A method for establishing a power grid framework topology change identification model suitable for multiple operation scenarios

Abstract

The application discloses a power grid network topology change identification model establishing method suitable for multiple operation scenes and relates to the field of power grid networking. The method solves the problem that the existing power grid network topology change identification model lacks regional pertinence and comprises the following steps: S1, performing static topology index analysis on a power region, obtaining identification model static analysis coefficients corresponding to different power regions according to an analysis result, and obtaining identification model static analysis data; S2, performing dynamic index analysis on the power region according to the identification model static analysis data, obtaining identification model dynamic analysis coefficients corresponding to different power regions according to an analysis result, and obtaining identification model dynamic analysis data; and S3, matching a model creation mode of the power region according to the identification model static analysis data and the identification model dynamic analysis data, and creating a regional topology change identification model. The application can improve the pertinence of the power grid network topology change identification model establishing method.

Description

Technical Field

[0001] This invention belongs to the field of power grid networking and relates to topology identification technology, specifically a method for establishing a power grid topology change identification model applicable to multiple operating scenarios. Background Technology

[0002] Existing methods for establishing power grid topology change identification models have the following drawbacks when creating topology change identification models for power areas:

[0003] 1. Existing methods for establishing power grid topology change identification models typically rely on manual analysis of topology data from historical periods to select the model creation method. They do not take into account the voltage levels, power node distribution, and topology complexity within the power area to conduct a comprehensive static analysis of the power area. Consequently, they cannot obtain static analysis coefficients of the identification model based on the analysis results to match the creation method of the power area, resulting in a lack of static specificity in the model creation method.

[0004] 2. Existing methods for establishing power grid topology change identification models do not perform continuous time-period topology change frequency index analysis on the power area. It is difficult to obtain dynamic analysis coefficients of the identification model based on the analysis results to match the creation method of the power area. As a result, the model creation method lacks dynamic specificity, leading to poor model application effect.

[0005] To address this, we propose a method for establishing a power grid topology change identification model applicable to multiple operating scenarios. Summary of the Invention

[0006] To address the shortcomings of existing technologies, the purpose of this invention is to provide a method for establishing a power grid topology change identification model applicable to multiple operating scenarios. This invention aims to improve the relevance of the power grid topology change identification model establishment method.

[0007] To achieve the above objectives, the present invention adopts the following technical solution: a method for establishing a power grid topology change identification model applicable to multiple operating scenarios, comprising the following steps:

[0008] Step S1: Perform static topology index analysis on the power area, obtain the static analysis coefficients of the identification model corresponding to different power areas based on the analysis results, and obtain the static analysis data of the identification model;

[0009] Step S2: Perform dynamic index analysis on the power area based on the static analysis data of the identification model, obtain the dynamic analysis coefficients of the identification model corresponding to different power areas based on the analysis results, and obtain the dynamic analysis data of the identification model.

[0010] Step S3: Match the model creation method of the power area based on the static analysis data and dynamic analysis data of the identification model, and create an area topology change identification model based on the matching results.

[0011] Furthermore, step S1 also includes the following steps:

[0012] Step S11: Obtain the power regions that need to be identified for power grid topology changes, obtain multiple power regions, and arbitrarily select one sample power region from the multiple power regions obtained;

[0013] Step S12: Perform structural complexity analysis on the sample power region, and obtain the static analysis coefficients of the identification model corresponding to the sample power region based on the analysis results;

[0014] Step S13: Obtain the static analysis coefficients of the identification model corresponding to each power area to obtain the static analysis data of the identification model.

[0015] Furthermore, step S12 also includes the following steps:

[0016] Step S121: Obtain the power line topology map corresponding to the sample power area to obtain the sample power topology map; obtain the regional map corresponding to the sample power area to obtain the sample regional map; fuse the sample power topology map and the sample regional map using indicators to obtain the regional topology fusion image.

[0017] Step S122: Obtain the topological layers corresponding to the sample power topology map to obtain multiple power topology levels. Obtain the geographical region corresponding to each power topology level in the regional topology fusion image to obtain the T1 topology sub-region to the Ta topology sub-region.

[0018] Step S123: Perform topological complexity analysis on the T1 topological sub-region, and obtain the structural complexity of the T1 region based on the analysis results;

[0019] Step S124: Obtain the regional structural complexity corresponding to the topological sub-regions T2 to Ta respectively, and obtain the regional structural complexity from T2 to Ta.

[0020] Furthermore, step S12 also includes the following steps:

[0021] Step S125: Obtain the area values ​​of topological sub-regions T1 to Ta in the regional topology fusion image, respectively, to obtain the area values ​​of region T1 to region Ta. Obtain the area values ​​of sample power regions in the regional topology fusion image, to obtain the area values ​​of sample regions. Calculate the ratio of each region area value to the sample region area value to obtain the area weights of region T1 to Ta.

[0022] Step S126: Obtain the upper limit of the layer voltage corresponding to the topological sub-regions from T1 to Ta respectively, and obtain the upper limit of the layer voltage from T1 to Ta. Sum the upper limit of the layer voltage from T1 to Ta to obtain the cumulative value of the layer voltage in the region. Calculate the ratio of the upper limit of the layer voltage of each layer to the cumulative value of the layer voltage in the region to obtain the voltage weight from T1 to Ta.

[0023] Step S127: Calculate the static analysis coefficients of the identification model corresponding to the sample power region by transforming the structural complexity of region T1 to the structural complexity of region Ta, the area weight of T1 to the area weight of Ta, and the voltage weight of T1 to the voltage weight of Ta.

[0024] Furthermore, step S123 also includes the following steps:

[0025] The power nodes existing in the T1 topological sub-region are obtained, from power node P1 to power node Pb.

[0026] The number of power line connections from power node P1 to power node Pb is obtained respectively, thus obtaining the number of power connection edges from P1 to Pb.

[0027] The structural complexity of region T1 is obtained by calculating the number of electrical connection edges from P1 to Pb.

[0028] The structural complexity of region T1 is calculated using the following formula:

[0029] ;

[0030] Where Hf1 is the structural complexity of region T1, Lbi is the number of power connection edges in Pi, and b is the number of power nodes.

[0031] Furthermore, step S2 also includes the following steps:

[0032] Step S21: Obtain static analysis data of the identification model, and obtain the sample power area based on the static analysis data of the identification model;

[0033] Step S22: Perform dynamic topology change analysis on the sample power area, and obtain the dynamic analysis coefficients of the identification model corresponding to the sample power area based on the analysis results;

[0034] Step S23: Obtain the dynamic analysis coefficients of the identification model corresponding to each power area to obtain the dynamic analysis data of the identification model.

[0035] Furthermore, step S22 also includes the following steps:

[0036] Step S221: Select several dynamic index analysis periods of equal duration from the historical working periods corresponding to the sample power area;

[0037] Step S222: Randomly select a sample dynamic time period from the multiple dynamic index analysis time periods, obtain the number of topology changes of the sample power area during the sample dynamic time period, obtain the time length corresponding to the sample dynamic time period, obtain the sample duration, calculate the ratio of the number of topology changes to the sample duration, and obtain the time period topology change frequency corresponding to the sample dynamic time period.

[0038] Step S223: Obtain the time period topology change frequency corresponding to each dynamic indicator analysis period.

[0039] Furthermore, step S22 also includes the following steps:

[0040] Step S224: Set any two consecutive time periods of dynamic indicator analysis as a dynamic indicator combination period to obtain multiple dynamic indicator combination periods. Randomly select a sample indicator combination segment from the multiple dynamic indicator combination periods.

[0041] Step S225: Perform time-period topology change frequency change analysis on the sample index combination time period, and obtain the change frequency continuity deviation based on the analysis results;

[0042] Step S226: Obtain the continuous deviation of the change frequency corresponding to each dynamic index combination time period, and calculate the average of the obtained continuous deviations of the change frequency to obtain the dynamic analysis coefficient of the identification model corresponding to the sample power area.

[0043] Furthermore, step S225 also includes the following steps:

[0044] The two dynamic indicator analysis periods contained in the sample indicator combination segment are respectively set as the first dynamic analysis period and the second dynamic indicator analysis period;

[0045] The frequency of topology changes for the first dynamic analysis period and the second dynamic index analysis period is obtained. The frequency of topology changes for the first period and the frequency of topology changes for the second period are obtained. The difference between the frequency of topology changes for the first period and the frequency of topology changes for the second period is calculated. The absolute value of the difference is taken, and the ratio of the absolute value of the difference to the frequency of topology changes for the first period is calculated to obtain the continuity deviation of the frequency of change corresponding to the sample index combination segment.

[0046] Furthermore, step S3 also includes the following steps:

[0047] Obtain static analysis data of the identification model, and obtain the static analysis coefficients of the identification model corresponding to each power area based on the static analysis data of the identification model; obtain dynamic analysis data of the identification model, and obtain the dynamic analysis coefficients of the identification model corresponding to each power area based on the dynamic analysis data of the identification model.

[0048] Obtain the static matching interval and the dynamic matching interval of the recognition model respectively;

[0049] If the static analysis coefficients of the identification model are in the static matching interval of the identification model and the dynamic analysis coefficients of the identification model are in the dynamic matching interval of the identification model, then the graph theory method is used to create a regional topology change identification model for the power area.

[0050] If the static analysis coefficients of the identification model are not within the static matching range of the identification model or the dynamic analysis coefficients of the identification model are not within the dynamic matching range of the identification model, then machine learning methods are used to create a regional topology change identification model for the power area.

[0051] In summary, due to the adoption of the above technical solution, the beneficial effects of the present invention are:

[0052] 1. This invention performs a comprehensive static analysis of a power region by combining the voltage level, power node distribution, and topological complexity within the power region. Based on the analysis results, it obtains the static analysis coefficients of the identification model to match the creation method of the power region, which can improve the lack of static specificity in the model creation method.

[0053] 2. This invention analyzes the frequency index of continuous time-period topology changes in power areas, and obtains dynamic analysis coefficients of the identification model based on the analysis results to match the creation method of the power area. This can improve the lack of dynamic targeting of the model creation method and the model application effect. Attached Figure Description

[0054] To facilitate understanding by those skilled in the art, the present invention will be further described below with reference to the accompanying drawings.

[0055] Figure 1 This is a diagram illustrating the implementation steps of the present invention;

[0056] Figure 2 This is a schematic diagram of the power node of the present invention. Detailed Implementation

[0057] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative effort are within the scope of protection of the present invention.

[0058] Example 1

[0059] Please see Figure 1 The power grid topology change identification model in this invention belongs to the category of artificial intelligence models. This invention provides a technical solution: a method for establishing a power grid topology change identification model applicable to multiple operating scenarios, comprising the following steps:

[0060] Step S1: Perform static topological index analysis on the target topology identification region, and obtain the static analysis coefficients of the identification model based on the analysis results;

[0061] Step S1 further includes the following steps:

[0062] The power regions that need to be identified for power grid topology changes are obtained, resulting in multiple power regions. Then, one sample power region is randomly selected from the multiple obtained power regions.

[0063] Structural complexity analysis is performed on the sample power region, and static analysis coefficients of the identification model corresponding to the sample power region are obtained based on the analysis results.

[0064] Specifically as follows:

[0065] The power line topology map corresponding to the sample power area is obtained to obtain the sample power topology map. The regional map corresponding to the sample power area is obtained to obtain the sample regional map. The topology indicators in the sample power topology map are marked in the sample regional map to obtain the regional topology fusion image.

[0066] The topological layers corresponding to the sample power topology map are obtained to obtain multiple power topology levels. The geographical regions corresponding to each power topology level in the regional topology fusion image are obtained to obtain multiple topological sub-regions. The obtained multiple topological sub-regions are named T1 topological sub-region to Ta topological sub-region respectively.

[0067] It should be noted here that:

[0068] In this application, the voltage stratification involved here is specifically based on the voltage level corresponding to the topology distribution network. The power topology stratification involved here includes high voltage layer (such as 500kV), medium voltage layer (such as 220kV / 110kV) and low voltage layer (such as 35kV / 10kV).

[0069] In this application, the hierarchical voltages corresponding to the T1 topological subregion to the Ta topological subregion decrease sequentially.

[0070] In this application, 1, 2, 3...a in the topological sub-regions T1 to Ta are the numbers corresponding to the topological sub-regions, and a is an integer greater than 0;

[0071] Perform topological complexity analysis on the T1 topological sub-region, and obtain the structural complexity of the T1 region based on the analysis results;

[0072] Specifically as follows:

[0073] The power nodes existing in the T1 topological sub-region are obtained, and the obtained power nodes are set as power nodes P1 to Pb respectively.

[0074] It should be noted here that:

[0075] In this application, the power nodes involved include, but are not limited to, substations, power plants, and power terminals;

[0076] In this application, P is the symbol corresponding to the power node, b is the quantity value corresponding to the power node, and b is an integer greater than 0.

[0077] Please see Figure 2 The number of power line connections from power node P1 to power node Pb is obtained respectively, thus obtaining the number of power connection edges from P1 to Pb.

[0078] The structural complexity of region T1 is obtained by calculating the number of electrical connection edges from P1 to Pb.

[0079] The structural complexity of region T1 is calculated using the following formula:

[0080] ;

[0081] Where Hf1 is the structural complexity of region T1, Lbi is the number of power connection edges in Pi, and b is the number of power nodes.

[0082] Repeat the process of obtaining the structural complexity of region T1, and obtain the structural complexity of the regions corresponding to the topological sub-regions T2 to Ta respectively, to obtain the structural complexity of region T2 to Ta.

[0083] The area values ​​of topological sub-regions T1 to Ta in the regional topology fusion image are obtained respectively, and the area values ​​of region T1 to region Ta are obtained. The area value of sample power region in the regional topology fusion image is obtained, and the area value of sample region is obtained. The ratio of area value of region T1 to area value of sample region is calculated to obtain the area weight of T1. The ratio of area value of region T2 to area value of sample region is calculated to obtain the area weight of T2. And so on, the ratio of area value of region Ta to area value of sample region is calculated to obtain the area weight of Ta.

[0084] The upper limits of the layer voltages corresponding to the topological sub-regions from T1 to Ta are obtained respectively. The upper limits of the layer voltages from T1 to Ta are then summed to obtain the cumulative value of the layer voltages in the region. The ratio of the upper limit of the layer voltage of T1 to the cumulative value of the layer voltage in the region is calculated to obtain the voltage weight of T1. The ratio of the upper limit of the layer voltage of T2 to the cumulative value of the layer voltage in the region is calculated to obtain the voltage weight of T2. And so on, the ratio of the upper limit of the layer voltage of Ta to the cumulative value of the layer voltage in the region is calculated to obtain the voltage weight of Ta.

[0085] The static analysis coefficients of the identification model corresponding to the sample power region are obtained by calculating the structural complexity of region T1 to region Ta, the area weight of T1 to area weight of Ta, and the voltage weight of T1 to voltage weight of Ta.

[0086] The static analysis coefficients of the identification model corresponding to the sample power area are calculated using the following formula:

[0087] ;

[0088] Where Jxs is the static analysis coefficient of the identification model corresponding to the sample power region, Fzdi is the structural complexity of the Ti region, Mqzi is the area weight of Ti, Vqzi is the voltage weight of Ti, and a is the quantity value corresponding to the topological sub-region.

[0089] It should be noted here that:

[0090] In this application, the structural complexity of the Ti region can be any one of the structural complexities of the T1 region to the structural complexity of the Ta region, the area weight of Ti can be any one of the area weights of the T1 region to the area weights of the Ta region, and the voltage weight of Ti can be any one of the voltage weights of the T1 region to the voltage weights of the Ta region.

[0091] Repeat the static analysis coefficients of the identification model corresponding to the sample power area, and obtain the static analysis coefficients of the identification model corresponding to each power area to obtain the static analysis data of the identification model;

[0092] Step S1 above combines the voltage level, power node distribution, and topological complexity within the power area to conduct a comprehensive static analysis of the power area. Based on the analysis results, the static analysis coefficients of the identification model are obtained to match the creation method of the power area. This systematically integrates the multi-dimensional characteristics of the area (such as the stability requirements of voltage levels, the density of node connections, and the potential for dynamic changes in the topology), thereby providing an objective basis for the selection of subsequent topology change identification models. This ensures that the model creation method is highly adapted to the actual characteristics of the area, avoiding resource waste caused by excessive model complexity and preventing the omission of key topological features due to model simplification, ultimately improving the accuracy of change detection and the efficiency of system response.

[0093] Step S2: Perform dynamic index analysis on the power area based on the static analysis data of the identification model, and obtain the dynamic analysis data of the identification model based on the analysis results;

[0094] Obtain static analysis data of the identification model, and obtain the sample power region based on the static analysis data of the identification model;

[0095] Dynamic topology change analysis is performed on the sample power area, and the dynamic analysis coefficients of the identification model corresponding to the sample power area are obtained based on the analysis results.

[0096] Specifically as follows:

[0097] Several dynamic indicator analysis periods of equal duration were selected from the historical working periods corresponding to the sample power area;

[0098] In the multiple dynamic index analysis periods obtained, an arbitrary sample dynamic period is selected. The number of topology changes in the sample power area during the sample dynamic period is obtained, and the time length corresponding to the sample dynamic period is obtained to obtain the sample duration. The ratio of the number of topology changes to the sample duration is calculated to obtain the topology change frequency of the period corresponding to the sample dynamic period.

[0099] Repeat the time period topology change frequency corresponding to the dynamic time period of the sample, and obtain the time period topology change frequency corresponding to each dynamic indicator analysis time period;

[0100] Set any two consecutive time periods of dynamic indicator analysis as a dynamic indicator combination period to obtain multiple dynamic indicator combination periods. Randomly select a sample indicator combination segment from the multiple dynamic indicator combination periods.

[0101] The two dynamic indicator analysis periods contained in the sample indicator combination segment are respectively set as the first dynamic analysis period and the second dynamic indicator analysis period;

[0102] It should be noted here that:

[0103] In this application, the time range corresponding to the first dynamic analysis period is earlier than the time range corresponding to the second dynamic indicator analysis period.

[0104] The topology change frequency corresponding to the first dynamic analysis period and the second dynamic index analysis period is obtained to obtain the topology change frequency of the first period and the topology change frequency of the second period. The difference between the topology change frequency of the first period and the topology change frequency of the second period is calculated. The absolute value of the obtained difference is taken, and the ratio of the absolute value of the obtained difference to the topology change frequency of the first period is calculated to obtain the continuity deviation of the change frequency corresponding to the sample index combination segment.

[0105] Repeat the process of obtaining the continuous deviation of the change frequency corresponding to the sample index combination segment, obtain the continuous deviation of the change frequency corresponding to each dynamic index combination time period, and calculate the average of the obtained multiple continuous deviations of the change frequency to obtain the dynamic analysis coefficient of the identification model corresponding to the sample power area.

[0106] Repeat the process of obtaining the dynamic analysis coefficients of the identification model corresponding to the sample power area, and obtain the dynamic analysis coefficients of the identification model corresponding to each power area to obtain the dynamic analysis data of the identification model.

[0107] Step S2 analyzes the frequency of topology changes in the power area over a continuous period of time to generate dynamic analysis coefficients to match the model creation method. This can capture the time-varying patterns of regional topology changes (such as periodic fluctuations or sudden anomalies), making the model creation method more in line with the dynamic characteristics of actual operation. It avoids model lag or oversensitivity caused by static configuration, while optimizing the allocation of computing resources. Ultimately, it improves the timeliness and adaptability of topology change identification and ensures that the system maintains a stable and efficient response capability in a dynamic environment.

[0108] Step S3: Match the model creation method of the power area with the static analysis data and dynamic analysis data of the identification model, and create an area topology change identification model;

[0109] Step S3 further includes the following steps:

[0110] Obtain static analysis data of the identification model, and obtain the static analysis coefficients of the identification model corresponding to each power area based on the static analysis data of the identification model; obtain dynamic analysis data of the identification model, and obtain the dynamic analysis coefficients of the identification model corresponding to each power area based on the dynamic analysis data of the identification model.

[0111] Obtain the static matching interval and the dynamic matching interval of the recognition model respectively;

[0112] It should be noted here that:

[0113] The dynamic matching interval and the static matching interval of the recognition model are obtained as follows:

[0114] Historical power regions for creating regional topology change identification models using graph theory methods are obtained. Static analysis coefficients of the identification model corresponding to each historical power region are obtained, resulting in multiple historical static analysis coefficients. The numerical interval formed by the minimum and maximum historical static analysis coefficients is set as the static matching interval of the identification model. Dynamic analysis coefficients of the identification model corresponding to each historical power region are obtained, resulting in multiple historical dynamic analysis coefficients. The numerical interval formed by the minimum and maximum historical dynamic analysis coefficients is set as the dynamic matching interval of the identification model.

[0115] If the static analysis coefficients of the identification model are in the static matching interval of the identification model and the dynamic analysis coefficients of the identification model are in the dynamic matching interval of the identification model, then the graph theory method is used to create a regional topology change identification model for the power area.

[0116] If the static analysis coefficients of the identification model are not within the static matching range of the identification model or the dynamic analysis coefficients of the identification model are not within the dynamic matching range of the identification model, then machine learning methods are used to create a regional topology change identification model for the power area.

[0117] It should be noted here that:

[0118] In this application, graph theory methods abstract the power system into a graph structure (nodes = devices, edges = connections), and detect topology changes (such as line disconnections and node merging) through connectivity analysis, DFS / BFS, and other algorithms. Machine learning methods train models based on historical data to learn the differences between normal and abnormal topology patterns. Supervised learning (such as SVM and random forest) classifies topology states, unsupervised learning (such as clustering and autoencoders) detects unlabeled anomalies, and deep learning (such as GNN and LSTM) captures spatiotemporal dependencies.

[0119] The preferred embodiments of the present invention disclosed above are merely illustrative of the invention. These preferred embodiments do not exhaustively describe all details, nor do they limit the invention to any specific implementation. Clearly, many modifications and variations can be made based on the content of this specification. This specification selects and specifically describes these embodiments to better explain the principles and practical applications of the invention, thereby enabling those skilled in the art to better understand and utilize the invention. The invention is limited only by the claims and their full scope and equivalents.

Claims

1. A method for establishing a grid topology change identification model suitable for multiple operation scenarios, characterized in that, include: Step S1: Perform static topology index analysis on the power area, obtain the static analysis coefficients of the identification model corresponding to different power areas based on the analysis results, and obtain the static analysis data of the identification model; Step S2: Perform dynamic index analysis on the power area based on the static analysis data of the identification model, obtain the dynamic analysis coefficients of the identification model corresponding to different power areas based on the analysis results, and obtain the dynamic analysis data of the identification model. Step S3: Match the model creation method of the power area based on the static analysis data and dynamic analysis data of the identification model, and create an area topology change identification model based on the matching results.

2. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 1, characterized in that, Step S1 further includes the following steps: Step S11: Obtain the power regions that need to be identified for power grid topology changes, obtain multiple power regions, and arbitrarily select one sample power region from the multiple power regions obtained; Step S12: Perform structural complexity analysis on the sample power region, and obtain the static analysis coefficients of the identification model corresponding to the sample power region based on the analysis results; Step S13: Obtain the static analysis coefficients of the identification model corresponding to each power area to obtain the static analysis data of the identification model.

3. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 2, characterized in that, Step S12 further includes the following steps: Step S121: Obtain the power line topology map corresponding to the sample power area to obtain the sample power topology map; obtain the regional map corresponding to the sample power area to obtain the sample regional map; fuse the sample power topology map and the sample regional map using indicators to obtain the regional topology fusion image. Step S122: Obtain the topological layers corresponding to the sample power topology map to obtain multiple power topology levels. Obtain the geographical region corresponding to each power topology level in the regional topology fusion image to obtain the T1 topology sub-region to the Ta topology sub-region. Step S123: Perform topological complexity analysis on the T1 topological sub-region, and obtain the structural complexity of the T1 region based on the analysis results; Step S124: Obtain the regional structural complexity corresponding to the topological sub-regions T2 to Ta respectively, and obtain the regional structural complexity from T2 to Ta.

4. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 3, characterized in that, Step S12 further includes the following steps: Step S125: Obtain the area values ​​of topological sub-regions T1 to Ta in the regional topology fusion image, respectively, to obtain the area values ​​of region T1 to region Ta. Obtain the area values ​​of sample power regions in the regional topology fusion image, to obtain the area values ​​of sample regions. Calculate the ratio of each region area value to the sample region area value to obtain the area weights of region T1 to Ta. Step S126: Obtain the upper limit of the layer voltage corresponding to the topological sub-regions from T1 to Ta respectively, and obtain the upper limit of the layer voltage from T1 to Ta. Sum the upper limit of the layer voltage from T1 to Ta to obtain the cumulative value of the layer voltage in the region. Calculate the ratio of the upper limit of the layer voltage of each layer to the cumulative value of the layer voltage in the region to obtain the voltage weight from T1 to Ta. Step S127: Calculate the static analysis coefficients of the identification model corresponding to the sample power region by transforming the structural complexity of region T1 to the structural complexity of region Ta, the area weight of T1 to the area weight of Ta, and the voltage weight of T1 to the voltage weight of Ta.

5. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 4, characterized in that, Step S123 further includes the following steps: The power nodes existing in the T1 topological sub-region are obtained, from power node P1 to power node Pb. The number of power line connections from power node P1 to power node Pb is obtained respectively, thus obtaining the number of power connection edges from P1 to Pb. The structural complexity of region T1 is obtained by calculating the number of electrical connection edges from P1 to Pb. The structural complexity of region T1 is calculated using the following formula: ； Where Hf1 is the structural complexity of region T1, Lbi is the number of power connection edges in Pi, and b is the number of power nodes.

6. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 1, characterized in that, Step S2 further includes the following steps: Step S21: Obtain static analysis data of the identification model, and obtain the sample power area based on the static analysis data of the identification model; Step S22: Perform dynamic topology change analysis on the sample power area, and obtain the dynamic analysis coefficients of the identification model corresponding to the sample power area based on the analysis results; Step S23: Obtain the dynamic analysis coefficients of the identification model corresponding to each power area to obtain the dynamic analysis data of the identification model.

7. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 6, characterized in that, Step S22 further includes the following steps: Step S221: Select several dynamic index analysis periods of equal duration from the historical working periods corresponding to the sample power area; Step S222: Randomly select a sample dynamic time period from the multiple dynamic index analysis time periods, obtain the number of topology changes of the sample power area during the sample dynamic time period, obtain the time length corresponding to the sample dynamic time period, obtain the sample duration, calculate the ratio of the number of topology changes to the sample duration, and obtain the time period topology change frequency corresponding to the sample dynamic time period. Step S223: Obtain the time period topology change frequency corresponding to each dynamic indicator analysis period.

8. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 7, characterized in that, Step S22 further includes the following steps: Step S224: Set any two consecutive time periods of dynamic indicator analysis as a dynamic indicator combination period to obtain multiple dynamic indicator combination periods. Randomly select a sample indicator combination segment from the multiple dynamic indicator combination periods. Step S225: Perform time-period topology change frequency change analysis on the sample index combination time period, and obtain the change frequency continuity deviation based on the analysis results; Step S226: Obtain the continuous deviation of the change frequency corresponding to each dynamic index combination time period, and calculate the average of the obtained continuous deviations of the change frequency to obtain the dynamic analysis coefficient of the identification model corresponding to the sample power area.

9. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 8, characterized in that, Step S225 further includes the following steps: The two dynamic indicator analysis periods contained in the sample indicator combination segment are respectively set as the first dynamic analysis period and the second dynamic indicator analysis period; The frequency of topology changes for the first dynamic analysis period and the second dynamic index analysis period is obtained. The frequency of topology changes for the first period and the frequency of topology changes for the second period are obtained. The difference between the frequency of topology changes for the first period and the frequency of topology changes for the second period is calculated. The absolute value of the difference is taken, and the ratio of the absolute value of the difference to the frequency of topology changes for the first period is calculated to obtain the continuity deviation of the frequency of change corresponding to the sample index combination segment.

10. The method for establishing a power grid topology change identification model applicable to multiple operating scenarios according to claim 1, characterized in that, Step S3 further includes the following steps: Obtain static analysis data of the identification model, and obtain the static analysis coefficients of the identification model corresponding to each power area based on the static analysis data of the identification model; obtain dynamic analysis data of the identification model, and obtain the dynamic analysis coefficients of the identification model corresponding to each power area based on the dynamic analysis data of the identification model. Obtain the static matching interval and the dynamic matching interval of the recognition model respectively; If the static analysis coefficients of the identification model are in the static matching interval of the identification model and the dynamic analysis coefficients of the identification model are in the dynamic matching interval of the identification model, then the graph theory method is used to create a regional topology change identification model for the power area. If the static analysis coefficients of the identification model are not within the static matching range of the identification model or the dynamic analysis coefficients of the identification model are not within the dynamic matching range of the identification model, then machine learning methods are used to create a regional topology change identification model for the power area.

Images