A transient voltage stabilization emergency control method, device and power control system

Abstract

The present invention discloses a transient voltage stability emergency control method, device and power control system, which belongs to the technical field of power system stability control. The method includes: the first stage of the present invention is to decide on the load shedding point, and the second stage is to decide on the load shedding amount. The first stage: determine the current voltage instability mode according to the current fault state information and mapping relationship in the actual operation of the power system; determine the screening and sorting of multiple load shedding points according to the current voltage instability mode; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode; the second stage: input the current fault state information into the lightweight gradient boosting machine LightGBM model to estimate the total load shedding amount; for the current voltage instability mode, perform emergency control based on the screening and sorting and the total load shedding amount. The present invention can reduce the time and expert workload for formulating offline emergency load shedding control measures, improve work efficiency, and has strong practicality.

Description

Technical Field

[0001] The present invention belongs to the technical field of power system stability control, and more specifically, relates to a transient voltage stability emergency control method, device and power control system. Background Art

[0002] As the second line of defense for the safe and stable operation of my country's power system, emergency control is an important and indispensable measure to maintain the transient safety and stability of the system after a serious fault or disturbance occurs. Load shedding control is the most effective and most commonly used measure for transient voltage instability. At present, the offline pre-decision-making and real-time condition matching execution mode is generally adopted for emergency load shedding control. Offline pre-decision-making is generally performed by power grid operation technical experts. They analyze transient simulation results based on domain knowledge and operation experience, and obtain effective emergency load shedding measures through continuous trial and error and adjustment. However, due to the large number of typical operating conditions and transient voltage instability faults in modern large-scale interconnected and complex power systems, the manual offline pre-setting working mode has become very time-consuming and labor-intensive.

[0003] In view of the above problems, the existing research methods to improve work efficiency can be roughly divided into three categories. The first is a static analysis method based on the power flow equation. Combined with the power system power flow balance equation, the active voltage sensitivity, the ratio of load impedance to system impedance, etc. are derived to guide the decision of load shedding point and load shedding amount. This process regards the voltage stability problem as a static problem and does not require system transient process information, but the resulting decision measures are often rough. The second is an automatic search method based on heuristic algorithms. Combined with the transient stability simulation calculation of the power system, considering the system dynamic process, using genetic algorithms, particle swarm algorithms and greedy algorithms to search for effective transient voltage emergency control measures. However, this type of method is difficult to solve the model and requires a large number of transient simulation calculations. When the power system is large, it is time-consuming. The third is an analysis method based on the sensitivity of the transient process control variable. The sensitivity of the load shedding amount to the transient stability margin, the trajectory sensitivity of the load shedding control vector, and the trajectory sensitivity of the node voltage relative to the control variable are used to convert the nonlinear model that is difficult to solve into other models that are easy to solve, and iterate the calculation repeatedly. This method has a simple model solution and high accuracy, but the trajectory sensitivity calculation requires additional transient stability simulation calculations, which is also time-consuming.

[0004] In recent years, with the improvement of computing power, artificial intelligence has made great progress. Deep reinforcement learning and artificial neural networks have been initially applied in voltage stability control. If the manifestations of voltage instability are not distinguished, it is often difficult to obtain effective control measures in some cases. Secondly, when artificial intelligence is applied to the formulation of emergency load shedding control measures for large power grids, there are many optional loads, continuous load shedding, and high-dimensional output variables, which lead to training difficulties. It is difficult to directly obtain effective control measures at one time. How to quickly and effectively formulate emergency load shedding control measures is a technical problem that needs to be solved urgently. Summary of the invention

[0005] In response to the above defects or improvement needs of the prior art, the present invention provides a transient voltage stability emergency control method, device and power control system, which aims to screen effective load nodes in the first stage and determine the amount of control measures in the second stage, thereby solving the technical problem that the existing emergency load shedding method is time-consuming and difficult to apply to the formulation of control measures for large-scale power systems.

[0006] To achieve the above object, according to one aspect of the present invention, a transient voltage stability emergency control method is provided, comprising:

[0007] S1: Based on the power system analysis comprehensive program simulation, the fault state information of the power system is obtained, and the mapping relationship between the fault state information and the voltage instability mode is obtained; the fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode;

[0008] S2: First stage: determining the current voltage instability mode according to the current fault state information in the actual operation of the power system and the mapping relationship; determining the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode;

[0009] S3: The second stage: inputting the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total amount of load shedding; for the current voltage instability mode, emergency control is performed based on the screening sorting and the total amount of load shedding; wherein the LightGBM model pre-establishes the connection between the fault status information and the total amount of load shedding.

[0010] In one embodiment,

[0011] When the current voltage instability mode is a pure voltage instability mode, the screening and sorting is: sorting the load shedding points from low to high according to the load voltage;

[0012] When the current voltage instability mode is the coupled voltage instability mode, the screening and sorting is as follows: sorting the load shedding points from large to small according to the reactive power mutation at the moment of the fault, and sorting the load shedding points from large to small according to the reactive power amplitude during the duration of the fault;

[0013] When the current voltage instability mode is a mixed voltage instability mode, the screening order is: first screen the effective load according to the coupled voltage instability mode, and then convert it to the pure voltage instability mode after removing a certain amount of load, and then screen the effective load according to the pure voltage instability mode.

[0014] In one embodiment, the S3 includes:

[0015] S31: splicing the bus voltage and generator power angle in the current fault state information according to the time dimension to form a hybrid matrix; performing one-dimensional expansion on the hybrid matrix to obtain a one-dimensional vector and inputting it into the LightGBM model so that it outputs the total amount of load shedding;

[0016] S32: For the current voltage imbalance mode, emergency control is performed based on the screening order and the total amount of load shedding.

[0017] In one embodiment, the S32 includes:

[0018] When in the pure voltage instability mode, the total load shedding amount is sequentially allocated to the multiple load shedding points according to the screening order, thereby realizing emergency control;

[0019] When in the coupling voltage instability mode,

[0020] S321: allocating the total load shedding amount to the multiple load shedding points in sequence according to the screening order, and determining initial load shedding control measures;

[0021] S322: verifying the effectiveness of the initial load shedding control measure based on transient stability simulation, adjusting the initial load shedding control measure using the effectiveness, and performing emergency control using the adjusted load shedding control measure;

[0022] Wherein, the mixed voltage instability mode includes the coupled voltage instability mode and the pure voltage instability mode.

[0023] In one embodiment, allocating the total load shedding amount to the multiple load shedding points in sequence according to the screening order comprises:

[0024] The multiple load shedding points are allocated according to the screening and sorting, firstly fully allocating the loads at the front of the sorting, and then screening the loads at the back of the sorting, until the total amount of load shedding is allocated.

[0025] In one embodiment, the S322 includes:

[0026] S322-1: Verify the initial load shedding control measure by using transient stability simulation to obtain the load shedding sensitivity corresponding to the two situations at the moment of the fault and during the duration of the fault;

[0027] S322-2: Select the screening order corresponding to the larger load shedding sensitivity and fine-tune the initial load shedding control measure; use the fine-tuned load shedding control strategy to perform emergency control on the current fault of the power system.

[0028] In one embodiment,

[0029] The calculation formula of the load shedding sensitivity is: Transient Voltage Sensitivity Index Where m is the total number of buses in the power system; T is the transient stability simulation time; t c is the fault clearing time; V i,t is the voltage amplitude of bus i at time t; V i,0 is the bus voltage amplitude before the fault; α is the safety threshold of voltage amplitude deviation; TVSI(0) represents the transient voltage stability when the load is not removed, TVSI(L i ) indicates L i The transient voltage stability after a part of the load is removed during the load sequence, s(L i ) represents the load shedding sensitivity of the i-th load sequence.

[0030] In one embodiment, the S322-2 includes:

[0031] Select the screening order corresponding to the larger load shedding sensitivity to fine-tune the initial load shedding control measure. If the system can recover stability after load shedding, gradually cut back a part of the load until the system is transiently unstable, and take the last stable load shedding measure as the final decision result; if the system cannot recover stability after load shedding, gradually increase the load shedding until the system is transiently stable, and take the last load shedding measure as the final decision result;

[0032] The final decision result is used to perform emergency control on the current fault of the power system.

[0033] According to another aspect of the present invention, there is provided a transient voltage stabilization emergency control device, comprising:

[0034] A simulation module is used to obtain fault state information of the power system based on the power system analysis comprehensive program simulation, and obtain the mapping relationship between the fault state information and the voltage instability mode; the fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode;

[0035] The first stage processing module is used to determine the current voltage instability mode according to the current fault state information in the actual operation of the power system and the mapping relationship; determine the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode;

[0036] The second stage processing module is used to input the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total load shedding; for the current voltage instability mode, emergency control is performed based on the screening sorting and the total load shedding; wherein the LightGBM model pre-establishes a connection between the fault status information and the total load shedding.

[0037] According to another aspect of the present invention, there is provided a power control system, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0038] In general, the above technical solutions conceived by the present invention can achieve the following beneficial effects compared with the prior art:

[0039] Since it takes a long time for large power grids to iterate and try out load shedding control measures, and there are many optional loads, continuous load shedding, and high-dimensional output variables when artificial intelligence is directly applied to decision-making, which leads to training difficulties. A two-stage control method is needed to improve the effectiveness, rapidity and accuracy of offline emergency control measures. Therefore, the present invention proposes a two-stage emergency control decision-making method that integrates artificial intelligence algorithms for the first time. In the first stage, different voltage instability manifestations are analyzed from simulation data, including pure voltage instability, coupled voltage instability and mixed voltage instability, and different effective load screening methods are used to improve the effectiveness of decision-making. In the second stage, a more accurate load shedding amount is estimated from the simulation data based on the LightGBM algorithm, and it is allocated to the sensitive load points of the first stage decision in sorting order, and the load shedding operation is executed. Then, the decision amount is fine-tuned according to whether the system is stable after load shedding, which improves the rapidity and accuracy of decision-making, greatly improves the efficiency of emergency load shedding control measures, and has strong practicality. BRIEF DESCRIPTION OF THE DRAWINGS

[0040] Figure 1 A framework diagram of a transient voltage stability emergency control method provided by an embodiment of the present invention;

[0041] Figure 2 A schematic diagram of a pure voltage instability mode provided by an embodiment of the present invention;

[0042] Figure 3 A schematic diagram of a coupling voltage instability mode provided by an embodiment of the present invention;

[0043] Figure 4 A schematic diagram of a transient voltage severity index provided by an embodiment of the present invention;

[0044] Figure 5 A schematic diagram of a hybrid voltage instability mode provided by an embodiment of the present invention;

[0045] Figure 6 A diagram showing the total initial decision estimate framework of LightGBM provided in one embodiment of the present invention;

[0046] Figure 7 A decision quantity deviation distribution diagram provided by an embodiment of the present invention;

[0047] Figure 8 This is a diagram showing the effect of an emergency load shedding control decision according to an embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to make the purpose, technical solutions and advantages of the present invention more clearly understood, the present invention is further described in detail below in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are only used to explain the present invention and are not intended to limit the present invention. In addition, the technical features involved in the various embodiments of the present invention described below can be combined with each other as long as they do not conflict with each other.

[0049] like Figure 1 As shown, the present invention provides a transient voltage stability emergency control method, comprising:

[0050] S1: Based on the power system analysis comprehensive program simulation, the fault state information of the power system is obtained, and the mapping relationship between the fault state information and the voltage instability mode is obtained; the fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode;

[0051] S2: The first stage: determine the current voltage instability mode according to the current fault state information and mapping relationship in the actual operation of the power system; determine the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode;

[0052] S3: The second stage: input the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total amount of load shedding; for the current voltage instability mode, emergency control is performed based on screening and sorting and the total amount of load shedding; wherein, the LightGBM model pre-establishes the connection between the fault status information and the total amount of load shedding.

[0053] Specifically, a two-stage emergency load shedding control mode is considered. In the first stage, the load shedding point is determined based on simulation data, and in the second stage, the total load shedding amount is determined based on simulation data and the LightGBM algorithm and distributed according to the ranking. In the first stage, three voltage instability modes are distinguished: pure voltage instability, coupled voltage instability, and mixed voltage instability, and different sensitive load screening methods are used for each mode.

[0054] In one embodiment,

[0055] When the current voltage instability mode is the pure voltage instability mode, the screening and sorting are as follows: sorting the load shedding points from low to high according to the load voltage;

[0056] When the current voltage instability mode is the coupled voltage instability mode, the screening and sorting are as follows: at the moment of the fault, the load shedding points are sorted from large to small according to the reactive power mutation amount, and during the duration of the fault, the load shedding points are sorted from large to small according to the reactive power amplitude amount;

[0057] When the current voltage instability mode is a mixed voltage instability mode, the screening order is as follows: first screen the effective load according to the coupled voltage instability mode, and then convert it to a pure voltage instability mode after removing a certain amount of load, and then screen the effective load according to the pure voltage instability mode.

[0058] In one embodiment, S3 includes:

[0059] S31: The bus voltage and generator power angle in the current fault status information are spliced ​​according to the time dimension to form a mixed matrix; the mixed matrix is ​​one-dimensionally expanded to obtain a one-dimensional vector and input into the LightGBM model to make it output the total amount of load shedding;

[0060] S32: For the current voltage imbalance mode, emergency control is performed based on screening sorting and total load shedding.

[0061] In one embodiment, S32 includes:

[0062] When in pure voltage instability mode, the total load shedding amount is distributed to multiple load shedding points in sequence according to the screening order, thereby achieving emergency control;

[0063] When in coupled voltage instability mode,

[0064] S321: allocating the total load shedding amount to multiple load shedding points in sequence according to the screening order, and determining initial load shedding control measures;

[0065] S322: verifying the effectiveness of the initial load shedding control measure based on the transient stability simulation, adjusting the initial load shedding control measure using the effectiveness, and performing emergency control using the adjusted load shedding control measure;

[0066] Among them, the mixed voltage instability mode includes the coupled voltage instability mode and the pure voltage instability mode.

[0067] Specifically, in the pure voltage instability mode, the voltage and power angle are expressed as follows Figure 2 As shown. It can be seen that under this operating condition and fault condition, after the fault is removed, the voltage of some buses continues to drop, and the relative power angles of all generators are stable. At this time, there is no coupling relationship between voltage instability and power angle instability, and the voltage instability in a certain area of ​​the system is caused by the fault. The load screening method is to sort the load voltage from low to high:

[0068]

[0069] In the formula, Represents the load node n sorted from low to high according to the bus voltage.

[0070] In the coupled voltage instability mode, the voltage and power angle are expressed as follows Figure 3 As shown. It can be seen that under this operating condition and fault condition, after the fault is removed, the bus voltage continues to decrease, and then oscillates and becomes unstable. As the bus voltage drops, the power angles of some generators are relatively swung apart, and the power angles become unstable. At this time, the power angle instability is coupled with the voltage instability, and the dominant instability mode is determined to be voltage instability. At this time, two load sequences are considered:

[0071]

[0072] In the formula, It represents the load node n in the sequence 1 obtained by sorting the reactive power change at the moment of fault from large to small; It represents the load node n in the sequence 2 obtained by sorting the magnitude of the load reactive power during the fault duration.

[0073] The indicator TVSI is used to evaluate the transient voltage stability. Its meaning is as follows: Figure 4 shown.

[0074]

[0075]

[0076] Where m is the total number of buses in the system; T is the transient stability simulation time; tc is the fault clearing time. V i,t is the voltage amplitude of bus i at time t; V i,0 is the bus voltage amplitude before the fault; α is the safety threshold of the voltage amplitude deviation, and its value is generally 10%.

[0077] Define load shedding sensitivity:

[0078] In the formula, TVSI(0) represents the transient voltage stability when the load is not removed, and TVSI(L i ) indicates L i The transient voltage stability after a part of the load is removed during the load sequence, s(L i ) represents the load shedding sensitivity of the i-th load sequence. Select the load sequence with large sensitivity.

[0079] In the mixed voltage instability mode, the voltage and power angle are expressed as follows Figure 5 As shown. It can be seen that under such operating conditions and fault conditions, after the fault is removed, the voltage of some buses continues to decrease, and then oscillates and becomes unstable, and the voltage of some buses continues to drop. As the bus voltage drops, the power angle of the generator swings relatively, and the power angle becomes unstable. At this time, the two aforementioned voltage instability manifestations are mixed together. This instability is also a system-level voltage instability. First, the sensitive load sequence is selected according to the method for processing coupled voltage instability, and the loads are cut in order of priority to monitor the changes in the voltage instability manifestation. When the load cut reaches a certain level, the voltage instability manifestation will be converted into pure voltage instability. Then the sensitive load sequence is selected according to the method of pure voltage instability.

[0080] In the second stage, the LightGBM algorithm is used to estimate the total amount of initial decisions. The framework diagram is as follows: Figure 6 First, the voltage and power angle two-dimensional information is sampled and intercepted from the simulation data, and then the voltage matrix and the power angle matrix are spliced ​​according to the time dimension to obtain a mixed matrix. Finally, the mixed matrix is ​​expanded into a one-dimensional vector along the voltage and power angle dimension and input into the LightGBM model.

[0081] It should be noted that since voltage instability and power angle instability are often coupled with each other, the power angle variable and the voltage variable have a complex nonlinear relationship with the estimated total load shedding, so both are used as inputs to the model. Specifically, a set of discrete points of bus voltage and generator power angle in the corresponding observation window is extracted from each set of original measurement data or simulation data, and one point is sampled every 0.01 seconds. After splicing according to the time series dimension, it is expanded into a one-dimensional vector in the form of voltage dimension and power angle dimension for easy processing by LightGBM:

[0082] f=[U 1,1 ,U1,2 …U 1,B …U T,1 ,U T,2 …U T,B δ 1,1 ,δ 1,2 …δ 1,M …δ T,1 ,δ T,2 …δ T,M ]

[0083] Where f is the one-dimensional vector of the input model; U i,j (i=1,2,…,T,j=1,2,…,B) represents the voltage of bus number j at time i; δ i,k (i=1,2,…,T,k=1,2,…,M) represents the power angle value of the kth generator at time i. T is the number of sampling points in the observation window, B is the total number of system buses, and M is the total number of system generators.

[0084] In one embodiment, the total load shedding amount is allocated to multiple load shedding points in sequence according to the screening order, including: allocating multiple load shedding points according to the screening order, first fully allocating the loads at the front of the order, and then screening the loads at the back of the order until the total load shedding amount is allocated.

[0085] In one embodiment, S322 includes:

[0086] S322-1: Use transient stability simulation to verify the initial load shedding control measures and obtain the load shedding sensitivity corresponding to the fault instant and the fault duration;

[0087] S322-2: Select the screening and sorting corresponding to the larger load shedding sensitivity to fine-tune the initial load shedding control measures; use the fine-tuned load shedding control strategy to perform emergency control on the current fault of the power system.

[0088] In one embodiment,

[0089] The calculation formula of load shedding sensitivity is:

[0090] Transient Voltage Sensitivity Index

[0091] Where m is the total number of buses in the power system; T is the transient stability simulation time; t c is the fault clearing time; V i,t is the voltage amplitude of bus i at time t; V i,0 is the bus voltage amplitude before the fault; α is the safety threshold of voltage amplitude deviation; TVSI(0) represents the transient voltage stability when the load is not removed, TVSI(L i) indicates L i The transient voltage stability after a part of the load is removed during the load sequence, s(L i ) represents the load shedding sensitivity of the i-th load sequence.

[0092] In one embodiment, S322-2 includes: selecting the screening and sorting corresponding to the larger load shedding sensitivity and fine-tuning the initial load shedding control measures; if the system can restore stability after the load shedding, gradually shear off a part of the load until the system is transiently unstable, and take the last stable load shedding measure as the final decision result; if the system cannot restore stability after the load shedding, gradually increase the load shedding until the system is transiently stable, and take the last load shedding measure as the final decision result; use the final decision result to perform emergency control on the current fault of the power system.

[0093] Specifically, in the second stage, the total number of decisions estimated by LightGBM is allocated according to the priority order of the load shedding points decided in the first stage, and the effectiveness of the control measures is verified and fine-tuned in combination with transient stability simulation.

[0094] It should be noted that when allocating decision quantities, the loads ranked first are fully allocated first, and then the decision quantities that have not been allocated are allocated to the last load. Combined with transient stability simulation analysis, if the load shedding control measures can restore the system to transient stability, considering that a part of the load may be over-cut, it is necessary to gradually cut back a part of the load until the system loses stability. At this time, the last measure that can restore the system to stability is taken as the final decision measure; and then the decision quantities that have not been allocated are allocated to the last load. Combined with transient stability simulation analysis, if the load shedding control measures cannot restore the system to transient stability, considering that the load shedding amount is not enough, it is necessary to gradually increase the load shedding until the system restores stability. At this time, the last control measure is taken as the final decision measure.

[0095] The following is a specific explanation of the above method using the Northeast Power Grid test system in my country. In the sample set generation stage, a temporary stability simulation is performed on the PSASP software to simulate the voltage data collected by the PMU and WAMS systems. Four typical power flow operation modes are selected; the load adopts a constant impedance parallel induction motor model, in which the proportion of the load motor component is set to 60%; the fault line traverses all 500kV AC lines; the fault location is 2% and 50% of the total length of each line; the fault duration is 0.1s, and 1808 voltage instability samples are generated.

[0096] The comprehensive screening method for distinguishing the manifestations of voltage instability was compared with two methods. The first is the traditional method of screening loads from low to high based on the voltage amplitude 0.5s after the fault, and the second is the method of screening loads from high to low based on the change of load admittance before and after the fault. The overall comparison of different screening methods in all voltage instability samples is shown in Table 1. It can be seen that the proposed comprehensive screening method has a higher overall efficiency of 97.5%; the average number of iterations is less, which is 10.33 times; the average time consumption is also less, which is 299.95s; the average load shedding is less, which is 1003.93MW, indicating that the screened load points are more effective and sensitive.

[0097] Table 1

[0098]

[0099] The deviation distribution diagram of LightGBM's estimated decision volume is as follows Figure 7 As shown, it can be seen that some of the estimated deviation values ​​are close to 0. If the step length is 30% of the load per step, the estimated error can be corrected within 2-3 steps. At the same time, the estimation method of LightGBM is compared with the estimation effect of Convolutional Neural Network (CNN) and Random Forest (RF). Define the mean absolute deviation (MAD) and mean relative deviation (MRD) of all samples:

[0100]

[0101] The comparison results are shown in Table 2. It can be seen that compared with CNN and RF, LightGBM has higher prediction accuracy and faster prediction speed.

[0102] Table 2

[0103]

[0104] Based on the simulation data, effective load shedding points are screened out. After a more accurate load shedding amount is determined based on LightGBM, the determined load shedding amount is allocated to the load shedding points according to the load shedding priority, and the final load shedding amount is fine-tuned in combination with transient stability simulation. Figure 8 As shown in the figure, the system is unstable before load shedding and stable after load shedding. The traversal simulation is performed on 519 test samples and compared with the loop iteration trial. The results are shown in Table 3.

[0105] Table 3

[0106]

[0107] It can be seen from Table 3 that since the estimated decision amount gives the base value of the iteration, the average number of iterations and the average time required are greatly reduced, the decision time is reduced, and the speed of decision-making is increased; fine-tuning is performed based on the estimated decision amount to avoid unnecessary over-cutting loads, and the decision results are more refined.

[0108] According to another aspect of the present invention, there is provided a transient voltage stabilization emergency control device, comprising:

[0109] A simulation module is used to obtain the fault state information of the power system based on the power system analysis comprehensive program simulation, and obtain the mapping relationship between the fault state information and the voltage instability mode; the fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode;

[0110] The first stage processing module is used to determine the current voltage instability mode according to the current fault state information and mapping relationship in the actual operation of the power system; determine the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode;

[0111] The second stage processing module is used to input the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total amount of load shedding; for the current voltage instability mode, emergency control is performed based on screening sorting and the total amount of load shedding; among them, the LightGBM model pre-establishes the connection between the fault status information and the total amount of load shedding.

[0112] According to another aspect of the present invention, there is provided a power control system, comprising a memory and a processor, wherein the memory stores a computer program, and the processor implements the steps of the above method when executing the computer program.

[0113] It will be easily understood by those skilled in the art that the above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. Any modifications, equivalent substitutions and improvements made within the spirit and principles of the present invention should be included in the protection scope of the present invention.

Claims

1. A transient voltage stability emergency control method, characterized in that: include: S1: obtaining fault status information of the power system based on the simulation of the power system analysis comprehensive program, and obtaining a mapping relationship between the fault status information and the voltage instability mode; The fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode; S2: First stage: determining the current voltage instability mode according to the current fault state information in the actual operation of the power system and the mapping relationship; determining the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode; S3: The second stage: inputting the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total amount of load shedding; for the current voltage instability mode, emergency control is performed based on the screening sorting and the total amount of load shedding; wherein the LightGBM model pre-establishes the connection between the fault status information and the total amount of load shedding.

2. The transient voltage stability emergency control method according to claim 1, characterized in that: When the current voltage instability mode is a pure voltage instability mode, the screening and sorting is: sorting the load shedding points from low to high according to the load voltage; When the current voltage instability mode is the coupled voltage instability mode, the screening and sorting is as follows: sorting the load shedding points from large to small according to the reactive power mutation at the moment of the fault, and sorting the load shedding points from large to small according to the reactive power amplitude during the duration of the fault; When the current voltage instability mode is a mixed voltage instability mode, the screening order is: first screen the effective load according to the coupled voltage instability mode, and then convert it to the pure voltage instability mode after removing a certain amount of load, and then screen the effective load according to the pure voltage instability mode.

3. The transient voltage stability emergency control method according to claim 2, characterized in that: The S3 includes: S31: splicing the bus voltage and generator power angle in the current fault state information according to the time dimension to form a hybrid matrix; performing one-dimensional expansion on the hybrid matrix to obtain a one-dimensional vector and inputting it into the LightGBM model so that it outputs the total amount of load shedding; S32: For the current voltage imbalance mode, emergency control is performed based on the screening order and the total amount of load shedding.

4. The transient voltage stability emergency control method according to claim 3, characterized in that: The S32 includes: When in the pure voltage instability mode, the total load shedding amount is sequentially allocated to the multiple load shedding points according to the screening order, thereby realizing emergency control; When in the coupling voltage instability mode, S321: allocating the total load shedding amount to the multiple load shedding points in sequence according to the screening order, and determining initial load shedding control measures; S322: verifying the effectiveness of the initial load shedding control measure based on transient stability simulation, adjusting the initial load shedding control measure using the effectiveness, and performing emergency control using the adjusted load shedding control measure; Wherein, the mixed voltage instability mode includes the coupled voltage instability mode and the pure voltage instability mode.

5. The transient voltage stability emergency control method according to claim 4, characterized in that: The step of allocating the total load shedding amount to the plurality of load shedding points in sequence according to the screening order comprises: The multiple load shedding points are allocated according to the screening and sorting, firstly fully allocating the loads at the front of the sorting, and then screening the loads at the back of the sorting, until the total amount of load shedding is allocated.

6. The transient voltage stability emergency control method according to claim 4, characterized in that: The S322 includes: S322-1: Verify the initial load shedding control measure by using transient stability simulation to obtain the load shedding sensitivity corresponding to the two situations at the moment of the fault and during the duration of the fault; S322-2: Select the screening order corresponding to the larger load shedding sensitivity and fine-tune the initial load shedding control measure; use the fine-tuned load shedding control strategy to perform emergency control on the current fault of the power system.

7. The transient voltage stability emergency control method according to claim 6, characterized in that: The calculation formula of the load shedding sensitivity is: Transient Voltage Sensitivity Index Where m is the total number of buses in the power system; T is the transient stability simulation time; t c is the fault clearing time; V i,t is the voltage amplitude of bus i at time t; V i,0 is the bus voltage amplitude before the fault; α is the safety threshold of voltage amplitude deviation; TVSI(0) represents the transient voltage stability when the load is not removed, TVSI(L i ) indicates L i The transient voltage stability after a part of the load is removed during the load sequence, s(L i ) represents the load shedding sensitivity of the i-th load sequence.

8. The transient voltage stability emergency control method according to claim 6, characterized in that: The S322-2 includes: Select the screening order corresponding to the larger load shedding sensitivity to fine-tune the initial load shedding control measure. If the system can recover stability after load shedding, gradually cut back a part of the load until the system is transiently unstable, and take the last stable load shedding measure as the final decision result; if the system cannot recover stability after load shedding, gradually increase the load shedding until the system is transiently stable, and take the last load shedding measure as the final decision result; The final decision result is used to perform emergency control on the current fault of the power system.

9. A transient voltage stability emergency control device, characterized in that: include: A simulation module, used to obtain fault status information of the power system based on the power system analysis comprehensive program simulation, and obtain a mapping relationship between the fault status information and the voltage instability mode; The fault state information includes: bus voltage and generator power angle; the voltage instability mode includes: pure voltage instability mode, coupled voltage instability mode and mixed voltage instability mode; The first stage processing module is used to determine the current voltage instability mode according to the current fault state information in the actual operation of the power system and the mapping relationship; determine the screening and sorting corresponding to multiple sensitive load shedding points according to the current voltage instability mode; The second stage processing module is used to input the current fault status information into the lightweight gradient boosting machine LightGBM model to estimate the total load shedding; for the current voltage instability mode, emergency control is performed based on the screening sorting and the total load shedding; wherein the LightGBM model pre-establishes a connection between the fault status information and the total load shedding.

10. A power control system, comprising a memory and a processor, wherein the memory stores a computer program, characterized in that: When the processor executes the computer program, the steps of the method according to any one of claims 1 to 8 are implemented.

Images