Frequency stabilization generator tripping and load shedding method and system capable of minimizing cost

Abstract

The invention discloses a cost-minimizing frequency stabilization generator tripping and load shedding method and system, and relates to the technical field of power system safety and stability control, and the method comprises the steps: carrying out the partitioning through a first algorithm, building a fault prediction set, and carrying out the screening through a first analysis method; based on the screening result of the first analysis method, constructing a strategy library of serious faults and a minimum cost optimization model; and solving through a first optimization algorithm, and obtaining a stable control generator tripping load shedding auxiliary service strategy based on a first analysis method. According to the method, the power grid is partitioned, the strategy library of serious faults and the minimum cost optimization model are constructed, the generator tripping and load shedding range is reasonably limited through partitioning, the frequency of the power grid is rapidly stabilized, the cost of the generator tripping and load shedding strategy is minimized by applying the particle swarm algorithm, the stable control of the frequency of the power grid can be more intelligent and refined, and the stability of the power grid is improved. The problems of compensation cost optimization and power grid operation efficiency in power grid frequency stability control under large interference are solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power system security and stability control, and particularly to a method and system for minimizing the cost of frequency-stable generator tripping and load shedding. Background Art

[0002] With the access of large-scale new energy and high-voltage DC transmission in the power system, the impact of unbalanced power suffered by the system has increased significantly. At the same time, the access of a large number of power electronic devices has led to a decline in the system inertia level and frequency regulation ability, thus weakening the frequency support ability of the system. The frequency stability of the power grid is directly related to the safe and stable operation of the power system. When a serious fault occurs in the power grid, such as a DC bipolar blocking fault, the frequencies of the sending-end and receiving-end AC power grids of the system may fluctuate greatly, which may lead to the collapse of the power grid and large-scale power outages.

[0003] Under the current situation, the first line of defense of the power grid cannot meet the adequacy and security requirements under the normal state of the power system, forcing the safety and stability control system in the second line of defense to take emergency control measures to ensure the stable operation of the power grid. With the development of the power market and the progress of smart grid technology, the requirements for the frequency response and optimal control of the power grid are continuously increasing. However, when dealing with the problem of power grid frequency stability, the traditional generator tripping and load shedding control methods do not fully consider the optimization of compensation costs. For the operation of the power system, reducing compensation costs is an important way to improve economic benefits. Therefore, it is necessary to study the method for obtaining grid market-oriented stability control auxiliary services to improve the frequency stability and operation efficiency of the power grid under the condition of minimizing compensation costs for emergency control. Summary of the Invention

[0004] In view of the above existing problems, the present invention is proposed.

[0005] Therefore, the technical problem solved by the present invention is: the existing power grid frequency response and optimal control methods have insufficient optimization of compensation costs, are not timely in response to frequency fluctuations, lack effective emergency control measures, and how to improve the frequency stability and operation efficiency of the power grid under the condition of minimizing compensation costs.

[0006] To solve the above technical problems, the present invention provides the following technical solution: A method for minimizing the cost of frequency-stable generator tripping and load shedding, including partitioning through a first algorithm, establishing a fault prediction set, and screening through a first analysis method; constructing a strategy library for severe faults and a minimum-cost optimization model based on the screening results of the first analysis method; solving through a first optimization algorithm, and obtaining a stability control generator tripping and load shedding auxiliary service strategy based on the first analysis method.

[0007] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the partitioning by the first algorithm includes calculating the power grid partitioning index and partitioning the power grid by the first algorithm.

[0008] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the establishment of the fault prediction set and the screening by the first analysis method include establishing the fault prediction set based on historical data and screening the severe faults in the fault prediction set by the first analysis method.

[0009] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the construction of the strategy library for severe faults includes the severe faults screened by the first analysis method and constructing the stable generator tripping and load shedding strategy according to the power grid partitioning result.

[0010] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the minimum cost optimization model includes constructing the minimum cost optimization model for the stable generator tripping and load shedding strategy.

[0011] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the solution by the first optimization algorithm includes solving the minimum cost optimization model by the first optimization algorithm.

[0012] As a preferred embodiment of the method for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: the obtaining of the generator tripping and load shedding auxiliary service strategy for stability control based on the first analysis method includes obtaining the perturbed frequency trajectory based on the first analysis method, calculating the fitness, and obtaining the generator tripping and load shedding auxiliary service strategy for stability control.

[0013] Another object of the present invention is to provide a system for minimizing cost of frequency-stable generator tripping and load shedding, which can construct a strategy library for severe faults and a minimum cost optimization model based on the screening results of the first analysis method, and solves the problem of insufficient compensation cost optimization in the current power grid frequency response and optimization control technology.

[0014] As a preferred embodiment of the system for minimizing cost of frequency-stable generator tripping and load shedding according to the present invention, wherein: it includes a partitioning and fault prediction module, a strategy library construction module, and an optimization solution module; the partitioning and fault prediction module is used for partitioning by the first algorithm, establishing the fault prediction set, and screening by the first analysis method; the strategy library construction module is used for constructing the strategy library for severe faults and the minimum cost optimization model based on the screening results of the first analysis method; the optimization solution module is used for solving by the first optimization algorithm and obtaining the generator tripping and load shedding auxiliary service strategy for stability control based on the first analysis method.

[0015] A computer device comprises a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of a method for frequency stabilization with minimum cost and load shedding.

[0016] A computer-readable storage medium stores a computer program, which, when executed by a processor, implements the steps of a method for frequency stabilization with machine shedding and load shedding that minimizes cost.

[0017] Beneficial effects of the present invention: The frequency stability cutting and load shedding method with minimized cost provided by the present invention uses a spectral clustering algorithm to partition the power grid, establishes a set of expected faults of severe power shortage (surplus) of the power grid, performs time domain simulation analysis on each fault one by one, calculates indicators such as frequency change rate, minimum frequency value, quasi-steady-state frequency deviation, and screens out severe frequency instability fault sets. For each severe frequency instability fault, a set of alternative generators and loads corresponding to its stable cutting and load shedding strategies are formed according to the power grid partitions. For each severe frequency instability fault, compensation cost models and parameters of alternative generators and loads are collected to form a total amount constraint of stable cutting and load shedding, and the corresponding minimum stable cutting and load shedding compensation cost optimization model is constructed respectively. For each severe frequency instability fault, a particle swarm algorithm is used. The problem of minimum compensation cost optimization for stable control cutting of generators and loads is solved, the disturbed frequency trajectory is obtained based on time domain simulation calculation, the fitness is calculated according to the frequency safety index, and the auxiliary service strategy for stable control cutting of generators and loads is obtained. The present invention considers the optimization of frequency stable control cutting of generators and loads for each severe frequency instability fault with minimum compensation cost under power grid partitioning, and reasonably limits the range of cutting of generators and loads by partitioning to achieve rapid stabilization of power grid frequency. The particle swarm algorithm is used to minimize the cost of the cutting of generators and loads strategy. Compared with the pre-set rules of traditional cutting of generators and loads, the stable control of power grid frequency can be made more intelligent and refined, and the problems of compensation cost optimization and power grid operation efficiency in stable control of power grid frequency under large interference are solved. The present invention achieves better results in terms of adaptability, economy and reliability. BRIEF DESCRIPTION OF THE DRAWINGS

[0018] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative work.

[0019] Figure 1 An overall flow chart of a frequency stabilization machine and load shedding method with minimized cost provided in the first embodiment of the present invention.

[0020] Figure 2Flowchart of a particle swarm algorithm for a method of minimizing cost and stabilizing frequency by load shedding and generator tripping provided in the first embodiment of the present invention.

[0021] Figure 3 Technical flowchart of a method of minimizing cost and stabilizing frequency by load shedding and generator tripping provided in the second embodiment of the present invention.

[0022] Figure 4 Schematic diagram of modules of a system for minimizing cost and stabilizing frequency by load shedding and generator tripping provided in the third embodiment of the present invention. Detailed implementation manners

[0023] To make the above objects, features, and advantages of the present invention more obvious and understandable, the following detailed description of the specific implementation manners of the present invention will be given in conjunction with the accompanying drawings of the specification. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all embodiments. All other embodiments obtained by those of ordinary skill in the art based on the embodiments of the present invention without creative efforts shall fall within the scope of protection of the present invention.

[0024] Embodiment 1. Refer to Figure 1 - Figure 2 , which is an embodiment of the present invention, and provides a method of minimizing cost and stabilizing frequency by load shedding and generator tripping, including:

[0025] S1: Partition through a first algorithm, establish a fault prediction set, and screen through a first analysis method.

[0026] Furthermore, partitioning through the first algorithm includes calculating a power grid partitioning index and partitioning the power grid through the first algorithm.

[0027] Calculating the power grid partitioning index includes, but is not limited to, calculating the electrical distance as the power grid partitioning index. The specific steps are as follows:

[0028] Set the sending-end power grid graph in the power grid as G Ι (V Ι , E Ι ), the receiving-end power grid graph as G ΙΙ (V ΙΙ , E ΙΙ ), where it includes vertex sets V Ι 、V ΙΙ and edge sets E Ι 、e ΙΙ , and there are n Ι 、n ΙΙ nodes as vertices in the corresponding graphs. Each edge in the edge set can be represented by E Ι (i, j) and E ΙΙ (i, j), representing the power line between node i and node j.

[0029] Taking the sending-end power grid diagram G Ι (V Ι ,E Ι ) as an example, the coupling degree between nodes is reflected by sensitivity, and the electrical distance is measured, expressed as:

[0030]

[0031] where α ij is the electrical sensitivity of the active source node i to the controlled node j, Δδ i is the phase angle change of node i, Δδ j is the phase angle change of node j, and ΔP j is the active power change of node j.

[0032] In the PQ decomposition method, the relationship between active power and phase angle is expressed as:

[0033] ΔP = B'Δδ

[0034] where ΔP is the node active power vector, B' is the system matrix, and Δδ is the node phase angle vector.

[0035] The sensitivity is solved by the successive recursion method, and the electrical sensitivity of node i to the other nodes is calculated, expressed as:

[0036]

[0037] b = [0, L, 0, 1, 0, L, 0] T

[0038] where ΔP i is the active power change of node i, b is a vector with only the i-th component being 1 and the other components being 0, L is the active source, and T is the transpose operation.

[0039] Each controlled node is mapped to the node where the active source is located, and the electrical sensitivity d ij of the active source node i to the controlled node j is obtained, expressed as:

[0040] d ij = -lg|α ij |

[0041] An electrical distance matrix D is established. Each active source node is used as one dimension of the matrix, and the electrical distance M of each active source node to each controlled node is used as the component in each dimension, expressed as:

[0042]

[0043] where D is an m×n matrix, m is the number of controlled nodes, and n is the number of active source nodes.

[0044] The electrical distance M between any two controlled nodes i and j ij , is expressed as:

[0045]

[0046] The power grid is partitioned by the first algorithm, where the first algorithm can be a spectral clustering algorithm, a density-based spatial clustering algorithm, or other methods capable of partitioning the power grid.

[0047] It should be noted that in the embodiment of the present application, when partitioning the power grid by the first algorithm, the spectral clustering algorithm is adopted, and the specific steps are as follows:

[0048] Using k-means clustering to partition the sending-end power grid includes:

[0049] A1. Determine the number K of clusters into which the data set is to be divided, that is, divide the power grid into K clusters, and the value range is determined by the scale of the active power source of the power grid, K = 2, 3,..., L. For each cluster, randomly select 1 data as the initial cluster center, denoted as c.

[0050] A2. Input the data set of the sending-end power grid, that is, the electrical distance M ij , which is denoted as x i . For each data in the data set, calculate its distance Q from each cluster center using the Euclidean distance, expressed as:

[0051]

[0052] where Ω j is the data set of cluster j, x i (j) is the i-th data in the j-th cluster, and c j is the j-th cluster center.

[0053] After the calculation is completed, assign each data in the cluster to the nearest cluster center to form a preliminary cluster.

[0054] A3. Update the position of the cluster center corresponding to the cluster. The new cluster center is the mean value of all data in the cluster:

[0055]

[0056] where N j is the total number of data in cluster j.

[0057] A4. Repeat steps A2 and A3 until the cluster assignment of all data does not change, and obtain the partition Γ of the sending-end power grid Ι .

[0058] The receiving-end grid partition Γ is calculated by the same steps: ΙΙ .

[0059] The result of partitioning the power grid is the sending-end power grid partition Γ Ι and receiving-end grid partition Γ ΙΙ .

[0060] It should be noted that, in an optional embodiment, the power grid is partitioned by a first algorithm, using a density-based spatial clustering algorithm, and the specific steps are:

[0061] First, in the actual scenario of power grid partitioning, it is critical to determine the appropriate parameter settings. For example, the radius (ε) is set as the upper limit of the electrical distance to ensure that clustering only considers nodes that are close to each other, which can effectively avoid the introduction of irrelevant data due to the long distance. At the same time, the minimum number of neighbors (minPts) is selected, which is usually taken as 10% to 20% of the maximum number of nodes in the data set, in order to ensure that there is sufficient data support in each cluster.

[0062] Next, in the data preprocessing stage, by collecting the electrical sensitivity and load status of each node in the power grid, a complete data set is constructed. When performing neighborhood queries, the neighbor set of each node can be quickly found based on efficient spatial indexes (such as quadtrees or KD trees). If the number of neighbors of a node exceeds minPts, it is marked as a core point and its importance in the power grid is identified, which usually means that the node has a strong influence on other nodes.

[0063] After the core points are determined, the cluster expansion process begins, and all density-reachable points are included in the same cluster by recursively checking the neighbors of each core point. This process will continue until no new points can be added. At the same time, non-core points adjacent to the core points are marked as boundary points in order to understand their relative position and impact in the power network.

[0064] After this series of processing, the points that cannot be classified will be identified as noise, thereby eliminating the interference of irrelevant data. Finally, after completing the zoning of the power grid, the areas with heavy load or affected by faults can be clearly identified.

[0065] Furthermore, establishing a fault prediction set and screening it by the first analysis method includes establishing a fault prediction set based on historical data and screening severe faults in the fault prediction set by the first analysis method.

[0066] Establishing a fault prediction set based on historical data includes establishing a grid severe power shortage (surplus) expected fault set Φ, which is expressed as:

[0067] Φ={φ1,…,φ n' ,…,φ N'}

[0068] wherein, φ n' is the n'-th severe power deficit (surplus) fault in the power grid, and N' is the total number of faults.

[0069] The severe faults in the fault prediction set are screened by the first analysis method. The first analysis method can be time-domain simulation analysis, fault prediction and screening based on machine learning, or other methods that can screen severe faults in the fault prediction set.

[0070] It should be noted that in the embodiment of the present application, when screening the severe faults in the fault prediction set by the first analysis method and using time-domain simulation analysis, the specific steps are as follows:

[0071] Perform time-domain simulation calculations on each fault in the pre-fault set to obtain the frequency perturbed trajectory, calculate its frequency stability index, determine the severity of frequency stability, and record the severe frequency instability faults.

[0072] The frequency stability index includes the rate of change of frequency RoCoF, the maximum value of transient frequency deviation, and the quasi-steady state frequency deviation.

[0073] B1. Define the rate of change of frequency RoCoF constraint, which is expressed as:

[0074]

[0075] wherein, f RoCoF and f RoCoF,max are respectively the actual value of the frequency change after the power grid is perturbed and the maximum allowable frequency change limit specified for the power grid.

[0076] B2. Define the maximum value of transient frequency deviation constraint, which is expressed as:

[0077] Δf TFD,min ≤Δf max ≤Δf TFD,max

[0078] wherein, Δf max is the maximum value of transient frequency deviation, Δf TFD,min and Δf TFD,max are the lower and upper limits of the transient frequency deviation specified for the power grid.

[0079] B3. Define the quasi-steady state frequency deviation constraint, which is expressed as:

[0080] Δf ss,min ≤Δf ss ≤Δf ss,max

[0081] wherein, Δf ss is the quasi-steady state frequency deviation, and the quasi-steady state frequency deviation Δfss is an index for measuring transient frequency stability, Δf ss,min and Δf ss,max are the lower and upper limits of the quasi-steady state frequency deviation specified for the power grid.

[0082] Screen severe frequency instability faults and construct a set Φ of severe frequency instability faults f , which is expressed as:

[0083] Φ f ={φ f,1 ,…, φ f,n' ,…, φ f,N'}

[0084] where φ f,n' is the n'-th severe frequency instability fault in the set of severe frequency instability faults.

[0085] It should be noted that in an alternative embodiment, severe faults in the fault prediction set are screened by the first analysis method, and fault prediction and screening based on machine learning are adopted. The specific steps are as follows:

[0086] First, construct a feature set using historical fault data and operation data, including but not limited to power flow data, historical fault types, frequency fluctuations, load changes, and environmental factors (such as temperature, humidity, etc.). Then, select a suitable machine learning algorithm, such as random forest, support vector machine (SVM), or long short-term memory network (LSTM), and divide it into a training set and a test set for model training.

[0087] Next, adjust the model parameters through cross-validation to improve the prediction accuracy and generalization ability. Then, perform feature engineering processing on the historical data, extract features that have a greater impact on fault prediction, such as frequency change trends, power deficits, etc., and construct a feature vector describing the operation state of the power grid.

[0088] After the model is established, the faults in the contingency set are input into the model for real-time status evaluation. The model will output the prediction probability of each fault according to the characteristics of the newly input data, and screen the faults with a prediction probability higher than the set threshold. This process can improve the accuracy of fault screening through continuous learning and updating, enabling grid managers to identify potential severe faults faster, and then take targeted preventive and response measures.

[0089] Finally, through subsequent analysis of the screened severe faults, a new fault mode library can be formed to further enrich and optimize the model, enabling it to more accurately reflect the actual operation state of the power grid in future predictions.

[0090] S2: Based on the screening results of the first analysis method, construct a strategy library for severe faults and a minimum cost optimization model.

[0091] Furthermore, constructing a severe fault strategy library includes severe faults screened based on the first analysis method, and constructing a stable generator tripping and load shedding strategy according to the power grid partition results.

[0092] Constructing a stable generator tripping and load shedding strategy includes, for each severe frequency instability fault, forming corresponding alternative generator and load sets for its stable generator tripping and load shedding strategy according to the power grid partition. The specific steps are as follows:

[0093] From the severe frequency instability fault set Φ f select, in sequence, all severe frequency faults φ f,n' that do not meet the requirements of the frequency stability index, and according to the sending-end power grid partition Γ Ι and the receiving-end power grid partition Γ ΙΙ results, form an alternative generator set Ω g,n' and an alternative load set Ω l,n' for its stable generator tripping and load shedding strategy.

[0094] It should be noted that the minimum cost optimization model includes the minimum cost optimization model for constructing a stable generator tripping and load shedding strategy. The specific steps are as follows:

[0095] For each severe frequency instability fault, collect the compensation cost models and parameters of the alternative generators and loads, form the total amount constraint of the stable control generator tripping and load shedding, and respectively construct its corresponding minimum stable control generator tripping and load shedding compensation cost optimization model.

[0096] Collect the compensation cost models and parameters of the alternative generator set Ω f,n' and the alternative load set Ω g,n' in φ l,n' to form the total amount constraint of the stable control generator tripping and load shedding, and construct the minimum stable control generator tripping and load shedding compensation cost optimization model with the minimum stable control generator tripping and load shedding compensation cost F as the optimization goal, expressed as:

[0097]

[0098] where c g,i , c l,j are respectively the unit compensation costs for the removal of the conventional unit i and the load j, P g,i , P l,j are respectively the powers of the conventional unit i in the alternative generator set Ω g,n' and the load j in the alternative load set Ω l,n' , u g,i , u l,j are respectively the conventional unit i in the alternative generator set Ω g,n' and the alternative load set Ω l,n'The removal identification variable for medium load j, 1 indicates participation, 0 indicates non - participation, N G and N L are respectively the set of candidate generators Ω g,n' the conventional units in it, the set of candidate loads Ω l,n' the total number of loads in it.

[0099] Define the stability - control action quantity constraint, expressed as:

[0100]

[0101] Among them, are respectively the stability - control action quantities of the conventional units and loads, that is, the active power change amounts after the power grid fault.

[0102] The system frequency safety constraint refers to Step B1, Step B2 and Step B3.

[0103] S3: Solve through the first - order optimization algorithm, and obtain the stability - control generator - tripping and load - shedding auxiliary service strategy based on the first - order analysis method.

[0104] Furthermore, solving through the first - order optimization algorithm includes solving the minimum - cost optimization model through the first - order optimization algorithm.

[0105] It should be noted that obtaining the stability - control generator - tripping and load - shedding auxiliary service strategy based on the first - order analysis method includes obtaining the disturbed frequency trajectory based on the first - order analysis method, calculating the fitness, and obtaining the stability - control generator - tripping and load - shedding auxiliary service strategy.

[0106] It should also be noted that solving through the first - order optimization algorithm includes, but is not limited to, using the particle - swarm algorithm to solve its minimum stability - control generator - tripping and load - shedding compensation cost optimization problem, and obtaining the stability - control generator - tripping and load - shedding auxiliary service strategy based on the first - order analysis method includes, but is not limited to, using time - domain simulation software to obtain the disturbed frequency trajectory, calculating the fitness, and obtaining the stability - control generator - tripping and load - shedding auxiliary service strategy. The specific steps are:

[0107] For φ f,n' The constructed minimum stability - control generator - tripping and load - shedding compensation cost optimization model is iteratively calculated based on time - domain simulation using the particle - swarm optimization algorithm to obtain the optimal strategy. The process of the particle - swarm algorithm is as Figure 2 shown, and the specific steps are:

[0108] C1. Initialize the particle - swarm parameters: Set the number of particles as N P , the dimension of the binary vector of each particle is N G +N L , the maximum number of iterations is H, the inertia weight is ω, and the acceleration constants are a1, a2.

[0109] C2. Form an initial population that satisfies the stability control action quantity constraint: Initialize the particle positions. For each particle k (generator and load shedding scheme), randomly generate a binary vector as the particle position, where each element is initialized to 0 or 1 with a probability of 50%. 0 indicates not shedding the generator or load, and 1 indicates shedding the generator or load, and eliminate the particles that violate the stability control action quantity constraint; Initialize the particle velocities, and set the velocity range to a random number between [-1, 1]; Initialize the individual best position of each particle That is its initial position; Initialize the global best position g best , which is the particle position that makes the objective function value optimal among all particles.

[0110] C3. Calculate the particle fitness and update the individual best and global best positions: For each particle k, perform a time-domain simulation of the corresponding fault to obtain the perturbed frequency trajectory, and calculate the fitness F Pe , which is expressed as:

[0111] F Pe = F + G RoCoF + G TFD + G ss

[0112]

[0113] where G RoCoF , G TFD , G ss are the penalty functions of RoCoF, the maximum transient frequency deviation, and the quasi-steady state frequency deviation respectively, and ξ RoCoF , ξ TFD , ξ ss are the penalty coefficients of RoCoF, the maximum transient frequency deviation, and the quasi-steady state frequency deviation respectively.

[0114] Take the fitness value of the current particle as the individual best position of the current particle k t is the number of iterations; Find the particle with the minimum fitness among all particles, and take its position as the global best position g best .

[0115] C4. Update the particle velocity and position: For each particle k, calculate the new velocity according to the velocity update formula, which is expressed as:

[0116]

[0117] where are the velocities of particle k at the (t + 1)-th and t-th iterations respectively, r1 and r2 are random numbers between [0, 1], is the position of particle k at the t-th iteration.

[0118] According to the position update formula, the calculated new position is expressed as:

[0119]

[0120] where is the position of particle k at the (t + 1)-th iteration.

[0121] C5. Determine whether the convergence condition is satisfied: whether the maximum number of iterations is reached or the global best position remains unchanged in multiple iterations. If the convergence condition is satisfied, end the loop and output the final global best position g best , which is the optimal auxiliary service strategy for generator tripping and load shedding, and at the same time output the corresponding minimized cost F value; if the convergence condition is not satisfied, repeat steps C3 and C4.

[0122] Discriminate the severe frequency instability fault set Φ f for each severe frequency instability fault φ f,n' in it to check whether the optimal auxiliary service strategy for generator tripping and load shedding for stability control is obtained. If the optimal strategies are all obtained, output the optimal auxiliary service strategy for generator tripping and load shedding for stability control; otherwise, repeat steps S2 and S3 until the optimal auxiliary service strategies for generator tripping and load shedding for stability control are obtained for all frequency instability faults.

[0123] Example 2. Referring to Figure 3 , which is an embodiment of the present invention, provides a method for minimizing the cost of frequency stability by generator tripping and load shedding. In order to verify the beneficial effects of the present invention, scientific demonstrations are carried out through economic benefit calculations and simulation experiments.

[0124] The technical process of this embodiment is as shown in Figure 3 . First, calculate the electrical distances between different nodes, and use the k-means clustering algorithm to partition the power grid to facilitate classifying similar nodes. Then, set the initial set of contingency faults and analyze the frequency trajectories of each fault to identify the frequency critical faults that have obvious impacts on the nodes, ensuring that the frequency trajectories of each fault in the set of contingency faults are all analyzed. Next, initialize the frequency instability faults. Then, according to the power grid partition, form the alternative generator and load sets for each frequency instability fault, and construct an optimization model for minimizing the compensation cost of generator tripping and load shedding for stability control. At the same time, use the particle swarm algorithm to improve the optimization efficiency of load regulation and solve the optimization problem of the compensation cost of generator tripping and load shedding for stability control, ensuring that the optimal auxiliary service strategies for generator tripping and load shedding for stability control are obtained for each frequency instability fault in the frequency instability fault set, and finally output the auxiliary service strategy for generator tripping and load shedding for stability control of the system.

[0125] Specifically, this embodiment involves the partitioning of the sending and receiving end power grids, the establishment of a fault prediction set, fault screening, and the optimization of a stability control strategy. First, a typical power grid is selected for analysis, such as the power supply network in a certain area, which contains 10 nodes. The sending end power grid is partitioned by the spectral clustering algorithm, with the number of clusters set to 2. The partitioning index of the power grid is calculated based on the electrical distance, and the electrical distance matrix is obtained by calculating the sensitivity between nodes. Subsequently, historical fault data is selected to construct a contingency set, and the data includes records of severe power deficit (surplus) faults that have occurred in the past. The fault set is defined as Φ, which contains the fault data that has occurred in history.

[0126] Secondly, time-domain simulations are performed on the recorded faults to calculate frequency stability indices, including the rate of change of frequency RoCoF, the maximum transient frequency deviation, and the quasi-steady state frequency deviation, to determine severe instability faults. For each fault, corresponding thresholds are set, and the corresponding index values are collected to form a severe frequency instability fault set Φ f 。

[0127] In this process, the validity and accuracy of the data for each step are ensured. Comparing the set frequency stability indices with historical data helps to more accurately predict impending faults and formulate corresponding generator and load shedding strategies to ensure the stable operation of the power grid.

[0128] The data records and analysis during the experiment are summarized into a table to show the frequency stability indices and fault prediction situations corresponding to different fault types for further analysis of the system performance.

[0129] Refer to Table 1 for a comparative analysis of the experimental data.

[0130] Table 1 Experimental data record table

[0131]

[0132] Through the data analysis in Table 1, first of all, by monitoring the rate of change of frequency (RoCoF), transient frequency deviation, and quasi-steady state frequency deviation of each fault type, it can be clearly seen that severe instability fault points can be screened out in a timely manner. When all frequency and stability indices are lower than the specified thresholds, the power grid will be in a safe area, and this setting effectively warns of future faults in the historical records.

[0133] Specifically, the RoCoF values of power deficit fault 1, fault 2, power surplus fault 1, and fault 2 are all higher than 0.5 Hz / s, indicating that they perform poorly in terms of large load fluctuations and affecting the power grid stability and belong to severe faults that need attention and treatment. Compared with the past treatment methods, this method makes the fault screening work more transparent and scientific by introducing the rate of change of frequency and maximum deviation criteria.

[0134] In terms of fault handling strategies, by combining the results of power grid zoning, strategies for generator tripping and load control are flexibly formulated, bringing obvious response capabilities. For example, for power deficit faults 1 and 2, timely implementation of the generator tripping strategy can effectively reduce the line pressure and avoid fault propagation over a larger area. According to historical data, this timely dynamic adjustment greatly reduces the risk of fault spread compared with static fault handling methods, ensuring the reliability of the power grid. Therefore, the present invention is creative.

[0135] Example 3, refer to Figure 4 , which is an embodiment of the present invention, provides a frequency-stable generator tripping and load shedding system for minimizing costs, including a zoning and fault prediction module, a strategy library construction module, and an optimization solution module.

[0136] Among them, the zoning and fault prediction module is used to perform zoning through the first algorithm, establish a fault prediction set, and screen through the first analysis method; the strategy library construction module is used to construct a strategy library for severe faults and a minimum cost optimization model based on the screening results of the first analysis method; the optimization solution module is used to solve through the first optimization algorithm and obtain a stable control generator tripping and load shedding auxiliary service strategy based on the first analysis method.

[0137] If the function is implemented in the form of a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on such an understanding, the technical solution of the present invention, in essence, or the part that contributes to the prior art, or a part of this technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions for causing a computer device (which can be a personal computer, a server, or a network device, etc.) to execute all or part of the steps of the methods of the various embodiments of the present invention. The aforementioned storage medium includes: various media such as USB flash drives, mobile hard disks, read-only memories (ROM, Read-Only Memory), random access memories (RAM, Random Access Memory), magnetic disks, or optical discs that can store program codes.

[0138] The logic and / or steps represented in the flowchart or otherwise described herein, for example, can be considered as a definable sequence list of executable instructions for implementing logical functions, and can be specifically implemented in any computer-readable medium for use by an instruction execution system, apparatus, or device (such as a computer-based system, a system including a processor, or other systems that can fetch and execute instructions from the instruction execution system, apparatus, or device), or used in conjunction with these instruction execution systems, apparatus, or devices. For the purposes of this specification, a "computer-readable medium" can be any device that can contain, store, communicate, propagate, or transport a program for use by or in conjunction with an instruction execution system, apparatus, or device.

[0139] More specific examples (non-exhaustive list) of computer-readable media include the following: an electrical connection portion (electronic device) having one or more wirings, a portable computer diskette (magnetic device), a random access memory (RAM), a read-only memory (ROM), an erasable programmable read-only memory (EPROM or flash memory), an optical fiber device, and a portable compact disc read-only memory (CDROM). Additionally, a computer-readable medium can even be paper or other suitable media on which a program can be printed, because the program can be obtained electronically, for example, by optically scanning the paper or other media, followed by editing, interpretation, or otherwise processing as appropriate, and then stored in a computer memory.

[0140] It should be understood that the various parts of the present invention can be implemented by hardware, software, firmware, or a combination thereof. In the above-described embodiments, multiple steps or methods can be implemented by software or firmware stored in a memory and executed by a suitable instruction execution system. For example, if implemented by hardware, as in another embodiment, any one or a combination of the following techniques well known in the art can be used: discrete logic circuits having logic gate circuits for implementing logical functions on data signals, application-specific integrated circuits having appropriate combinational logic gate circuits, programmable gate arrays (PGAs), field programmable gate arrays (FPGAs), etc. It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

[0141] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention rather than to limit them. Although the present invention has been described in detail with reference to the preferred embodiments, those of ordinary skill in the art should understand that the technical solutions of the present invention can be modified or equivalently replaced without departing from the spirit and scope of the technical solutions of the present invention, and they should all be covered within the scope of the claims of the present invention.

Claims

1. A method for minimizing the cost of frequency-stable generator tripping and load shedding, characterized in that, Including: Partition by the first algorithm, establish a fault prediction set, and screen by the first analysis method; Based on the screening results of the first analysis method, construct a policy library for severe faults and a minimum-cost optimization model; Solve by the first optimization algorithm, and obtain the auxiliary service strategy for stabilizing control of generator tripping and load shedding based on the first analysis method.

2. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 1, wherein: The partition by the first algorithm includes calculating the power grid partition index and partitioning the power grid by the first algorithm.

3. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 2, characterized in that: The establishment of the fault prediction set and the screening by the first analysis method include establishing a fault prediction set based on historical data and screening severe faults in the fault prediction set by the first analysis method.

4. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 3, characterized in that: The construction of the policy library for severe faults includes severe faults screened by the first analysis method, and constructing a strategy for stabilizing generator tripping and load shedding according to the power grid partition result.

5. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 4, characterized in that: The minimum-cost optimization model includes constructing a minimum-cost optimization model for the strategy of stabilizing generator tripping and load shedding.

6. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 5, characterized in that: The solution by the first optimization algorithm includes solving the minimum-cost optimization model by the first optimization algorithm.

7. The method for minimizing the cost of frequency-stable generator tripping and load shedding according to claim 6, wherein: The obtaining of the auxiliary service strategy for stabilizing control of generator tripping and load shedding based on the first analysis method includes obtaining the disturbed frequency trajectory based on the first analysis method, calculating the fitness, and obtaining the auxiliary service strategy for stabilizing control of generator tripping and load shedding.

8. A system adopting the method for minimizing the cost of frequency-stable load shedding by load shedding with a breaker as described in any one of claims 1 to 7, characterized in that: Including a partition and fault prediction module, a policy library construction module, and an optimization solution module; The partition and fault prediction module is used to partition by the first algorithm, establish a fault prediction set, and screen by the first analysis method; The policy library construction module is used to construct a policy library for severe faults and a minimum-cost optimization model based on the screening results of the first analysis method; The optimization solution module is used to solve by the first optimization algorithm and obtain the auxiliary service strategy for stabilizing control of generator tripping and load shedding based on the first analysis method.

9. A computer device, comprising a memory and a processor, the memory storing a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for stabilizing frequency by generator tripping and load shedding with minimized cost according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for stabilizing frequency by generator tripping and load shedding with minimized cost according to any one of claims 1 to 7.

Images