A method for active defense of smart grid cascading failures

Abstract

The application discloses a kind of active defense methods of smart grid cascading failure, comprising the following steps: step 1) before disaster occurs, system obtains the disaster distribution data of affected smart grid;Step 2) system is based on the failure probability of disaster distribution data calculation grid node or grid element, the capacity reduction amount of the failure effect equivalence of grid node or element is calculated;Step 3) system is according to capacity reduction amount predictively promotes the capacity of smart grid node or element;Step 4) when disaster occurs, system obtains the disaster data of smart grid node and judges whether grid node fails;Step 5) system is according to whether the failure of smart grid node or grid element executes the cascading failure defense of grid node or grid element;The beneficial effects of the application: can effectively active defense to grid cascading failure.

Description

Technical Field

[0001] This invention relates to the field of smart grid safe operation technology, and specifically to a proactive defense method for smart grid cascading faults. Background Technology

[0002] Extreme natural disasters often cause widespread, clustered faults in power transmission lines, leading to major power outages. This has prompted governments and scholars worldwide to increasingly emphasize the role of smart grids in mitigating disaster-causing faults. In fact, under extreme natural disasters, once certain components or lines in the power grid trigger cascading faults, their propagation speed is extremely rapid, potentially causing system disconnection and widespread blackouts within a very short time. Therefore, it is necessary to develop a system for mitigating smart grid faults under natural disasters to reduce power outage losses caused by such events.

[0003] In most current research on cascading failure prevention in coupled networks, the propagation of faults in coupled nodes is often handled in a simplistic way: if one end node of a coupled link fails, the other end node of that link is assumed to also fail. However, in smart grids, cascading failure propagation is often not so straightforward. For example, the failure of a power node does not necessarily cause the failure of its coupled communication node (because the communication node is coupled to other power nodes); similarly, the failure of a communication node does not necessarily cause the failure of its coupled power node (because the power node is also coupled to other communication nodes, through which it can transmit control commands or status information). Furthermore, in coupled network scenarios like smart grids, once a cascading failure occurs at either end of a coupled link, although it can propagate to the other side of the network, the resulting impacts are not necessarily the same. In fact, when one end node of a coupled link fails, the entire coupled link also fails, leading to the other coupled end node losing the functional support of the coupled link and propagating the impact of the failure to the network on the side where that end node is located. Formally, this impact is equivalent to reducing the processing capacity (i.e., capacity) of that end node. The extent to which node capacity is reduced is closely related to the severity of the failure. In fact, if the current load exceeds the node's capacity after capacity is reduced, the node will fail due to overload; otherwise, the node will continue to function normally. Summary of the Invention

[0004] The technical problem this invention aims to solve is how to construct the equivalence of node capacity reduction under the interdependent coupling effect between the power grid and the power communication network due to disaster-causing faults; based on capacity reduction, how to improve the predictability of differentiated node capacity to proactively defend against the impact of disaster-causing faults caused by natural disasters; and when deviations occur in the accuracy of capacity improvement due to enhanced predictability, leading to ineffective proactive defense, how to allow grid nodes to shed loads to reduce the adverse consequences of load shedding while maintaining the stable and reliable operation of the grid. The purpose of this invention is to provide a proactive defense method for cascading faults in smart grids, solving the above problems.

[0005] This invention is achieved through the following technical solution:

[0006] A proactive defense method for cascading faults in smart grids includes the following steps:

[0007] Step 1): Before a disaster occurs, the system acquires various disaster distribution data, such as the types and severity of disasters affecting the smart grid, the frequency of the same disaster occurring in the same geographical area, and the functional / life failure cycles of various grid components.

[0008] Step 2): The system calculates the failure probability of power grid nodes or power grid components based on disaster distribution data, and calculates the capacity reduction amount of the fault effect equivalence of power grid nodes or components based on the failure probability.

[0009] Step 3): The system proactively increases the capacity of smart grid nodes or components based on the capacity reduction amount;

[0010] Step 4): When a disaster occurs, the system acquires disaster data of smart grid nodes or grid components and determines whether the grid nodes or grid components have failed.

[0011] Step 5): The system executes cascading fault defense for the smart grid node or grid component based on whether the node or component has failed.

[0012] If a power grid node or power grid component fails, the system will proactively disconnect part of the overloaded power grid node or power grid component and distribute the remaining load of the failed node to surrounding nodes to perform cascading failure defense of the power grid node or power grid component.

[0013] If the grid node or grid component does not fail, the cascading failure defense for the grid node or grid component ends.

[0014] Specifically, in step 1), the disaster probability distribution of the smart grid node or grid element is obtained, and the disaster probability distribution of the grid node or grid element is calculated according to formula (1-1):

[0015] P = f1(x1, x2, ... x) n}(1-1)

[0016] Where, x i These represent various factors that influence the probability of disasters at power grid nodes or components.

[0017] In step 2), the failure probability of smart grid nodes under natural disasters is quantified by equivalent quantification with the impact effect of smart grid node failures. This quantifies the impact effect of each smart grid node failure under natural disasters, and calculates the capacity reduction of smart grid nodes or grid components whose disaster-causing failures can be quantified. Based on the equivalent quantification method of failure impact, the capacity reduction ratio of grid nodes can be calculated according to formula (2-2):

[0018] K i =f2(P i s1, s2...s n (2-2)

[0019] Among them, K i P represents the rate of capacity decay at power grid nodes. i Let s represent the probability distribution of disaster at power grid node i. i This represents the set of topological and electrical characteristic parameters of power grid node i.

[0020] Furthermore, based on the quantification of the aforementioned equivalence, the amount of capacity reduction when smart grid nodes encounter disasters is predicted, thereby proactively increasing the capacity of each smart grid node affected by the disaster to resist capacity reduction and reduce the transmission of cascading failures in smart grid nodes.

[0021] In step 2), under the disaster-causing effect, to quantify the magnitude of the disaster-causing effect on each node of the smart grid, the fault impact effect on one smart grid node is equivalent to a reduction in the capacity of that node. The amount by which the capacity of each node is reduced is closely related to the magnitude of the fault impact effect. The more severe the fault effect, the greater the impact on the grid node. Therefore, under the equivalent fault impact effect, the capacity reduction of each grid node after fault equivalence is calculated according to formula (3-3):

[0022] C′=C×(1-K) (3-3)

[0023] Where C is the capacity of the grid node before failure, C′ is the capacity of the grid node after reduction in the equivalent row, and K is the reduction ratio. Different capacity reduction methods correspond to different reduction ratios K. If the reduction is based on the topology parameters or electrical characteristic parameters of the grid node, the reduction ratio K is determined by the following formula (4-4):

[0024]

[0025] Where d1 is the result of the topology or electrical characteristic parameters of the coupled grid node after minimum-maximum normalization, and a is the equivalent reduction adjustment parameter.

[0026] In step 3), the predictability improvement of the grid node capacity is used to reduce the impact of disaster-causing faults. Specifically, the calculation formula for the predictability improvement of the grid node capacity affected by disaster-causing faults is (5-5):

[0027] C1=Cλ=(1+α)L0λ (5-5)

[0028] Where C is the initial capacity of the grid node that may be affected by a catastrophic fault, L0 is the initial load of the grid node, α represents the tolerance coefficient, and C1 represents the equivalent capacity of the grid node after being affected by a catastrophic fault. The grid node will only be in normal working condition when its capacity is greater than its load. To ensure that the grid node still works normally under the influence of natural disasters, λ needs to satisfy formulas (6-6) and (7-7):

[0029] (1+α)L0λ×K>L (6-6)

[0030] Right now:

[0031]

[0032] Where L represents the current actual load of the grid node, and K is the reduction ratio. The more important the grid node, the more capacity needs to be added. To consider costs, a capacity increase threshold λ is set. c When λ > λ c When, let λ = λ c .

[0033] Furthermore, the predictive improvement ratio of the grid node capacity can be calculated according to formula (8-8):

[0034] λ i =f3(K i ,α) (8-8)

[0035] Where, λ i It is the percentage increase in node capacity, K i It is the node capacity decay ratio calculated based on the equivalence quantification method, where α is the network tolerance coefficient.

[0036] In step 5), to prevent the failed power grid node from causing other normal power grid nodes to become overloaded due to increased load during the load transfer process, thus generating a series of cascading faults, a load shedding method is set for the overloaded power grid nodes during disasters to reduce the impact of cascading faults on power nodes and reduce the transmission of cascading faults in power grid nodes.

[0037] Furthermore, the load shedding is based on the current node failure status, transferring the load after load shedding to other grid nodes to prevent the fault from continuing to iterate and affecting more grid nodes to the greatest extent. The load shedding is determined based on the number of iterations of the current fault, the degree distribution of the grid nodes, and the failure probability of the grid nodes under the disaster. The more iterations and the higher the failure probability, the less the remaining load of the grid node. The greater the connectivity of the grid node, the more remaining load.

[0038] The load shedding reduces the widespread propagation of cascading faults. The amount of load shedding depends on the number of iterations β of the cascading fault within the power grid, the grid node connectivity k, and the grid node failure probability P. The more iterations and the higher the failure probability, the less load remains at the grid node. The higher the grid node connectivity, the more load remains. The load shedding ratio can be calculated according to formula (9-9):

[0039]

[0040] in, It is the proportion of load cut off by grid node i.

[0041] Compared with the prior art, the present invention has the following advantages and beneficial effects:

[0042] 1. This invention takes into account the equivalence between the impact of faults and the reduction of node capacity under natural disasters. Before the disaster occurs, the capacity of each node in the power grid is increased in a differentiated manner to reduce the impact of natural disasters. The equivalence quantification model for the impact of faults caused by natural disasters is designed as follows: the model quantifies the equivalence between the probability of node failure and the effect of node faults under natural disasters. That is, the impact of faults caused by power grid nodes under extreme weather is equivalent to the reduction of the capacity of faulty nodes, thereby quantifying the magnitude of the impact of faults on each node in the power grid under natural disasters.

[0043] 2. This invention improves the capacity of smart grid nodes or grid components through pre-disaster prediction: Before a natural disaster occurs, the capacity of grid nodes needs to be increased in a targeted manner according to the capacity reduction ratio of the grid nodes in order to reduce the impact of the natural disaster. The amount of capacity increase mainly depends on the reduction ratio K of the node capacity, and is also affected by the network cost (tolerance coefficient α). The more the capacity is reduced (the larger the reduction ratio) and the smaller the tolerance coefficient of the grid node, the more capacity prediction improvement is required.

[0044] 3. The load shedding mechanism of this invention needs to transfer the remaining load after shedding the current grid node to other grid nodes based on the failure status of the current grid node (the load allocation ratio depends on the conservative capacity ratio of other nodes after equivalent quantification) in order to prevent the fault from continuing to iterate and affecting more grid nodes to the greatest extent. Therefore, the load shedding scheme needs to be determined based on the number of iterations of the current fault, the degree distribution of the node, and the failure probability of the node under the disaster. The more iterations and the higher the failure probability, the less the remaining load of the node. The higher the node connectivity, the more the remaining load, so as to reduce the adverse consequences of load shedding and maintain the stable and reliable operation of the power grid.

[0045] 4. The present invention designs a proactive pre-disaster defense method to prevent the impact of natural disaster-induced failures: Based on an equivalence quantification scheme for the impact of disaster-induced failures, the impact of extreme disasters on power grid nodes can be quantified as a reduction in the capacity of the power grid nodes. Simultaneously, the power grid node may fail due to real-time load exceeding the reduced capacity, leading to the propagation of cascading failures. Therefore, before an extreme natural disaster occurs, the amount of capacity reduction when a node encounters a disaster can be predicted based on the equivalence quantification scheme. This allows for preventative capacity increases in affected nodes to resist capacity reduction and reduce the propagation of cascading failures. Attached Figure Description

[0046] To more clearly illustrate the technical solutions of the exemplary embodiments of the present invention, the accompanying drawings used in the embodiments will be briefly described below. It should be understood that the following drawings only show some embodiments of the present invention and should not be considered as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort. In the drawings:

[0047] Figure 1 This is a flowchart of the method of the present invention;

[0048] Figure 2 This is a diagram showing the one-to-one matching and coupling of the information network and the power grid in this invention;

[0049] Figure 3 This is a diagram showing the one-to-one heterogeneous coupling between the information network and the power grid in this invention;

[0050] Figure 4 This is a diagram showing the co-coupling of the information network and the power grid under a coupling strength of 0.7 according to the present invention.

[0051] Figure 5 This is a diagram showing the heterogeneous coupling of the information network and the power grid under a coupling strength of 0.7 according to the present invention.

[0052] Figure 6 This diagram illustrates the active defense effect of the present invention. Detailed Implementation

[0053] To make the objectives, technical solutions, and advantages of the present invention clearer, the present invention will be further described in detail below with reference to the embodiments and accompanying drawings. The illustrative embodiments and descriptions of the present invention are only used to explain the present invention and are not intended to limit the present invention.

[0054] Example 1

[0055] Current research on cascading fault defense in coupled networks typically assumes that if a smart grid node at one end of a coupled link fails, the smart grid node at the other end of the link will also fail. This approach fails to consider the existence of redundant power supply and communication nodes in the network. Furthermore, in coupled network scenarios like smart grids, while cascading faults at both ends of the coupled link can propagate to the other side of the network, the resulting impacts may not be identical. Therefore, this simplistic approach cannot accurately characterize fault propagation in power information networks. In addition, traditional methods for increasing grid node capacity rarely consider the differences between grid nodes, leading to resource waste by uniformly increasing capacity for all nodes. This invention differentiates nodes, reducing the impact of cascading faults while saving network investment costs.

[0056] like Figure 1 As shown, this invention proposes an active defense method for cascading faults in smart grids based on a dependent coupled network model and load distribution strategy. Specifically, it proposes an analysis method for the propagation impact of cascading faults in coupled nodes based on the equivalence quantification of fault impacts, and thereby proposes an active defense method for cascading faults that combines predictive increase of node capacity with proactive load shedding.

[0057] This invention addresses the equivalence between the impact of faults and node capacity reduction under natural disasters by differentially increasing the capacity of each node in the power grid before a disaster occurs, thereby mitigating the impact of natural disasters. Simultaneously, when proactive capacity enhancement defense strategies fail to prevent node failures, in-disaster load shedding is used to reduce the propagation of cascading faults. The invention includes the following steps:

[0058] For the design of an equivalent quantitative model of the impact of natural disaster-induced faults: Under the influence of extreme natural disasters, power grid nodes have a certain probability of failure due to damage; in addition, the failure of nodes in the power information network will also lead to the loss of all or part of the functions of the power grid nodes at the other end of the coupled link. In order to quantify the magnitude of the damage caused by natural disasters to each node in the power grid, this invention designs a model that quantifies the equivalent quantitative relationship between the probability of node failure under natural disasters and the effect of node failure. That is, the impact of the disaster-induced faults of power grid nodes under extreme weather is equivalent to the reduction of the capacity of the faulty node, thereby quantifying the magnitude of the fault effect of each node in the power grid under natural disasters.

[0059] Design of proactive pre-disaster defense methods to prevent the impact of natural disaster-induced failures: Based on the equivalence quantification scheme of the impact of disaster-induced failures, the impact of extreme disasters on power grid nodes can be quantified as the capacity reduction of the nodes; at the same time, the power grid node may fail due to the real-time load exceeding the reduced capacity, leading to the propagation of cascading failures; therefore, before the occurrence of extreme natural disasters, the amount of capacity reduction of nodes when encountering disasters can be predicted according to the equivalence quantification scheme, thereby preventing the capacity reduction of nodes affected by natural disasters from being increased, in order to resist the capacity reduction and reduce the propagation of cascading failures.

[0060] Design of a load shedding strategy to reduce the impact of natural disaster-induced failures: Due to the bias in prediction accuracy and the limitations of network costs, when extreme disasters occur, if the proactive defense strategy fails to prevent node failure due to faults, under the effect of the equivalent quantification scheme, the impact of natural disasters on power grid nodes will be equivalent to a reduction in node capacity, leading to the failure of some nodes due to overload. In order to prevent the failed nodes from causing other normal nodes to become overloaded due to increased load during the load transfer process, thereby generating a series of cascading failures, it is necessary to set appropriate load shedding methods for overloaded nodes in the power grid during disasters, so as to reduce the impact of cascading failures on power nodes, reduce the transmission of cascading failures, and improve the robustness of the power grid.

[0061] Example 2

[0062] like Figure 1 As shown, a proactive defense method for cascading faults in a smart grid includes the following steps:

[0063] Step 1): Before a disaster occurs, the system acquires various disaster distribution data, such as the types and severity of disasters affecting the smart grid, the frequency of the same disaster occurring in the same geographical area, and the functional / life failure cycles of various grid components.

[0064] Step 2): The system calculates the failure probability of power grid nodes or power grid components based on disaster distribution data, and calculates the capacity reduction amount of the fault effect equivalence of power grid nodes or components based on the failure probability.

[0065] Step 3): The system proactively increases the capacity of smart grid nodes or components based on the capacity reduction amount;

[0066] Step 4): When a disaster occurs, the system acquires disaster data of smart grid nodes or grid components and determines whether the grid nodes or grid components have failed.

[0067] Step 5): The system executes cascading fault defense for the smart grid node or grid component based on whether the node or component has failed.

[0068] If a power grid node or power grid component fails, the system will proactively disconnect part of the overloaded power grid node or power grid component and distribute the remaining load of the failed node to surrounding nodes to perform cascading failure defense of the power grid node or power grid component.

[0069] If the grid node or grid component does not fail, the cascading failure defense for the grid node or grid component ends.

[0070] The smart grid is simulated using the IEEE 39-node system, which includes 10 generators, 39 nodes, 19 load points, and 46 lines. The information network is a scale-free network with 39 nodes and a one-to-one coupling method.

[0071] Example 3

[0072] 1. Design of an equivalent quantitative model for the impact of disaster-induced failures

[0073] In power cyber-physical systems, power nodes and information nodes influence each other. On the one hand, the failure of an information node weakens the grid's dispatching capability; on the other hand, the failure of a power node also affects information nodes due to the lack of direct power supply nodes. When a power node fails, it needs to transmit messages and provide corresponding feedback through communication nodes. However, in coupled network scenarios like smart grids, when cascading failures occur at both ends of the coupled link, although the fault can propagate to the other side of the network, the resulting impact may not be the same.

[0074] Under extreme natural disasters, to quantify the magnitude of the disaster's impact on each node of the power grid, the effect of a fault on one node is considered equivalent to a reduction in its processing capacity (i.e., capacity). The extent of the capacity reduction at a node is closely related to the magnitude of the fault's impact. Generally, the more severe the fault, the greater the impact on the node. Therefore, considering the equivalent effect of the fault's impact, the capacity reduction of the node after the fault is equivalent is calculated as shown in formula (1-1):

[0075] C′=C×(1-K) (1-1)

[0076] Where C is the capacity of the grid node before failure, C′ is the capacity of the grid node after equivalent reduction, and K is the reduction ratio. Different capacity reduction methods correspond to different reduction ratios K. Here, we consider reducing the capacity according to the node's topology parameters (e.g., degree value of information nodes) or electrical characteristic parameters (e.g., electrical betweenness of power nodes). The reduction ratio K is then determined by formula (1-2):

[0077]

[0078] Where d1 is the result of the topology or electrical characteristic parameters of the coupled grid node after minimum-maximum normalization, and a is the equivalent reduction adjustment parameter.

[0079] 2. Design of pre-disaster capacity prediction strategies for nodes susceptible to disaster-causing failures.

[0080] Under extreme natural disaster conditions, power grid nodes in the same area have roughly the same failure probability. Considering the equivalence between the impact of faults under natural disasters and the reduction of power grid node capacity, this invention proposes an active defense scheme based on a differentiated node capacity predictability enhancement strategy to reduce the impact of disaster-causing faults caused by natural disasters.

[0081] Specifically, the formula for calculating the predictive capacity improvement of nodes that may be affected by catastrophic failures is (2-1):

[0082] C1=Cλ=(1+α)L0λ (2-1)

[0083] Where C represents the initial capacity of the node potentially affected by a catastrophic failure, L0 represents the initial load of the node, α represents the tolerance coefficient, and C1 represents its capacity after equivalent reduction under the influence of a catastrophic failure. Since a node will only be in normal working condition when its capacity is greater than its load, λ needs to satisfy formulas (2-2) and (2-3) to ensure that the node can still function normally under the influence of natural disasters:

[0084] (1+α)L0λ×K>L (2-2)

[0085] Right now:

[0086]

[0087] Here, L represents the current actual load of the node. Therefore, the more important the grid node, the more capacity needs to be added. At the same time, to consider costs, a capacity increase threshold λ is set. c When λ > λ c When, let λ = λ c .

[0088] 3. Design of proactive load shedding strategies for power nodes during disasters to reduce the impact of cascading failures.

[0089] Due to limitations in prediction accuracy and network costs, proactive defense strategies that enhance capacity through anticipation cannot completely prevent grid node failures. Therefore, when natural disasters occur, there are two reasons for grid node failures: the disaster itself can cause the node to fail directly, or the failure of the power information network node at the other end of the coupling link can cause the power node to fail. After a grid node fails, in order to maintain stable generator output, the load of the failed node must be borne by surrounding nodes. This increases the likelihood that the real-time load of surrounding grid nodes will exceed their maximum load capacity, which may lead to a series of cascading failures and seriously affect the safe and stable operation of the power grid.

[0090] During natural disasters, load shedding is necessary to prevent large-scale power outages. A reasonable and effective load shedding strategy must reduce the adverse consequences of load shedding while maintaining the stable and reliable operation of the power grid. Most existing load shedding mechanisms are based on linear programming to find the optimal solution. Due to the cumbersome solution process, this invention proposes a load shedding scheme to reduce the impact of cascading failures. After a node fails due to overload but before load transfer, to reduce the propagation of cascading faults, a portion of the load on the power grid node needs to be shedding. The load shedding strategy needs to transfer the load after shedding to other power grid nodes based on the current failure status of the node, in order to minimize the impact of the fault on more power grid nodes. Therefore, the load shedding scheme needs to be determined based on the number of iterations of the current fault, the degree distribution of the node, and the failure probability of the node under the disaster. The more iterations and the higher the failure probability, the less remaining load the node has; conversely, the higher the node connectivity, the more remaining load it has.

[0091] 4. Implementation of proactive defense methods against cascading failures caused by natural disasters.

[0092] Based on the above analysis, the specific method for proactively preventing cascading failures caused by natural disasters, as described in this patent, is as follows:

[0093] Step 1): Obtain the disaster probability distribution of smart grid nodes or grid components: The disaster probability of a node or component depends on various factors, such as the type and intensity of the disaster, the frequency of the same disaster occurring in the same geographical area, and the functional / lifetime failure cycle of each grid component. Therefore, the disaster probability distribution of a grid node or grid component can be expressed in the form of formula (4-1):

[0094] P = f1(x1, x2, ... x) n}(4-1)

[0095] Where, x i This refers to various factors that affect the probability of a node or component causing a disaster.

[0096] Step 2): Calculate the capacity reduction of smart grid nodes or grid components affected by quantifiable disaster-causing faults: Based on the equivalence quantification method of fault impact, the capacity reduction of grid nodes mainly depends on the fault probability distribution of the node, the node's topological parameters (such as degree, betweenness), or electrical characteristic parameters (such as power, voltage, current, etc.). Furthermore, the higher the fault probability and the more important the structure of the node, the greater the capacity reduction. Therefore, the capacity reduction ratio of node i can be expressed in the form of formula (4-2):

[0097] K i =f2(P i s1, s2...s N (4-2)

[0098] Among them, K i P represents the rate of capacity decay of a node. i Let s represent the probability distribution of node i causing disaster. i This represents the set of topological and electrical characteristic parameters of node i.

[0099] Step 3): Preemptive capacity enhancement of smart grid nodes or grid components before natural disasters: Before a natural disaster occurs, the capacity of grid nodes needs to be increased in a targeted manner based on the capacity reduction ratio of the nodes to reduce the impact of the natural disaster. The amount of capacity increase mainly depends on the node capacity reduction ratio K, and is also affected by network costs (tolerance coefficient α). The greater the capacity reduction (the larger the reduction ratio) and the smaller the tolerance coefficient of the node, the greater the required capacity increase. Therefore, the preemptive capacity enhancement ratio of grid nodes can be expressed in the form of equation (4-3):

[0100] λ i =f3(K i ,α)(4-3)

[0101] Where, λ i It is the percentage increase in node capacity, K i It is the node capacity decay ratio calculated based on the equivalence quantification method, where α is the network tolerance coefficient.

[0102] Step 4): Proactive load shedding of smart grid nodes or components during disasters: During natural disasters, load shedding strategies reduce the widespread propagation of cascading faults. The amount of load shedding depends on the number of iterations β of the cascading fault within the grid, the node connectivity (number of neighboring nodes), and the node failure probability. The more iterations and the higher the failure probability, the less load remains at the node; conversely, the higher the node connectivity, the more load remains. Therefore, the load shedding ratio can be expressed as formula (4-4):

[0103]

[0104] in, β is the proportion of load cut off at node i, β is the number of iterations of the cascading fault within the power grid, and k is the load shedding rate at node i. i It represents the degree distribution of the current node.

[0105] 5. Feasibility verification of the method and its implementation effect

[0106] The complex interconnections of the power information network conform to the general characteristics of complex networks. Based on complex network theory, the system can be abstracted as a set of points and edges, where edges include connecting edges and dependent edges. For the load transfer mechanism, this invention, drawing on the classic ML model, considers the mutual influence between the network states on each side of the coupled network when a cascading failure occurs. Based on the betweenness distribution of the end grid nodes in the coupled link, the betweenness of the grid nodes and the active power consumed by the nodes in the power information network and the power grid are respectively taken as the load. When a node fails due to a natural disaster, the capacity of the failed node drops to 0, and all its load needs to be distributed to neighboring nodes. The distribution ratio depends on the conservative capacity of the node after equivalence quantification. After receiving the load distributed by the failed node, the real-time load of the neighboring node is equal to the sum of its previous load and the additional load it received. When the load of a neighboring node exceeds its capacity, a new round of load redistribution will occur until there are no overloaded nodes.

[0107] Furthermore, to measure a network's ability to withstand cascading failures under attack, a resilience metric needs to be defined. This invention considers the failure of node i, which initially has the highest load, and assumes that the total number of remaining nodes affected by node i is N. f Then the resilience of a network is defined as N. f The ratio to the total number of nodes N in the network is shown in formula (5-1):

[0108]

[0109] The smaller the S-value, the better the network's resilience.

[0110] 5.1 Feasibility Verification of the Method for Quantifying the Equivalence of Disaster-Causing Fault Impacts

[0111] Considering that coupled power grid nodes have a certain probability of failure under different extreme natural disasters, this invention quantifies the magnitude of natural disaster damage to each node of the power grid by reducing the capacity of the faulty node, and verifies the feasibility of this equivalent method through simulation analysis.

[0112] This section compares the resilience indices of fault equivalent schemes and probabilistic failure schemes under different coupling methods. Figure 2 The coupling method is a one-to-one matching coupling between the information network and the power grid. Figure 3 The coupling method is a one-to-one heterogeneous coupling between the information network and the power grid. Figure 4The coupling method is that the information network and the power grid are coupled in the same manner under the condition of coupling strength of 0.7. Figure 5 The coupling method is a heterogeneous coupling between the information network and the power grid under a coupling strength of 0.7. The coupling strength is used to adjust the ratio of coupled edges; a higher coupling strength results in more coupled edges. Specifically, when the coupling strength is 1, the network is a one-to-one coupling. Simulation results show that the proportion of remaining nodes is roughly the same in the equivalent cases of probabilistic failure of coupled nodes and faulty coupled nodes. Under the same coupling method, attacking the same node causes relatively consistent damage to the power information network in both methods. Since the two methods can achieve roughly similar effects, the fault equivalence method can be considered to replace the handling method for probabilistic failure of coupled nodes under a certain natural disaster.

[0113] 5.2 Analysis of the effectiveness of proactive defense against cascading disaster failures

[0114] Considering the equivalence between the impact of faults and node capacity reduction under natural disasters, this patent proposes to differentiate the capacity of each node in the power grid before a disaster occurs, based on the amount of capacity reduction at the time of a fault, in order to reduce the impact of natural disasters on the power grid; if proactive defense fails to prevent node failure, a load shedding strategy is adopted during the disaster to reduce the propagation of cascading faults. Simulation comparison experiments are designed and the results are as follows:

[0115] Figure 6 With a fixed tolerance coefficient α = 0.4, the comparison charts before and after network hardening are obtained by changing the attacking power grid node each time. Figure 5 The image shows a comparison of the network before and after hardening, with the network's tolerance coefficient changed for each of the 39 nodes in a fixed attack on the power grid. This demonstrates that the proposed scheme can significantly reduce the impact of cascading failures, while also providing different hardening schemes for different power grid nodes, thus saving network resources.

[0116] The specific embodiments described above further illustrate the purpose, technical solution, and beneficial effects of the present invention. It should be understood that the above description is only a specific embodiment of the present invention and is not intended to limit the scope of protection of the present invention. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the scope of protection of the present invention.

Claims

1. A proactive defense method for cascading faults in smart grids, characterized in that, Includes the following steps: Step 1): Before a disaster occurs, the system acquires the types and severity of disasters affecting the smart grid, the frequency of the same disaster occurring in the same geographical area, and various disaster distribution data of the functional / life failure cycles of each power grid component. Step 2): Calculate the failure probability of power grid nodes or power grid components based on disaster distribution data, and calculate the capacity reduction amount of the fault effect equivalence of power grid nodes or power grid components based on the failure probability. Step 3): The system proactively increases the capacity of smart grid nodes or grid components based on the capacity reduction amount; wherein, the proactive increase in the capacity of the grid nodes is to reduce the impact of disaster-causing faults, specifically, the calculation formula for the proactive capacity increase of grid nodes affected by disaster-causing faults is (5-5): （5-5）； in, The initial capacity of grid nodes that may be affected by catastrophic faults. The initial load of the power grid node. Represents the tolerance coefficient. This represents the equivalent capacity reduction of a power grid node under the influence of a catastrophic fault. A power grid node will only operate normally when its capacity exceeds its load. This is to ensure that power grid nodes continue to function normally under the influence of natural disasters. Formulas (6-6) and (7-7) must be satisfied: （6-6）； Right now: （7-7）； Where L represents the current actual load of the grid node, and K is the reduction ratio. The more important the grid node, the more capacity needs to be added. To consider costs, a capacity increase threshold is set. ,when season ; Step 4): When a disaster occurs, the system acquires disaster data of smart grid nodes or grid components and determines whether the grid nodes or grid components have failed. Step 5): The system executes cascading fault defense for the smart grid node or grid component based on whether the node or component has failed. If a power grid node or power grid component fails, the system will proactively disconnect part of the overloaded power grid node or power grid component and distribute the remaining load of the failed node to surrounding nodes to perform cascading failure defense of the power grid node or power grid component. If the grid node or grid component does not fail, the cascading failure defense for the grid node or grid component ends.

2. The active defense method for cascading faults in a smart grid according to claim 1, characterized in that, In step 1), the disaster probability distribution of the smart grid node or grid element is obtained, and the disaster probability distribution of the grid node or grid element is calculated according to formula (1-1): （1-1）； in, These represent various factors that influence the probability of disasters occurring at power grid nodes or components.

3. The active defense method for cascading faults in a smart grid according to claim 1, characterized in that, In step 2), the failure probability of smart grid nodes under natural disasters is quantified by equivalent quantification with the impact effect of smart grid node failures. This quantifies the impact effect of each smart grid node failure under natural disasters, and calculates the capacity reduction of smart grid nodes or grid components whose disaster-causing failures can be quantified. Based on the equivalent quantification method of failure impact, the capacity reduction ratio of grid nodes can be calculated according to formula (2-2): （2-2）； in, This indicates the rate of capacity decay at power grid nodes. Represents a power grid node The probability distribution of disasters, Represents a power grid node A set of topological and electrical characteristic parameters.

4. The active defense method for cascading faults in a smart grid according to claim 3, characterized in that, Based on the quantification of the aforementioned equivalence, the amount of capacity reduction of smart grid nodes when they encounter disasters is predicted, thereby preventively increasing the capacity of each smart grid node affected by the disaster to resist the capacity reduction and reduce the transmission of cascading failures of smart grid nodes.

5. The active defense method for cascading faults in a smart grid according to claim 3, characterized in that, In step 2), under the disaster-causing effect, to quantify the magnitude of the disaster-causing effect on each node of the smart grid, the fault impact effect on one smart grid node is equivalent to a reduction in the capacity of that node. The amount by which the capacity of each node is reduced is closely related to the magnitude of the fault impact effect. The more severe the fault effect, the greater the impact on the grid node. Therefore, under the equivalent fault impact effect, the capacity reduction of each grid node after fault equivalence is calculated according to formula (3-3): （3-3）； Where C is the capacity of the power grid node before failure. K represents the capacity of a grid node after equivalent reduction, and K is the reduction ratio. Different capacity reduction methods correspond to different reduction ratios K. When reducing capacity according to the topology parameters or electrical characteristic parameters of the grid node, the reduction ratio K is determined by the following formula (4-4): （4-4）； in, It is the result of the topology parameters or electrical characteristic parameters of the coupled power grid nodes after minimum-maximum normalization, and 'a' is the equivalent reduction adjustment parameter.

6. The active defense method for cascading faults in a smart grid according to claim 1, characterized in that, The predictable increase ratio of the grid node capacity can be calculated according to formula (8-8): （8-8）； in, It is the percentage increase in node capacity. The node capacity decay ratio is calculated based on the equivalence quantification method. It is the network's tolerance coefficient.

7. The active defense method for cascading faults in a smart grid according to claim 1, characterized in that, In step 5), to prevent the failed power grid node from causing other normal power grid nodes to become overloaded due to increased load during the load transfer process, thus generating a series of cascading faults, a load shedding method is set for the overloaded power grid nodes during disasters to reduce the impact of cascading faults on power nodes and reduce the transmission of cascading faults in power grid nodes.

8. The active defense method for cascading faults in a smart grid according to claim 7, characterized in that, The load shedding is based on the current node failure status, transferring the load after load shedding to other grid nodes to prevent the fault from continuing to iterate and affecting more grid nodes to the greatest extent. The load shedding is determined based on the current fault iteration number, the degree distribution of grid nodes, and the failure probability of grid nodes under disaster. The more iterations and the higher the failure probability, the less load remains on the grid node. The greater the grid node connectivity, the more load remains.

9. The active defense method for cascading faults in a smart grid according to claim 8, characterized in that, The load shedding reduces the widespread propagation of cascading faults, and the amount of load shedding depends on the number of iterations of the cascading fault within the power grid. The higher the grid node connectivity k and the grid node failure probability P, the more iterations there are, the higher the failure probability, and the lower the remaining load of the grid node. The higher the grid node connectivity, the higher the remaining load. The load shedding ratio can be calculated according to formula (9-9): （9-9）； in, It is the proportion of load cut off by grid node i.

Images