vulnerability matrix to obtain the second-order critical vulnerability value vector. This method is repeated iteratively until the convergence condition is met, that is, the difference between the critical propagation value vectors after two adjacent iterations is less than the set maximum error value, and the difference between the critical vulnerability value vectors after two adjacent iterations is less than the set specified error value. The final line critical propagation value vector and line critical vulnerability value vector are obtained. The specific calculation method is as follows: Calculate the propagation value vector π of the path using the defined propagation matrix and vulnerability matrix. h and the fragile value vector π a The calculation formula is as follows: (10) In the formula, This is the propagation value vector of the path during the m-th iteration, which is composed of the propagation values ​​of all paths. This is the vulnerability value vector of the line at the m-th iteration, which is composed of the vulnerability value vectors of all lines. Repeat equation (10) until the convergence condition is met. ,in The maximum error value is the π obtained after convergence. h and π a , which are the propagation value vector and vulnerability value vector of the power grid line, respectively.

5. The method according to claim 1, characterized in that, The key value calculation mentioned in S6 specifically refers to: Define a key value index π(k) for line k to comprehensively reflect the propagation capability and vulnerability of line deterioration faults: (11) In the formula, ω1 is the weight of line k as a propagation line; ω2 is the weight of line k as a vulnerable line.

6. The method according to claim 1, characterized in that, The method for calculating the comprehensive key value described in S7 is as follows: Comprehensive key value of line k based on topology and operational characteristics : (12) In the formula, C Dk (p) and C Dk (q) represents the degree of integration between nodes p and q at both ends of line k.

7. The method according to claim 1, characterized in that, S8 specifically includes the following: After obtaining the comprehensive critical value of all power grid lines, the obtained comprehensive critical values ​​are sorted from largest to smallest to obtain the order of the comprehensive critical values ​​of the lines. The higher the ranking of the lines, the higher their criticality in the power grid. The lines ranked at the top are identified as critical lines.

Images