CN116720093A - Method for dividing multiple constant value regions of protection device based on topological similarity analysis - Google Patents

Abstract

The application relates to the technical field of relay protection of power systems, in particular to a method for dividing a multi-constant-value region of a protection device based on topology similarity analysis, which comprises the steps of constructing a matrix representation of the topology of the power system based on the topology similarity analysis, and obtaining a plurality of singular values containing topology information through a singular value decomposition matrix; establishing a topological similarity index based on the singular value of a topological matrix of the power system, and dividing all topologies into a plurality of topological groups by adopting a K-Medoids clustering algorithm to obtain a central topology; and (3) establishing a double-target setting optimization model by taking the minimum total network protection action time and the minimum probability of losing selectivity as double targets, and optimally solving the topological setting values in each topological group by adopting a multi-target particle swarm optimization algorithm. According to the method, the relation between the singular value of the topology matrix and the topology similarity is integrated into the clustering model, and the obtained topology group is closer in setting, so that the requirements of different operation modes are better met.

Description

A method for dividing multiple setting areas of protection devices based on topological similarity analysis

Technical field

The present application relates to the technical field of power system relay protection, and in particular to a method of dividing multiple fixed value areas of a protection device based on topological similarity analysis.

Background technique

Existing power system relay protection protection devices generally adopt the setting mode of "offline setting, online unchanged". When performing offline setting calculations, different topologies of the system need to be considered to ensure that the setting values can still work normally under different topologies. Work. However, as the system scale expands, the number of topologies that need to be considered increases greatly. If all possible topologies are still considered at the same time, the performance of the protection settings will be greatly reduced.

Online tuning is a way to solve the above problems. The online tuning mode only needs to be tuned for the current topology without considering all possible topologies of the system. While reducing the complexity of the tuning calculation, it also puts forward high timeliness requirements. . However, in actual situations, the power system may change at any time. The online tuning system requires a certain amount of time from data acquisition, topology analysis to online tuning calculation, and this time increases with the expansion of the system scale, making it difficult to meet the needs of online tuning. High timeliness requirements. Therefore, a compromise setting scheme between online setting and offline setting - multi-setting area setting is very necessary to improve the setting and protection performance of the protection device.

In view of this, a compromise setting scheme between online setting and offline setting is needed - a method of dividing multiple setting areas of protection devices based on topological similarity analysis to improve the setting and protection performance of protection devices.

Contents of the invention

In view of the shortcomings in the existing technology that all possible topological structures of the system need to be considered during offline tuning calculations, resulting in complicated tuning calculations, and the timeliness requirements of the online tuning mode are too high, the present invention provides a protection based on topological similarity analysis. The method of dividing multiple setting areas of the device can compromise the advantages and disadvantages of online tuning and offline tuning without considering all possible topologies, and at the same time reduce the timeliness requirements. The specific technical solutions are as follows:

A method for dividing multiple setting areas of protection devices based on topological similarity analysis, including the following steps:

Step S1: Obtain the power system topology sample and construct an adjacency matrix representation;

Step S2: Obtain singular values containing topological information through singular value decomposition of the adjacency matrix;

Step S3: Establish a topological similarity index based on singular values containing topological information;

Step S4: According to the topology similarity index, use the K-Medoids clustering algorithm to divide all topologies into multiple topology groups and obtain the corresponding topology center points;

Step S5: Establish a dual-objective tuning optimization model with the goal of minimizing the entire network protection action time and minimizing the probability of losing selectivity; Step S6: Use the multi-objective particle swarm optimization algorithm to optimize the topology tuning values in each topology group.

Preferably, step S1 specifically includes the following steps:

S11. Obtain several topology samples of the power system under actual operating conditions and obtain actual operating topology samples;

S11. In the simulation environment, anticipate possible topology methods and obtain simulation topology samples;

S13. Use the actual running topology sample and the simulated expected topology sample as a topology sample set;

S14. Abstract each topology sample of the topology sample set described in step S13 into a topology structure graph consisting of nodes and branches, and represent the topology structure graph with an adjacency matrix.

Preferably, the method of expressing the topological structure graph with an adjacency matrix in step S14 is as follows:

For a new topology, let the topology diagram of the power system be G(V,E), where V(v ₁ ,v ₂ ,…,v _n ) represents the node set, E(e ₁ , e ₂ ,…, e _n ) represents the line set, then the elements in the adjacency matrix A of the topological structure graph G(V,E) are defined as

In the formula, a _ij is the element in the i-th row and j-th column of the adjacency matrix A, i=1,2,3,…,n; j=1,2,3,…,n, n is the number of nodes in the entire network, ( v _i , v _j ) represents the edge connecting node v _i and node v _j .

Preferably, the singular value decomposition adjacency matrix in step S2 is as follows:

Assuming the adjacency matrix A∈R _m,n , based on the singular value decomposition theory, the adjacency matrix A can be decomposed into the product of three matrices:

A=USV ^T (2)

In the formula, W = diag (λ ₁ ,...λ _r ), λ ₁ , λ ₂ ,...λ _r are the singular values of the adjacency matrix A and are arranged in descending order, r is the rank of the adjacency matrix A, and U is the left singular matrix, It consists of the eigenvectors of AA ^T , V ^T is the right singular matrix, and it consists of the eigenvectors of A ^T A;

For the obtained singular values, the smaller singular values are eliminated and only the larger m singular values are retained. The principle of retention is to retain the singular values from large to small until the sum of squares of the retained singular values occupies 95% of the total sum of squares of singular values. %above.

Preferably, the method for establishing the topological similarity index in step S3 is specifically:

For two different topologies, the similarity is determined by calculating the root mean square of the singular value sequence. Let the two singular value sequences be respectively and/> Each reservation retains the larger m singular values, then the root mean square of the two singular value sequences is

In the formula, eta is the root mean square of the two singular value sequences. The smaller the eta is, the more similar the two topologies are. is the i-th singular value in the singular value sequence ρ,/> is the i-th singular value in the singular value sequence ω, and θ _i is the i-th weight coefficient corresponding to the i-th singular value.

Preferably, the algorithm flow steps of the K-Medoids clustering algorithm described in step S4 include:

S41. Specify the number of topological groups k, and randomly select k singular value sequences as center points;

S42. Calculate the topological similarity index between the remaining singular value sequence and each center point, and then assign each topology sample to the topology group with the smallest similarity index;

S43. After dividing all remaining topological samples into k topological groups, calculate the cost function in each topological group. The formula is as shown in Equation (5):

in,

In the formula, t _p is the cost after this grouping, C _j is the topological group represented by the jth center point y _j , is the similarity index between the topological sample point x in C _j and the sample center point y _j of C _j ;

S44. In each divided topological group, select any topological sample point as the new center point and exchange it with the original center point, and calculate the exchange cost, as shown in Equation (6), if the exchange cost is less than 0, then the new center point The center point exchanges the original center point, otherwise the original center point is maintained;

In the formula, is the exchange cost between the topological sample point x and the j-th center point y _j , t _q is the cost calculated by the new center point, and t _p is the cost calculated by the original center point;

S45. Follow steps S42-S44 again to perform grouping of topology samples until the center point no longer changes or reaches a preset number of iterations, thereby dividing all topology samples into multiple topology groups and obtaining corresponding topology center points.

Preferably, the method for establishing the dual-objective tuning optimization model in step S5 is specifically:

A dual-objective function is formed by minimizing the entire network protection action time and minimizing the probability of losing selectivity, as shown below:

(1) Objective function one

In the formula, minOF ₁ is the objective function composed of the minimum protection action time of the entire network, n is the number of protection devices in the system, l is the topology group number, C _l is the lth topology group, j is the topology group C _l j topological modes, k is the number of fixed value areas and is equal to the number k of topological groups, is the protection device action time corresponding to the j-th topology mode of the i-th protection device;

(2) Objective function two

In the formula, minOF ₂ is the objective function with the lowest probability of losing selectivity, and p _j (l) is the probability of the protection device losing selectivity under the jth topology mode in the lth topology group.

Preferably, step S6 specifically includes the following steps:

S61. Initialize the position and speed of the particles, calculate the initial values of minOF ₁ and minOF ₂ , and initially set the optimal position x _{best of} individual particles and the best position g _best of the group;

S62. Initial screening of non-inferior solutions flj, find the Pareto optimal solution among particles and store it in the non-inferior solution set flj; S63. Update the position and speed of each particle;

S64. Update the values of minOF ₁ and minOF ₂ , and update the optimal position x _best of individual particles;

S65. Find the Pareto optimal solution in the particle and store it in the non-inferior solution set flj, and update and filter the non-inferior solution flj; S66. Use the roulette method to select g _best from flj;

S67. Repeat steps S63--S66 until the maximum number of iterations i _max is reached, exit the algorithm and output the Pareto optimal solution and Pareto frontier.

Compared with the prior art, this application has at least the following beneficial effects:

This application performs tuning based on obtaining topology samples under actual operating conditions, which is a characteristic of the online tuning mode. It has the disadvantage of avoiding the "offline tuning, online unchanged" tuning mode that needs to consider all possible topology methods; and in In the simulation environment, possible topological methods are anticipated, appropriate sample supplements are performed, and sample diversity is increased, which improves the ability of the fixed value to adapt to system changes, thereby reducing the timeliness requirements of the online tuning mode and improving the timeliness of the online tuning mode. The shortcoming of high requirements; with the dual goals of minimizing the entire network protection action time and the lowest probability of losing selectivity, a dual-objective tuning optimization model is established, and a multi-objective particle swarm optimization algorithm is used to optimize the topology tuning settings within each topology group. Obtain multiple fixed value areas that adapt to multiple topological groups, thereby adapting to different topologies. Among them, this application integrates the relationship between the singular value of the topological matrix and the topological similarity into the clustering model, and the resulting topological group is closer in tuning, thereby better meeting the requirements of different operating modes.

Description of the drawings

In order to explain the specific embodiments of the present invention more clearly, the drawings needed to describe the specific embodiments will be briefly introduced below.

Figure 1 is a schematic flow chart of a method for dividing multiple fixed value areas of a protection device based on topological similarity analysis according to an embodiment of the present application;

Figure 2 is a schematic diagram of the power grid topology shown in an embodiment of the present application;

Figure 3 is a flow chart of the multi-objective particle swarm optimization algorithm shown in the embodiment of the present application.

Figure 4 is a schematic diagram of target power flow data samples after visual preprocessing shown in the embodiment of the present application;

Detailed ways

The technical solutions in the embodiments of the present application will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, rather than all of the embodiments. Based on the embodiments in this application, all other embodiments obtained by those of ordinary skill in the art without creative efforts fall within the scope of protection of this application.

Please refer to FIG. 1 , which is a schematic flow chart of a method for dividing multiple fixed value areas of a protection device based on topological similarity analysis provided by an embodiment of the present application. A method of dividing multiple fixed value areas of a protection device based on topological similarity analysis in the embodiment of the present application can be applied to computer equipment, including but not limited to smart phones, notebook computers, tablet computers, desktop computers, physical Equipment such as servers and cloud servers. As shown in Figure 1, the method for dividing multiple fixed value areas of a protection device based on topological similarity analysis in this embodiment includes steps S1 to S6, which are detailed as follows:

Step S1: Obtain the power system topology sample and construct an adjacency matrix representation.

In some embodiments, step S1 includes:

S11. Obtain several topology samples of the power system under actual operating conditions and obtain actual operating topology samples;

S11. In the simulation environment, anticipate possible topology methods and obtain simulation topology samples;

S13. Use the actual running topology sample and the simulated expected topology sample as a topology sample set; S14. Abstract each topology sample in the topology sample set described in step S13 into a topology structure diagram composed of nodes and branches. Represent the topological structure graph as an adjacency matrix.

In this embodiment, taking Figure 2 as an example, the topological structure diagram is represented by an adjacency matrix, specifically as follows:

First, for a new topology, let the topology diagram of the power system be G(V,E), where V(v ₁ , v ₂ ,..., v _n ) represents the node set, and E(e ₁ , e ₂ , ...,e _n ) represents the line set, then the elements in the adjacency matrix A of the topological structure graph G(V,E) are defined as

where a _ij is the element in the i-th row and j-th column of the adjacency matrix A (i=1,2,3,…,n; j=1,2,3,…,n), n is the number of nodes in the entire network, (v _i , v _j ) represents the edge connecting node v _i and node v _j , and E is the line set.

Then the adjacency matrix of the 9-node system shown in Figure 2 is expressed as

Step S2: Obtain singular values containing topological information through singular value decomposition of the adjacency matrix.

In this embodiment, still taking Figure 2 as an example, several singular values containing topological information can be obtained by singular value decomposition of the above adjacency matrix. The calculation of the singular values is as follows:

Assuming the adjacency matrix A∈R _m,n , based on the singular value decomposition theory, A can be decomposed into the product of three matrices:

A=USV ^T (2)

In the formula: W = diag (λ ₁ ,...λ _r ), λ ₁ , λ ₂ ,...λ _r are the singular values of the adjacency matrix A and are arranged in descending order, and r is the rank of the adjacency matrix A. U is the left singular matrix, consisting of the eigenvectors of AA ^T , and V ^T is the right singular matrix, consisting of the eigenvectors of A ^T A.

Then for the 9-node system shown in Figure 2, the singular values obtained by the singular value decomposition of the adjacency matrix are 2.2361, 2.2361, 1.4142, 1.4142, 1.4142, 1.4142, 8.8827×10 ^-17 , 4.3312×10 ^-17 , 9.2997×10 ^{- 18} .

According to the singular value decomposition theory, the larger the singular value λ _i , the more important the characteristic information it contains. Therefore, in practical applications, only the first m singular values are retained. Eliminating smaller singular values will not only increase the error, but also Less computational effort. In order to ensure that the first m singular values can completely represent all the information of the adjacency matrix, the principle of m retention is to retain singular values in order from large to small until the sum of squares of the retained singular values accounts for more than 95% of the total sum of squares of singular values.

Step S3: Establish a topological similarity index based on singular values containing topological information.

In some embodiments, step S3 includes: for two different topologies, determining the similarity by calculating the root mean square of the singular value sequence.

In this embodiment, the similarity determination is performed by calculating the root mean square of the singular value sequence. The details are as follows: for two different topologies, the similarity determination is performed by calculating the root mean square of the singular value sequence. Let the two singular value sequences be and/> Each reservation retains the larger m singular values, then the root mean square of the two singular value sequences is

In the formula, eta is the root mean square of the two singular value sequences. The smaller the eta is, the more similar the two topologies are. is the i-th singular value in the singular value sequence ρ,/> is the i-th singular value in the singular value sequence ω, and θ _i is the i-th weight coefficient corresponding to the singular value. According to the singular value decomposition theory, the larger the singular value is, the more important the feature information it contains is, and it can better reflect the topological characteristics, so the corresponding weight coefficient θ _i is also larger.

Step S4, according to the topological similarity index, use the K-Medoids clustering algorithm to divide all topologies into multiple topological groups and obtain the corresponding topological center points;

The K-Medoids clustering algorithm, the algorithm process includes:

S41. Specify the number of topological groups k, and randomly select k singular value sequences as center points;

S42. Calculate the topological similarity index between the remaining singular value sequence and each center point, and then assign each topological sample point to the topological group with the smallest similarity index;

S43. After dividing all remaining topological sample points into k topological groups, calculate the cost function within each topological group. Formula (5) is:

in,

In the formula, t _p is the cost after this grouping, C _j is the topological group represented by the jth center point y _j , is the similarity index between the topological sample point x in C _j and the sample center point y _j of C _j ;

S44. In each divided topological group, select any topological sample point as the new center point and exchange it with the original center point, and calculate the exchange cost, as shown in Equation (6), if the exchange cost is less than 0, then the new center point The center point exchanges the original center point, otherwise the original center point is maintained;

In the formula, is the exchange cost between the topological sample point x and the j-th center point y _j , t _q is the cost calculated by the new center point, and t _p is the cost calculated by the original center point;

S45. Follow steps S42-S44 again to perform grouping of topology samples until the center point no longer changes or reaches a preset number of iterations, thereby dividing all topology samples into multiple topology groups and obtaining corresponding topology center points.

Step S5: Establish a dual-objective tuning optimization model with the goal of minimizing the entire network protection action time and minimizing the probability of losing selectivity.

In some embodiments, step S5 includes: taking the minimum protection action time of the entire network and the minimum probability of losing selectivity as dual objectives, forming a dual-objective function to satisfy sensitivity as a constraint.

In this embodiment, a dual-objective function is formed by minimizing the entire network protection action time and minimizing the probability of losing selectivity, as shown below:

(1) Objective function one

Among them, minOF ₁ is the objective function composed of the minimum protection action time of the entire network, n is the number of protection devices in the system, l is the topology group number, C _l is the lth topology group, j is the jth in topology group C _l A topological method, k is the number of fixed value areas and is equal to the number k of topological groups, is the action time of the protection device corresponding to the j-th topology mode of the i-th protection device in the l-th topology group.

(2) Objective function two

Among them, minOF ₁ is the objective function composed of the lowest probability of losing selectivity, and P _j (l) is the probability of the protection device losing selectivity under the jth topology mode in the lth topology group.

Step S6: Use the multi-objective particle swarm optimization algorithm to optimize the topology setting values in each topology group.

As shown in Figure 3, it is a flow chart of the multi-objective particle swarm optimization algorithm. In some embodiments, the multi-objective particle swarm optimization algorithm is used to separately solve all topology tuning optimization solutions in each topology group to obtain a solution that satisfies all topologies in the topology group. a set of fixed values.

The step S6 includes:

S61. Initialize the position and speed of the particles, calculate the initial values of minOF ₁ and minOF ₂ , and initially set the optimal position x _best of individual particles and the best position g _best of the group;

S62. Initial screening of non-inferior solutions flj, find the Pareto optimal solution among particles and store it in the non-inferior solution set flj; S63. Update the position and speed of each particle;

S64. Update the values of minOF ₁ and minOF ₂ , and update the optimal position x _best of individual particles;

S65. Find the Pareto optimal solution in the particle and store it into the non-inferior solution set flj, and update and screen the non-inferior solution flj; S66. Use the roulette method to select g _best from flj;

S67. Repeat steps S63--S66 until the maximum number of iterations i _max is reached, exit the algorithm and output the Pareto optimal solution and Pareto frontier.

As an example but not a limitation, take the IEEE 39-node power system as an example. The topology diagram of the power system is shown in Figure 4 . The offline simulation uses the N-3 principle to generate samples, that is, considering the operating modes after the outage of 3 lines (in practical applications, the operating modes that have occurred in the past period of time and the expected possible operating modes of the power grid are considered), and 100 samples are set to be generated. Data, as shown in Table 1.

Table 1 Operating mode number

After dividing the 100 topology methods into C ₁ , C ₂ ···C ₇ topology groups, each topology group contains several topologies as shown in Table 2, and then use the MOPSO algorithm to optimize and solve each topology group. Each topology group obtains a set of protection settings and corresponding action time. The sum is the protection action time of the entire network as shown in the table below. The proportion of the number of protections that have lost selectivity in each topology mode to the total number of protections is: The probability of losing selectivity. The probability of losing selectivity shown in the table below is the maximum value of the probability of losing selectivity among several topological methods of a topological group. (Note: The specific data calculation process is calculated by the now mature MOPSO algorithm program.)

Table 2 Operation mode clustering results

After processing by the above proposed method, the comparison results with traditional offline tuning are shown in Table 3. It can be seen that the protection setting performance after multi-setting area processing has been improved and can better adapt to the operating mode.

Table 3 Comparison results with traditional offline tuning

Finally, it should be noted that the above embodiments are only used to illustrate the technical solution of the present invention, but not to limit it. Although the present invention has been described in detail with reference to the foregoing embodiments, those of ordinary skill in the art should understand that: The technical solutions described in the foregoing embodiments can still be modified, or some or all of the technical features can be equivalently replaced; and these modifications or substitutions do not deviate from the essence of the corresponding technical solutions from the technical solutions of the embodiments of the present invention. scope, they should be covered by the claims and the scope of the description of the present invention.

Claims (8)

1. A method for dividing multiple fixed-value areas of protection devices based on topological similarity analysis, which is characterized by including the following steps: Step S1: Obtain the power system topology sample and construct an adjacency matrix representation; Step S2: Obtain singular values containing topological information through singular value decomposition of the adjacency matrix; Step S3: Establish a topological similarity index based on singular values containing topological information; Step S4: According to the topology similarity index, use the K-Medoids clustering algorithm to divide all topologies into multiple topology groups and obtain the corresponding topology center points; Step S5: Establish a dual-objective tuning optimization model with the goal of minimizing the entire network protection action time and minimizing the probability of losing selectivity; Step S6: Use the multi-objective particle swarm optimization algorithm to optimize the topology setting values in each topology group. 2. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that step S1 specifically includes the following steps: S11. Obtain several topology samples of the power system under actual operating conditions and obtain actual operating topology samples; S11. In the simulation environment, anticipate possible topology methods and obtain simulation topology samples; S13. Use the actual running topology sample and the simulated expected topology sample as a topology sample set; S14. Abstract each topology sample of the topology sample set described in step S13 into a topology structure graph consisting of nodes and branches, and represent the topology structure graph with an adjacency matrix. 3. The construction method of adjacency matrix representation according to claim 2, characterized in that step S14 is specifically: For a new topology, let the topology diagram of the power system be G(V,E), where V(v ₁ ,v ₂ ,…,v _n ) represents the node set, E(e ₁ , e ₂ ,…, e _n ) represents the line set, then the elements in the adjacency matrix A of the topological structure graph G(V,E) are defined as In the formula, a _ij is the element in the i-th row and j-th column of the adjacency matrix A, i=1,2,3,…,n; j=1,2,3,…,n, n is the number of nodes in the entire network, ( v _i , v _j ) represents the edge connecting node v _i and node v _j . 4. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that step S2 is specifically: Assuming the adjacency matrix A∈R _m,n , based on the singular value decomposition theory, the adjacency matrix A can be decomposed into the product of three matrices: A=USV ^T (2) In the formula: W = diag (λ ₁ ,...λ _r ), λ ₁ , λ ₂ ,...λ _r are the singular values of the adjacency matrix A and are arranged in descending order, r is the rank of the adjacency matrix A; U is the left singular matrix, It consists of the eigenvectors of AA ^T , V ^T is the right singular matrix, and it consists of the eigenvectors of A ^T A; For the obtained singular values, the smaller singular values are eliminated and only the larger m singular values are retained. The principle of retention is to retain the singular values from large to small until the sum of squares of the retained singular values occupies 95% of the total sum of squares of singular values. %above. 5. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that step S3 is specifically: For two different topologies, the similarity is determined by calculating the root mean square of the singular value sequence. Let the two singular value sequences be respectively and/> Each reservation retains the larger m singular values, then the root mean square of the two singular value sequences is In the formula, eta is the root mean square of the two singular value sequences. The smaller the eta is, the more similar the two topologies are. is the i-th singular value in the singular value sequence ρ,/> is the i-th singular value in the singular value sequence ω, and θ _i is the i-th singular value corresponding to the i-th weight coefficient. 6. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that the K-Medoids clustering algorithm in step S4 includes: S41. Specify the number of topological groups k, and randomly select k singular value sequences as center points; S42. Calculate the topological similarity index between the remaining singular value sequence and each center point, and then assign each topology sample to the topology group with the smallest similarity index; S43. After dividing all remaining topological samples into k topological groups, calculate the cost function within each topological group. The formula is as shown in Equation (5): in, In the formula, t _p is the cost after this grouping, C _j is the topological group represented by the jth center point y _j , is the similarity index between the topological sample point x in C _j and the sample center point y _j of C _j ; S44. In each divided topological group, select any topological sample point as the new center point and exchange it with the original center point, and calculate the exchange cost, as shown in Equation (6), if the exchange cost is less than 0, then the new center point The center point exchanges the original center point, otherwise the original center point is maintained; In the formula, is the exchange cost between the topological sample point x and the j-th center point y _j , t _q is the cost calculated by the new center point, and t _p is the cost calculated by the original center point; S45. Follow steps S42-S44 again to perform grouping of topology samples until the center point no longer changes or reaches a preset number of iterations, thereby dividing all topology samples into multiple topology groups and obtaining corresponding topology center points. 7. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that step S5 is specifically: A dual-objective function is formed by minimizing the entire network protection action time and minimizing the probability of losing selectivity, as shown below: (1) Objective function one In the formula, minOF ₁ is the objective function composed of the minimum protection action time of the entire network, n is the number of protection devices in the system, l is the topology group number, C _l is the lth topology group, j is the topology group C _l j topological modes, k is the number of fixed value areas and is equal to the number k of topological groups, t _i ^j (l) is the action of the protection device corresponding to the jth topological mode of the i-th protection device in the l-th topological group time; (2) Objective function two In the formula, minOF ₂ is the objective function with the lowest probability of losing selectivity, and p _j (l) is the probability of the protection device losing selectivity under the jth topology mode in the lth topology group. 8. The method for dividing multiple fixed value areas of a protection device according to claim 1, characterized in that step S6 specifically includes the following steps: S61. Initialize the position and speed of the particles, calculate the initial values of minOF ₁ and minOF ₂ , and initially set the optimal position x _best of individual particles and the best position g _best of the group; S62. Initial screening of non-inferior solutions flj, finding the Pareto optimal solution in the particles and storing it in the non-inferior solution set flj; S63. Update the position and speed of each particle; S64. Update the values of minOF ₁ and minOF ₂ , and update the optimal position x _best of individual particles; S65. Find the Pareto optimal solution in the particle and store it into the non-inferior solution set flj, and update and filter the non-inferior solution flj; S66. Use the roulette method to select g _best from flj; S67. Repeat steps S63--S66 until the maximum number of iterations i _max is reached, exit the algorithm and output the Pareto optimal solution and Pareto frontier.