An online emergency control strategy matching method adaptive to abrupt change of power grid mode

Abstract

The application discloses an online emergency control strategy matching method suitable for sudden change of power grid mode, which comprises the following steps: in response to change of the power grid mode, obtaining the changed power grid operation state data and a plurality of sets of historical power grid operation state data corresponding to a plurality of operation conditions and known online emergency control strategies; respectively calculating a time-varying index based on each set of obtained power grid operation state data; clustering the power grid mode according to the time-varying index to obtain a group to which the real-time power grid mode belongs; respectively determining a key parameter feature vector for the real-time power grid mode and the historical power grid mode in the group; calculating the similarity between the real-time power grid mode and each historical power grid mode based on the key parameter feature vector; and taking the online emergency control strategy corresponding to the historical power grid mode with the largest similarity as the strategy matching result. The online emergency control strategy after the sudden change of the power grid mode is obtained through twice matching, and the speed and accuracy of the strategy matching are improved, so that the safe and stable operation of the power grid is ensured.

Description

TECHNICAL FIELD

[0001] The present application relates to the technical field of automatic operation control of power systems, and particularly relates to an online emergency control strategy matching method suitable for sudden changes in power grid modes. BACKGROUND

[0002] With the rapid development of ultra-high voltage direct current, new energy and power markets, the operation mode of power systems is changing and the uncertainty is significantly enhanced. There are fluctuations of tens of millions of kilowatts in the power transmission section of the power grid within a day, and the power change on the minute time scale can reach about 1% of the installed capacity (about millions of kilowatts). The influence on the transient time scale, especially after the superposition of power grid faults, cannot be ignored.

[0003] The methods currently used in actual systems generally include time domain simulation method, machine learning method and historical data mining. In online safety and stability analysis with a period of 5-15 minutes, the time domain simulation method is generally used, but when the power grid mode suddenly changes, the time domain simulation search strategy method which requires a long time cannot be used, which may cause a long period of vacancy of the power grid safety and stability strategy. The machine learning technology is used for transient stability evaluation, which needs to establish the mapping relationship between some key system parameters such as generator power angle, active power and transient stability margin, but the machine learning method based on local information has extremely limited generalization ability in the power system with strong nonlinearity and strong time variability, and it is difficult to meet the requirements of actual engineering application. The historical data mining method is used, that is, the current mode characteristics are matched with the historical modes, but due to the complex change characteristics of the power grid, it is often impossible to match the completely consistent power grid operation mode.

[0004] Therefore, there is an urgent need for a power grid safety and stability strategy matching method suitable for sudden changes in power grid modes to improve the matching precision of the strategy and ensure the continuous safe and stable operation of the power grid. SUMMARY

[0005] The purpose of the present application is to provide an online emergency control strategy matching method suitable for sudden changes in power grid modes, to improve the matching precision of the power grid operation mode, to efficiently and quickly obtain a suitable online emergency control strategy, and to ensure the continuous safe and stable operation of the power grid. The technical solution adopted by the present application is as follows.

[0006] In one aspect, the present application provides an online emergency control strategy matching method suitable for sudden changes in power grid modes, comprising:

[0007] In response to a change in the power grid mode, obtaining the changed power grid operation state data and historical database data, wherein the historical database data includes multiple sets of historical power grid operation state data corresponding to multiple operation conditions and respectively known online emergency control strategies;

[0008] Based on the acquired data, the transient stability characteristics corresponding to the changed power grid operation status data and each set of historical power grid operation status data are determined respectively.

[0009] Based on the transient stability characteristics, the power grid operation status data after the change and multiple sets of historical power grid operation status data are clustered to obtain the group to which the real-time power grid mode corresponding to the changed power grid operation status data belongs, and the group is used as the target group for secondary matching.

[0010] Based on the changed power grid operation status data within the secondary matching target group and the historical power grid operation status data of each set, the key parameter feature vectors of the corresponding real-time power grid mode and historical power grid mode are determined respectively.

[0011] Based on the key parameter feature vectors, the similarity between the real-time power grid mode and each historical power grid mode is calculated;

[0012] The online emergency control strategy corresponding to the historical power grid mode with the highest similarity to the real-time power grid mode is taken as the strategy matching result.

[0013] Optionally, the method further includes: substituting the online emergency control strategy corresponding to the historical power grid mode with the highest similarity to the real-time power grid mode into the real-time power grid mode for simulation verification.

[0014] If there are no safety or stability issues, the online emergency control strategy will be sent to the execution device for power grid control. If there are safety issues, based on control experience and the control measures corresponding to the online emergency control strategy, the control measure increment will be determined, the control measure increment will be simulated and verified, and the control measure increment will be sent to the execution device for power grid control after the simulation verification is passed.

[0015] Therefore, based on the rapid acquisition of online security and stability control strategies, we can further ensure the applicability of the strategies to real-time power grid conditions and ensure the safe operation of the power grid.

[0016] Optionally, the power grid operating status data includes conventional power grid characteristics. and transient stability characteristics, among which, This represents the cross-sectional power characteristic of the power grid system in the k-th set of data. This represents the start-up mode that characterizes the power grid system in the k-th set of data. This represents the load situation that characterizes the power grid system in the k-th set of data. This represents the topology characterizing the power grid system in the k-th set of data. Transient stability features include the SEEAC margin η. kSE The SEEAC margin can be calculated using simulation software; specific details can be found in existing technologies. The aforementioned conventional power grid characteristics provide the data foundation for the relevant calculations in this invention.

[0017] Optionally, the calculation formula of the changed power grid operation state data and the time-varying index corresponding to each set of the historical power grid operation state data is:

[0018]

[0019] In the formula, σ k represents the time-varying index corresponding to the kth set of power grid operation state data, η kDE and η kSE respectively represent the DSEEAC margin and the SEEAC margin corresponding to the kth set of power grid operation state data. The time-varying index, the DSEEAC margin and the SEEAC margin as the transient stability characteristics of the power grid mode can represent the power grid transient trajectory information, and based on this, the matching speed and the matching accuracy of the power grid mode can be improved.

[0020] Optionally, the power grid mode clustering of the changed power grid operation state data and the multiple sets of historical power grid operation state data according to the transient stability characteristics comprises:

[0021] According to the time-varying index, the DSEEAC margin and the SEEAC margin, the K-means algorithm is used for clustering, wherein the Euclidean distance calculation formula of the real-time power grid mode and the cluster center is:

[0022]

[0023] In the formula, d(x, C i ) represents the Euclidean distance of the real-time power grid mode x and the cluster center C i , σ x represents the time-varying index of the real-time power grid mode, represents the time-varying index of the cluster center power grid mode, represents the SEEAC margin corresponding to the real-time power grid mode, represents the DEEAC margin corresponding to the cluster center power grid mode.

[0024] In the above scheme, when the K-means algorithm clustering is performed, the number of cluster centers and the cluster centers can be adjusted offline according to different stability control defense regions, so that the historical modes close to the current power grid mode exist in the aggregation region, and the number in the aggregation region is avoided to be too large, and the modes with large difference in time-varying index are matched.

[0025] Optionally, the key parameter feature vector of the corresponding real-time power grid mode and the historical power grid mode is determined based on the changed power grid operation state data and each set of historical power grid operation state data in the secondary matching target group, comprising:

[0026] 1) Determine the key parameter B for the real-time power grid mode and the key parameter A for each historical power grid mode;

[0027] 2) Calculate the participation factor λ for each key parameter;

[0028] 3) The key parameter feature vector D for real-time power grid mode and the key parameter feature vector C for historical power grid mode are determined as follows:

[0029]

[0030]

[0031] In the formula, A k , λ k C k Let represent the key parameters, participation factors, and feature vectors of the key parameters corresponding to the k-th historical power grid operation status data, respectively. This is the set of key parameters for the k-th historical power grid mode. For the corresponding set of participating factors, Key parameters The participating factors; {b1,b2,…,b n} represents the set of key parameters for the real-time power grid mode. This represents the corresponding set of participating factors.

[0032] To reduce the impact of the absolute magnitude of key parameters on the matching results, this invention employs a similarity matching method that considers participation factors. It introduces a feature vector that takes participation factors as weights into the secondary matching target group obtained after clustering, performing secondary matching on real-time and historical power grid modes to further improve matching accuracy. Different power grids can determine the key parameters of their mode characteristics based on actual operating experience.

[0033] Optionally, the key parameters include the transient power angle stability characteristics of the synchronous power source, the transient power angle stability characteristics of the asynchronous power source, and the load sensitivity to the transient power angle stability margin; wherein,

[0034] The formula for calculating the participation factor of the transient power angle stability characteristics of a synchronous power source is:

[0035] In the formula, λ n E represents the participation factor of the nth generator. kn E represents the acceleration kinetic energy of the nth generator at the dynamic saddle point (DSP). kmax This represents the maximum acceleration kinetic energy of all units at the DSP point;

[0036] The formula for calculating the participation factor of the transient power angle stability characteristics of asynchronous power sources is as follows:

[0037] In the formula, λ i a represents the transient power angle stability participation factor of wind farm i. j The variable j represents the transient power angle stability participation factor of conventional units, N represents the number of units in group S or group A, and x represents the number of units in group A. i,j This represents the equivalent reactance between the grid connection point bus of wind farm i and the conventional unit j bus;

[0038] The formula for calculating the participation factor of the load in the transient power angle stability margin sensitivity is as follows:

[0039] Where, λ k Δη represents the sensitivity of load k to transient power angle stability margin, Δη represents the change in stability margin, and ΔC represents the change in load k.

[0040] Optionally, calculating the similarity between the real-time power grid mode and each historical power grid mode based on the key parameter feature vector includes: using a cosine similarity calculation method to calculate the cosine similarity between the key parameter feature vector of the real-time power grid mode and the key parameter feature vector of each historical power grid mode, with the calculation formula as follows:

[0041]

[0042] In the formula, cosθ k C represents the cosine similarity between the real-time power grid mode and the k-th historical power grid mode. k Let represent the key parameter feature vector of the k-th historical power grid mode, and D represent the key parameter feature vector of the real-time power grid mode.

[0043] The cosine similarity method is a commonly used method for comparing the relationship between vectors. This invention calculates the cosine value of the feature vector after the power grid mode change and the feature vector of all other historical power grid modes in the secondary matching target group. The closer the result is to 1, the smaller the angle between the feature vectors of the power grid modes, and the closer the power grid modes are.

[0044] Secondly, the present invention provides an online emergency control strategy matching device adapted to sudden changes in power grid operation, comprising:

[0045] The data acquisition module is configured to acquire the changed power grid operation status data and historical database data in response to changes in the power grid mode. The historical database data includes multiple sets of historical power grid operation status data corresponding to various operating conditions and for which online emergency control strategies are known.

[0046] The time-varying index calculation module is configured to determine the transient stability characteristics corresponding to the changed power grid operation state data and each set of historical power grid operation state data based on the acquired data, respectively.

[0047] The first matching module is configured to perform power grid mode clustering on the changed power grid operation state data and the multiple sets of historical power grid operation state data according to the transient stability characteristics, to obtain a group to which a real-time power grid mode corresponding to the changed power grid operation state data belongs, and to take the group as a second matching target group.

[0048] The key parameter determination module is configured to determine the key parameter feature vectors of the real-time power grid mode and the historical power grid mode, respectively, based on the changed power grid operation state data and each set of historical power grid operation state data in the second matching target group.

[0049] The second matching module is configured to calculate the similarity of the real-time power grid mode and each historical power grid mode based on the key parameter feature vectors.

[0050] The strategy selection module is configured to take an online emergency control strategy corresponding to a historical power grid mode with the greatest similarity to the real-time power grid mode as a strategy matching result.

[0051] Further, the online emergency control strategy matching device further includes a strategy checking module configured to substitute the online emergency control strategy corresponding to the historical power grid mode with the greatest similarity to the real-time power grid mode into the real-time power grid mode for simulation checking verification.

[0052] If there is no security and stability problem, the online emergency control strategy is issued to an execution device for power grid control. If there is a security problem, a control measure increment is determined based on the control measures corresponding to the online emergency control strategy according to control experience, the control measure increment is simulated and checked, and the control measure increment is issued to the execution device for power grid control after the simulation checking verification is passed.

[0053] Optionally, the first matching module performs power grid mode clustering on the changed power grid operation state data and the multiple sets of historical power grid operation state data according to the transient stability characteristics, and includes:

[0054] The K-means algorithm is used for clustering according to the time-varying index, the DSEEAC margin and the SEEAC margin, wherein a Euclidean distance calculation formula of the real-time power grid mode and a clustering center is:

[0055]

[0056] In the formula, d(x, C i ) represents the Euclidean distance between the real-time power grid mode x and the clustering center Ci Euclidean distance of the real-time power grid mode, σ x denotes a time-varying index of the real-time power grid mode, denotes a time-varying index of the clustering center power grid mode, denotes a SEEAC margin corresponding to the real-time power grid mode, denotes a DEEAC margin corresponding to the clustering center power grid mode.

[0057] In a third aspect, the present application provides a computer readable storage medium, which stores a computer program, and the computer program is executed by a processor to implement the online emergency control strategy matching method according to the first aspect.

[0058] Advantages

[0059] The online emergency control strategy fast matching method according to the present application firstly performs the first strategy matching by using the time-varying index which can be quickly calculated and the power grid mode clustering, so as to narrow the matching range, and then performs the second matching by using the similarity of the key parameter feature vectors, so as to obtain the final online emergency control strategy. The online emergency control strategy fast matching method according to the present application can quickly complete the matching of the power grid safety and stability control measures when the power grid flow and topology change, greatly reduces the blank period of the power grid safety and stability control measures, and the two times of matching improves the accuracy of the strategy matching, so as to guarantee the safe and stable operation of the power grid. BRIEF DESCRIPTION OF DRAWINGS

[0060] Figure 1 Fig. 1 shows a flowchart of an embodiment of the online emergency control strategy matching method according to the present application;

[0061] Figure 2 Fig. 2 shows a flowchart of a specific implementation of the online emergency control strategy matching method according to the present application. DETAILED DESCRIPTION

[0062] The present application will be further described below in combination with the drawings and specific embodiments.

[0063] Embodiment 1

[0064] This embodiment introduces an online emergency control strategy matching method which is suitable for power grid mode mutation, and the method is based on the online emergency control strategy matching method according to the present application. Figure 1 The method comprises the following steps:

[0065] It is judged whether the power grid mode changes or not, and in response to the change of the power grid mode, the changed power grid operation state data and the historical database data are obtained, wherein the historical database data comprises multiple sets of historical power grid operation state data corresponding to multiple operation conditions and respectively known online emergency control strategies. The signal of the change of the power grid mode can be received from the outside;

[0066] Based on the acquired data, the transient stability characteristics corresponding to the changed power grid operation state data and each set of historical power grid operation state data are determined respectively;

[0067] According to the transient stability characteristics, the power grid mode clustering is performed on the changed power grid operation state data and the multiple sets of historical power grid operation state data, to obtain a group to which a real-time power grid mode corresponding to the changed power grid operation state data belongs, and the group is taken as a secondary matching target group;

[0068] Based on the changed power grid operation state data and each set of historical power grid operation state data in the secondary matching target group, the key parameter feature vectors of the corresponding real-time power grid mode and historical power grid mode are determined respectively;

[0069] Based on the key parameter feature vectors, the similarity between the real-time power grid mode and each historical power grid mode is calculated;

[0070] The online emergency control strategy corresponding to the historical power grid mode with the largest similarity to the real-time power grid mode is taken as a strategy matching result.

[0071] Reference Figure 2 As shown in the figure, the method of the embodiment involves the following aspects when applied.

[0072] I. Establishing a historical database

[0073] The embodiment adopts a method of combining online real-time data and offline data to build a historical database. In the building process, online data needs to completely cover all working conditions of daily operation, but in order to ensure the universality of the historical database, offline construction needs to be used for some special situations of the power grid, so as to expand the application range of the power grid emergency control.

[0074] For each set of power grid operation state data in the database, a mode word for describing the conventional power grid mode needs to be added to form a conventional power grid feature index library Among them, represents the section power representing the characteristics of the power grid system in the kth set of historical data, represents the start mode representing the characteristics of the power grid system in the kth set of historical data, represents the load condition representing the characteristics of the power grid system in the kth set of historical data, represents the topological structure representing the characteristics of the power grid system in the kth set of historical data.

[0075] In addition, the transient stability characteristics of each set of power grid operation state data also need to be added to form a transient stability feature index library S k {η kSE ,σ k}, wherein η kSErepresents the SEEAC margin result corresponding to the kth set of power grid operation state data calculated by simulation software, σ k represents the power grid time-varying index result corresponding to the kth set of power grid operation state data.

[0076] For each set of historical power grid operation state data in the historical database, the SEEAC margin and the time-varying index can be pre-calculated, or the time-varying index can also be calculated together with the real-time power grid operation state data of the online emergency control strategy to be matched before the power grid mode clustering. Among them, the time-varying index σ k The calculation formula is as follows:

[0077]

[0078] In the formula, η kDE represents the DSEEAC margin result corresponding to the kth set of power grid operation state data calculated by simulation software. The DSEEAC margin result of the historical operation state data can also be pre-calculated.

[0079] In actual online emergency control strategy matching for power grid mode change scenarios, after obtaining the real-time power grid operation state data, the corresponding characteristic values are determined according to the above-mentioned conventional power grid characteristics and transient stability characteristics extraction method of the historical operation state data, which are used for subsequent logical operation.

[0080] II. Power grid mode clustering

[0081] In this embodiment, the power grid mode clustering is realized according to the transient stability characteristics to determine the first matching, so as to determine the range of historical power grid modes similar to the real-time power grid operation state, that is, to determine the target group for the second matching.

[0082] In this embodiment, the K-means algorithm is used for clustering. When clustering, the number of clustering centers and the prepared clustering centers can be adjusted offline according to different stability control defense regions, so as to ensure that there are historical modes close to the current power grid mode in the aggregated region, and avoid that there are too many in the aggregated region and the modes with large differences in time-varying index are matched. Specifically, according to the needs of different power grids, the clustering center and the initial value of the number of clustering centers are set according to the operation experience and the index calculation result, and then the clustering center and the number thereof are obtained based on repeated tests on the power grid modes in the historical database.

[0083] In this embodiment, when the K-means algorithm is used for clustering according to the time-varying index, the DSEEAC margin and the SEEAC margin, the clustering index adopts the Euclidean distance, and the Euclidean distance calculation formula of the real-time power grid mode and the clustering center is as follows:

[0084]

[0085] In the formula, d(x, Ci represents the Euclidean distance between real-time power grid mode x and cluster center C i represents the time-varying index of the cluster center power grid mode, x represents the time-varying index of the real-time power grid mode, represents the time-varying index of the cluster center power grid mode, represents the SEEAC margin corresponding to the real-time power grid mode, represents the DEEAC margin corresponding to the cluster center power grid mode.

[0086] After completing the K-means clustering, the power grid mode is divided into K regions, at this time, all historical modes in the region to which the post-mutation power grid mode belongs have close transient stability characteristics with the real-time power grid mode, so all historical modes in the region are candidate object modes for further strategy matching, and the region is taken as a secondary matching target group.

[0087] III. Calculate the key parameter feature vector for secondary matching

[0088] After determining the secondary matching target group, the matching range is narrowed down, and only the key parameter feature vectors of the real-time power grid mode and the historical power grid mode corresponding to the post-change power grid operation state data and each set of historical power grid operation state data in the secondary matching target group need to be determined, and further refined power grid mode matching is realized based on the key parameter feature vectors. The specific steps are as follows.

[0089] 3.1 Determine the key parameter B of the real-time power grid mode and the key parameter A of each historical power grid mode respectively

[0090] The types of key parameters in the power grid key parameter feature vector can be selected according to the actual power grid and experience. Here, typical key parameters include transient power angle stability characteristics of synchronous power sources, transient power angle stability characteristics of non-synchronous power sources, and sensitivity of load to transient power angle stability margin, etc.

[0091] 3.2 In order to reduce the influence of the absolute size of each key parameter on the matching result, the participation factor is introduced as the weight feature vector in this embodiment. For the foregoing example of each key parameter, the calculation method of their participation factor λ is as follows:

[0092] The participation factor calculation formula of the transient power angle stability characteristics of synchronous power sources is:

[0093] In the formula, λ n represents the participation factor of the nth generator, E kn represents the acceleration kinetic energy of the nth generator at the dynamic saddle point (DSP), E kmax represents the maximum acceleration kinetic energy of all units at the DSP point;

[0094] The formula for calculating the participation factor of the transient power angle stability characteristics of asynchronous power sources is as follows:

[0095] In the formula, λ i a represents the transient power angle stability participation factor of wind farm i. j The variable j represents the transient power angle stability participation factor of conventional units, N represents the number of units in group S or group A, and x represents the number of units in group A. i,j This represents the equivalent reactance between the grid connection point bus of wind farm i and the conventional unit j bus;

[0096] The formula for calculating the participation factor of the load in the transient power angle stability margin sensitivity is as follows:

[0097] Where, λ k Δη represents the sensitivity of load k to transient power angle stability margin, Δη represents the change in stability margin, and ΔC represents the change in load k.

[0098] 3.3 The key parameter feature vectors of the real-time power grid mode after the mutation and each historical power grid mode within the secondary matching target group can be determined from the selected and determined key parameters and their participating factors:

[0099]

[0100]

[0101] In the formula, A k , λ k C k Let represent the key parameters, participation factors, and feature vectors of the key parameters corresponding to the k-th historical power grid operation status data, respectively. This is the set of key parameters for the k-th historical power grid mode. For the corresponding set of participating factors, Key parameters The participating factors; {b1,b2,…,b n} represents the set of key parameters for the real-time power grid mode. This represents the corresponding set of participating factors.

[0102] IV. Quadratic matching based on key parameter feature vectors

[0103] After obtaining the key parameter feature vectors of the real-time power grid mode after the mutation and each historical power grid mode in the secondary matching target group, this embodiment adopts a cosine similarity matching method that takes into account the participation factor to perform secondary matching of power grid mode and online emergency control strategy.

[0104] Specifically, based on the key parameter feature vector, the similarity between the real-time power grid mode and each historical power grid mode is calculated, including: using the cosine similarity calculation method, the cosine similarity between the key parameter feature vector of the real-time power grid mode and the key parameter feature vector of each historical power grid mode is calculated respectively, and the calculation formula is:

[0105]

[0106] In the formula, cosθ k represents the cosine similarity between the real-time power grid mode and the kth historical power grid mode, C k represents the key parameter feature vector of the kth historical power grid mode, and D represents the key parameter feature vector of the real-time power grid mode.

[0107] The cosine similarity method is a commonly used method for comparing the relationship between vectors. In the present application, the cosine value between the feature vector after the sudden change of the power grid mode and the feature vectors of all other historical power grid modes in the secondary matching target group is calculated. The closer the result is to 1, the smaller the angle between the feature vectors of the power grid mode, and the closer the power grid mode is.

[0108] At this point, the historical power grid mode closest to the real-time power grid mode after the sudden change can be taken as the matching result, and the online emergency control strategy corresponding to the historical power grid mode is taken as the online emergency control strategy of the real-time power grid mode.

[0109] After obtaining the strategy matching result, in order to further ensure the adaptability of the online emergency control strategy to the power grid after the sudden change, the online emergency control strategy corresponding to the historical power grid mode with the largest similarity to the real-time power grid mode obtained by matching is substituted into the real-time power grid mode for simulation verification: if there is no security and stability problem in the verification process, the online emergency control strategy is issued to the execution device for power grid control; if there is a security problem, the control measure increment is determined based on the control measures corresponding to the online emergency control strategy according to the control experience, the control measure increment is simulated and verified, and the control measure increment is issued to the execution device for power grid control after the simulation verification is passed.

[0110] In summary, the online emergency control strategy matching after the sudden change of the power grid operation mode is completed, the target range is quickly narrowed through the first matching, and the matching result is locked by using the cosine similarity based on the key feature vector through the second matching. The present application can realize efficient, rapid and accurate control strategy matching after the sudden change of the power grid mode, make up for the blank period of control measures in the traditional control process, and ensure the safe and stable operation of the power grid.

[0111] Embodiment 2

[0112] Based on the same inventive concept as example 1, this embodiment introduces an online emergency control strategy matching device adapting to sudden change of power grid mode, which comprises:

[0113] a data acquisition module configured to acquire changed power grid operation state data and historical database data in response to change of power grid mode, wherein the historical database data comprises multiple sets of historical power grid operation state data corresponding to multiple operation conditions and respectively known online emergency control strategies;

[0114] a time-varying index calculation module configured to calculate time-varying indexes corresponding to the changed power grid operation state data and each set of the historical power grid operation state data respectively based on the acquired data;

[0115] a primary matching module configured to perform power grid mode clustering on the changed power grid operation state data and the multiple sets of historical power grid operation state data according to the time-varying indexes, to obtain a group to which a real-time power grid mode corresponding to the changed power grid operation state data belongs, and to take the group as a secondary matching target group;

[0116] a key parameter determination module configured to determine key parameter feature vectors of the real-time power grid mode and the historical power grid mode respectively based on the changed power grid operation state data and each set of the historical power grid operation state data in the secondary matching target group;

[0117] a secondary matching module configured to calculate similarities between the real-time power grid mode and each historical power grid mode based on the key parameter feature vectors;

[0118] and a strategy selection module configured to take an online emergency control strategy corresponding to a historical power grid mode having the greatest similarity with the real-time power grid mode as a strategy matching result.

[0119] Further, the online emergency control strategy matching device of this embodiment further comprises a strategy checking module configured to substitute the online emergency control strategy corresponding to the historical power grid mode having the greatest similarity with the real-time power grid mode into the real-time power grid mode for simulation checking verification:

[0120] If there is no security and stability problem, the online emergency control strategy is issued to an execution device for power grid control; if there is a security problem, a control measure increment is determined based on a control measure corresponding to the online emergency control strategy according to control experience, the control measure increment is simulated and checked, and the online emergency control strategy is issued to the execution device for power grid control after the simulation checking verification is passed.

[0121] The specific implementation of each functional module is referred to the corresponding description in example 1, which will not be repeated. It is particularly pointed out that:

[0122] The one-time matching module performs power grid mode clustering on the changed power grid operation state data and the multiple sets of historical power grid operation state data according to the time-varying index, and includes the following steps.

[0123] According to the calculated time-varying index, K-means algorithm is used for clustering, wherein the Euclidean distance calculation formula of the real-time power grid mode and the clustering center is as follows:

[0124]

[0125] In the formula, d(x, C i ) represents the Euclidean distance of the real-time power grid mode x and the clustering center C i , σ x represents the time-varying index of the real-time power grid mode, represents the time-varying index of the clustering center power grid mode, represents the SEEAC margin corresponding to the real-time power grid mode, represents the DEEAC margin corresponding to the clustering center power grid mode.

[0126] Embodiment 3

[0127] This embodiment introduces a computer readable storage medium, which stores a computer program. When the computer program is executed by a processor, the online emergency control strategy matching method as introduced in Embodiment 1 is realized.

[0128] In summary of the above embodiments, the online emergency control strategy fast matching method of the present application can quickly complete the matching of power grid safety and stability control measures when the power grid flow and topology change, make up for the blank period of power grid safety and stability control measures, and improve the precision of strategy matching, and guarantee the safe and stable operation of the power grid.

[0129] Those skilled in the art should understand that the embodiments of the present application can be provided as a method, a system, or a computer program product. Therefore, the present application can be in the form of a complete hardware embodiment, a complete software embodiment, or an embodiment combining software and hardware aspects. Moreover, the present application can be in the form of a computer program product implemented on one or more computer usable storage media (including but not limited to disk storage, CD-ROM, optical storage, etc.) containing computer usable program code.

[0130] This application is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of this application. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart... Figure 1 One or more processes and / or boxes Figure 1 A device that provides the functions specified in one or more boxes.

[0131] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0132] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0133] The embodiments of the present invention have been described above with reference to the accompanying drawings. However, the present invention is not limited to the specific embodiments described above. The specific embodiments described above are merely illustrative and not restrictive. Those skilled in the art can make many other forms under the guidance of the present invention without departing from the spirit and scope of the claims. All of these forms are within the protection scope of the present invention.

Claims

1. A method for matching online emergency control strategies to adapt to sudden changes in power grid operation, characterized in that, include: In response to changes in the power grid mode, the system acquires the changed power grid operation status data and historical database data, wherein the historical database data includes multiple sets of historical power grid operation status data corresponding to various operating conditions and for which online emergency control strategies are known. Based on the acquired data, the transient stability characteristics corresponding to the changed power grid operation status data and each set of historical power grid operation status data are determined respectively. Based on the transient stability characteristics, the power grid operation status data after the change and multiple sets of historical power grid operation status data are clustered to obtain the group to which the real-time power grid mode corresponding to the changed power grid operation status data belongs, and the group is used as the target group for secondary matching. Based on the changed power grid operation status data within the secondary matching target group and the historical power grid operation status data of each set, the key parameter feature vectors of the corresponding real-time power grid mode and historical power grid mode are determined respectively. Based on the key parameter feature vectors, the similarity between the real-time power grid mode and each historical power grid mode is calculated; The online emergency control strategy corresponding to the historical power grid mode that has the greatest similarity to the real-time power grid mode is taken as the strategy matching result; The calculation formula used for the time-varying indicators corresponding to the changed power grid operating status data and each set of historical power grid operating status data is as follows: ， In the formula, Indicates the first Time-varying indicators corresponding to power grid operation status data, and They represent the first The DSEEAC margin and SEEAC margin corresponding to the power grid operation status data; The method of clustering power grid operation mode based on transient stability characteristics of changed power grid operation status data and multiple sets of historical power grid operation status data includes: Based on the time-varying indices, DSEEAC margin, and SEEAC margin, K-means algorithm is used for clustering. The Euclidean distance between the real-time power grid and the cluster centers is calculated using the following formula: ， In the formula, Representing real-time power grid mode With cluster center European distance, Indicators representing the time-varying nature of real-time power grid operation. Indicators representing the time-varying nature of cluster-centric power grid patterns This indicates the SEEAC margin for the corresponding real-time power grid mode. This represents the DEEAC margin for the corresponding cluster center grid method.

2. The method according to claim 1, characterized in that, it further... include: The online emergency control strategy corresponding to the historical power grid mode that has the greatest similarity to the real-time power grid mode is substituted into the real-time power grid mode for simulation verification: If there are no safety and stability issues, the online emergency control strategy will be sent to the execution device for power grid control. If a safety issue exists, based on control experience and the control measures corresponding to the online emergency control strategy, the incremental control measures are determined, the incremental control measures are verified by simulation, and after the simulation verification is passed, they are sent to the execution device for power grid control.

3. The method according to claim 1, characterized in that, The power grid operation status data includes conventional power grid characteristics. and transient stability characteristics, among which, Indicates the first The cross-sectional power in the dataset characterizing the power grid system. Indicates the first The start-up methods that characterize the power grid system in the set of data. Indicates the first The load conditions that characterize the power grid system in the set of data. Indicates the first The topology characterizing the power grid system in the dataset, including transient stability features such as SEEAC margin. .

4. The method according to claim 1, characterized in that, The key parameter feature vectors for determining the corresponding real-time grid mode and historical grid mode based on the changed grid operation status data within the secondary matching target group and each set of historical grid operation status data include: 1) Determine the key parameters for real-time power grid modes respectively. Key parameters of various historical power grid modes ; 2) Calculate the participation factors for each key parameter. ; 3) Determine the key parameter feature vectors of the real-time power grid mode. Key parameter feature vectors of historical power grid methods for: ， ， In the formula, , , They represent the first The first set of historical power grid operation status data corresponding to the Key parameters of historical power grid modes, key parameter participation factors, and key parameter feature vectors. For the first A set of key parameters for historical power grid modes For the corresponding set of participating factors, Key parameters Participating factors; This is the set of key parameters for the real-time power grid mode. This represents the corresponding set of participating factors.

5. The method according to claim 4, characterized in that, The key parameters include the transient power angle stability characteristics of synchronous power sources, the transient power angle stability characteristics of asynchronous power sources, and the load sensitivity to transient power angle stability margin; among which... The formula for calculating the participation factor of the transient power angle stability characteristics of a synchronous power source is: , In the formula, Indicates the first The participation factors of each generator, Indicates the first The acceleration kinetic energy of a generator at the dynamic saddle point DSP This represents the maximum acceleration kinetic energy of all units at the DSP point; The formula for calculating the participation factor of the transient power angle stability characteristics of asynchronous power sources is as follows: , In the formula, Indicates wind farm Transient work angle stability participation factor, Indicates conventional units Transient work angle stability participation factor Indicates the number of S-group units or A-group units. Indicates wind farm Grid connection point bus and conventional units Equivalent reactance between busbars; The formula for calculating the participation factor of the load in the transient power angle stability margin sensitivity is as follows: , in, Indicates load Sensitivity to transient work angle stability margin This represents the change in stability margin. Indicates load The change in quantity.

6. The method according to claim 1, characterized in that, The step of calculating the similarity between the real-time power grid mode and each historical power grid mode based on the key parameter feature vector includes: using the cosine similarity calculation method to calculate the cosine similarity between the key parameter feature vector of the real-time power grid mode and the key parameter feature vector of each historical power grid mode, respectively. The calculation formula is as follows: ， In the formula, Indicates the real-time power grid mode and the first Cosine similarity between historical power grid methods Indicates the first Key parameter feature vectors of historical power grid modes The key parameter feature vectors representing real-time power grid modes.

7. An online emergency control strategy matching device for adapting to sudden changes in power grid operation, employing the online emergency control strategy matching method according to any one of claims 1 to 6, characterized in that, include: The data acquisition module is configured to acquire the changed power grid operation status data and historical database data in response to changes in the power grid mode. The historical database data includes multiple sets of historical power grid operation status data corresponding to various operating conditions and for which online emergency control strategies are known. The time-varying index calculation module is configured to determine the transient stability characteristics corresponding to the changed power grid operating state data and each set of historical power grid operating state data based on the acquired data. The primary matching module is configured to perform grid mode clustering on the changed grid operation status data and multiple sets of historical grid operation status data based on the transient stability characteristics, to obtain the group to which the real-time grid mode corresponding to the changed grid operation status data belongs, and to use the group as the target group for secondary matching. The key parameter determination module is configured to determine the key parameter feature vectors of the corresponding real-time power grid mode and the historical power grid mode based on the changed power grid operation status data within the secondary matching target group and the historical power grid operation status data of each set. The secondary matching module is configured to calculate the similarity between the real-time power grid mode and each historical power grid mode based on the key parameter feature vector. Additionally, the strategy selection module is configured to use the online emergency control strategy corresponding to the historical power grid mode that has the greatest similarity to the real-time power grid mode as the strategy matching result.

8. The online emergency control strategy matching device according to claim 7, characterized in that, It also includes a strategy verification module, which is configured to: substitute the online emergency control strategy corresponding to the historical power grid mode with the highest similarity to the real-time power grid mode into the real-time power grid mode for simulation verification. If there are no safety and stability issues, the online emergency control strategy will be sent to the execution device for power grid control. If a safety issue exists, based on control experience and the control measures corresponding to the online emergency control strategy, the incremental control measures are determined, the incremental control measures are verified by simulation, and after the simulation verification is passed, they are sent to the execution device for power grid control.

9. The online emergency control strategy matching device according to claim 7, characterized in that, The first-level matching module performs grid mode clustering on the changed grid operation status data and multiple sets of historical grid operation status data based on transient stability characteristics, including: Based on the time-varying indices, DSEEAC margin, and SEEAC margin, K-means algorithm is used for clustering. The Euclidean distance between the real-time power grid and the cluster centers is calculated using the following formula: ， In the formula, Representing real-time power grid mode With cluster center European distance, Indicators representing the time-varying nature of real-time power grid operation. Indicators representing the time-varying nature of cluster-centric power grid patterns This indicates the SEEAC margin for the corresponding real-time power grid mode. This represents the DEEAC margin for the corresponding cluster center grid method.

10. A computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the online emergency control strategy matching method as described in any one of claims 1-6.

Images