1. A cascading failure risk assessment method based on a network dependency relationship is characterized by comprising the following steps:

step 1, constructing a power information physical system comprising a power grid and an information grid, and improving a dependency relationship condition between the power grid and the information grid;

step 2, constructing a risk transfer function of the fault;

the risk transfer function includes: the risk transfer function of the initial failure of the information node, the risk transfer function of the failure of the power node, the risk transfer function of the subsequent failure of the information node, the risk transfer function of the initial line fault of the power grid and the risk transfer function of the subsequent line fault of the power grid;

step 3, forming an initial fault set of nodes in the information network according to the risk transfer function of the initial failure of the information nodes in the risk transfer function;

defining a variable S, and initializing S to be 1;

step 4, setting the S-th node in the initial fault set as a failure node, and enabling the S-th failure node to be positioned in the first link of the S-th accident chain;

in the S accident chain, obtaining a cross-space propagation process of the fault between the power grid and the information grid according to a risk transfer function of failure of the power node and a risk transfer function of subsequent failure of the information node in the risk transfer functions;

step 5, the power grid carries out regional power rebalance control:

if the power grid is disconnected, firstly, performing active power output adjustment on the generator in the disconnection area, and then performing load flow calculation on the power grid;

if the power grid is not split, directly carrying out load flow calculation on the power grid;

step 6, selecting an initial fault line of the power grid after power rebalancing according to a risk transfer function of the power grid initial line fault in the risk transfer function, and selecting a subsequent fault line according to a risk transfer function of the power grid subsequent line fault in the risk transfer function, thereby forming an accident chain path;

if the on-off information of the fault line is successfully uploaded to the dispatching center, the dispatching center issues corresponding control measures; otherwise, the dispatching center has no control;

step 7, judging whether the accident chain searching meets the ending condition, if so, executing step 9, otherwise, executing step 8;

the ending condition of the accident chain search is as follows: the load loss rate of the power grid reaches a set threshold value or the number of the power grid disconnections reaches a set island threshold value;

step 8, assigning the S +1 to the S and returning to the step 4;

and 9, recording and outputting the relevant data of the whole accident chain, wherein the data comprises the following steps: and calculating the risk of the accident chain by adopting the control measures and control quantity of each link of the accident chain and the accident occurrence probability.

2. The cascading failure risk assessment method based on the network dependency relationship as claimed in claim 1, wherein the dependency relationship condition between the power grid and the information grid in step 1 is improved according to the following process:

step 1.1, except for the autonomous nodes, other information nodes in the information network are regarded as effective nodes and reserved if the dependent nodes supply power, or else, regarded as invalid and deleted;

except for the autonomous nodes, other power nodes in the power grid are regarded as effective nodes and reserved if the dependent nodes of the other power nodes are controlled, and otherwise, the other power nodes are regarded as invalid and deleted;

step 1.2, when the line in the power grid is heavily overloaded, the overload protection device disconnects the corresponding line;

when lines connected with power nodes in a power grid are disconnected, the power nodes become isolated nodes and become invalid, and the invalid power nodes are deleted;

the node load is represented by the node intermediaries in the information network, and the constraint conditions of the information nodes in the information network are constructed by using the formula (1):

in the formula (1), P_cFor the tolerance factor of the jth information node in the information network,for the current load of the jth information node in the information network,the initial load of the jth information node in the information network;

if the load of the jth information node does not meet the constraint condition, the corresponding jth information node is indicated to be overloaded and invalid and is deleted;

step 1.3, when the power grid is decomposed into a plurality of subsets, if at least one generator and one load exist in the subsets, the corresponding subsets can continue to work, the power nodes in the subsets are still effective, and otherwise, the power nodes in the subsets are deleted;

when the information network is split into a plurality of subsets, if the access layer nodes in the subsets can still upload information to the core layer through the communication links, the corresponding subsets can continue to work, the information nodes in the subsets are still effective, otherwise, the information nodes in the subsets are deleted.

3. The cascading failure risk assessment method based on network dependency relationship as claimed in claim 1, wherein the risk transfer function in step 2 comprises: the risk transfer function of the initial failure of the information node, the risk transfer function of the failure of the power node, the risk transfer function of the subsequent failure of the information node, the risk transfer function of the initial line fault of the power grid and the risk transfer function of the subsequent line fault of the power grid; obtaining a risk transfer function of the initial failure of the information node by using a formula (2) to a formula (5); obtaining a risk transfer function of power node failure by using the formula (6); obtaining a risk transfer function of subsequent failure of the information node by using the formula (7);

obtaining the jth information node cause topological structure x by using the formula (2)₁Probability of failure caused

In the formula (2), k_jThe number of the jth information node is Nc, and Nc is the total number of the information network nodes;

obtaining the information occupancy rate x of the jth information node by using the formula (3)₂Probability of failure caused

In the formula (3), n_jThe number of data packets of the jth information node;

obtaining the network attack x of the jth information node by using the formula (4)₃Probability of failure caused

In the formula (4), g_iThe betweenness of the jth information node;

obtaining the risk transfer function of the initial failure of the jth information node by using the formula (5)

Obtaining the risk transfer function of the failure of the jth power node by using the formula (6)

In the formula (6), the reaction mixture is,representing the risk factor x of the jth power node depending on the information node Vc due to the power node Vp₄Probability of failure caused;indicates the j power node is due to the power nodeRisk factor x of point isolation₅Probability of failure caused;

obtaining the risk transfer function of the subsequent failure of the jth information node by using the formula (7)

In the formula (7), the reaction mixture is,the risk factor x representing that the jth information node depends on the power node Vp due to the information node Vc₆Probability of failure caused;representing the risk factor x of the jth information node due to overload of the information node₇Probability of failure caused.

4. The cyber dependency-based cascading failure analysis and risk assessment method for the electric power information physical system according to claim 1, wherein the risk of the accident chain in the step 9 is calculated according to the following steps:

step 9.1, obtaining the ith link S in the S accident chain by using the formula (13)_iResult of Sev (S)_i)：

In formula (13): v'_cl(S_i)、V′_pl(S_i) The ith link S of the S accident chain in the information network and the power grid respectively_iOf the existing load, V_cl、V_plInitial loads of an information network and a power grid respectively;

step 9.2, obtaining the risk R (S) of the S accident chain by using the formula (14):

in the formula (14), p_iFor the ith link S of the S-th accident chain_iThe probability of occurrence.

Background

The power grid and the information grid in the electric power information physical system are interdependent, and faults are transmitted between the two networks in a cross-space mode, so that the power grid and the information grid are damaged to different degrees, and serious economic loss is caused. Most of students only consider the influence of the network structure characteristics on the electric power information physical system when researching the network dependency relationship at present, and have less research on the characteristics of the power grid and the information grid. In the research on the electric power information physical system, the cascading failure caused by single network attack or the failure of the information network is mainly aimed at, and the two are not considered in combination. When risk assessment is carried out on cascading failures of the electric power information physical system, the existing research only considers the risk of an information network or a power grid, and the electric power information physical system is not considered as a whole.

Disclosure of Invention

The invention aims to overcome the defects in the prior art, and provides a cascading failure risk assessment method based on a network dependency relationship, so that an accident chain can be predicted more accurately, the accuracy of risk prediction can be further improved when cascading failures are subjected to risk assessment, and theoretical guidance is provided for the prevention and control of cascading failures of a power information physical system.

In order to achieve the purpose, the invention adopts the following technical scheme:

the invention relates to a cascading failure risk assessment method based on a network dependency relationship, which is characterized by comprising the following steps of:

step 1, constructing a power information physical system comprising a power grid and an information grid, and improving a dependency relationship condition between the power grid and the information grid;

step 2, constructing a risk transfer function of the fault;

the risk transfer function includes: the risk transfer function of the initial failure of the information node, the risk transfer function of the failure of the power node, the risk transfer function of the subsequent failure of the information node, the risk transfer function of the initial line fault of the power grid and the risk transfer function of the subsequent line fault of the power grid;

step 3, forming an initial fault set of nodes in the information network according to the risk transfer function of the initial failure of the information nodes in the risk transfer function;

defining a variable S, and initializing S to be 1;

step 4, setting the S-th node in the initial fault set as a failure node, and enabling the S-th failure node to be positioned in the first link of the S-th accident chain;

in the S accident chain, obtaining a cross-space propagation process of the fault between the power grid and the information grid according to a risk transfer function of failure of the power node and a risk transfer function of subsequent failure of the information node in the risk transfer functions;

step 5, the power grid carries out regional power rebalance control:

if the power grid is disconnected, firstly, performing active power output adjustment on the generator in the disconnection area, and then performing load flow calculation on the power grid;

if the power grid is not split, directly carrying out load flow calculation on the power grid;

step 6, selecting an initial fault line of the power grid after power rebalancing according to a risk transfer function of the power grid initial line fault in the risk transfer function, and selecting a subsequent fault line according to a risk transfer function of the power grid subsequent line fault in the risk transfer function, thereby forming an accident chain path;

if the on-off information of the fault line is successfully uploaded to the dispatching center, the dispatching center issues corresponding control measures; otherwise, the dispatching center has no control;

step 7, judging whether the accident chain searching meets the ending condition, if so, executing step 9, otherwise, executing step 8;

the ending condition of the accident chain search is as follows: the load loss rate of the power grid reaches a set threshold value or the number of the power grid disconnections reaches a set island threshold value;

step 8, assigning the S +1 to the S and returning to the step 4;

and 9, recording and outputting the relevant data of the whole accident chain, wherein the data comprises the following steps: and calculating the risk of the accident chain by adopting the control measures and control quantity of each link of the accident chain and the accident occurrence probability.

The cascading failure risk assessment method based on the network dependency relationship is also characterized in that the dependency relationship condition between the power grid and the information grid in the step 1 is improved according to the following process:

step 1.1, except for the autonomous nodes, other information nodes in the information network are regarded as effective nodes and reserved if the dependent nodes supply power, or else, regarded as invalid and deleted;

except for the autonomous nodes, other power nodes in the power grid are regarded as effective nodes and reserved if the dependent nodes of the other power nodes are controlled, and otherwise, the other power nodes are regarded as invalid and deleted;

step 1.2, when the line in the power grid is heavily overloaded, the overload protection device disconnects the corresponding line;

when lines connected with power nodes in a power grid are disconnected, the power nodes become isolated nodes and become invalid, and the invalid power nodes are deleted;

the node load is represented by the node intermediaries in the information network, and the constraint conditions of the information nodes in the information network are constructed by using the formula (1):

in the formula (1), P_cFor the tolerance factor of the jth information node in the information network,for the current load of the jth information node in the information network,the initial load of the jth information node in the information network;

if the load of the jth information node does not meet the constraint condition, the corresponding jth information node is indicated to be overloaded and invalid and is deleted;

step 1.3, when the power grid is decomposed into a plurality of subsets, if at least one generator and one load exist in the subsets, the corresponding subsets can continue to work, the power nodes in the subsets are still effective, and otherwise, the power nodes in the subsets are deleted;

when the information network is split into a plurality of subsets, if the access layer nodes in the subsets can still upload information to the core layer through the communication links, the corresponding subsets can continue to work, the information nodes in the subsets are still effective, otherwise, the information nodes in the subsets are deleted.

The risk transfer function in step 2 comprises: the risk transfer function of the initial failure of the information node, the risk transfer function of the failure of the power node, the risk transfer function of the subsequent failure of the information node, the risk transfer function of the initial line fault of the power grid and the risk transfer function of the subsequent line fault of the power grid; obtaining a risk transfer function of the initial failure of the information node by using a formula (2) to a formula (5); obtaining a risk transfer function of power node failure by using the formula (6); obtaining a risk transfer function of subsequent failure of the information node by using the formula (7);

obtaining the jth information node cause topological structure x by using the formula (2)₁Probability of failure caused

In the formula (2), k_jThe number of the jth information node is Nc, and Nc is the total number of the information network nodes;

obtaining the information occupancy rate x of the jth information node by using the formula (3)₂Probability of failure caused

In the formula (3), n_jThe number of data packets of the jth information node;

obtaining the network attack x of the jth information node by using the formula (4)₃Probability of failure caused

In the formula (4), g_iThe betweenness of the jth information node;

obtaining the risk transfer function of the initial failure of the jth information node by using the formula (5)

Obtaining the risk transfer function of the failure of the jth power node by using the formula (6)

In the formula (6), the reaction mixture is,representing the risk factor x of the jth power node depending on the information node Vc due to the power node Vp₄Probability of failure caused;representing the risk factor x of the jth power node due to the isolation of the power node₅Probability of failure caused;

obtaining the risk transfer function of the subsequent failure of the jth information node by using the formula (7)

In the formula (7), the reaction mixture is,the risk factor x representing that the jth information node depends on the power node Vp due to the information node Vc₆Probability of failure caused;representing the risk factor x of the jth information node due to overload of the information node₇Probability of failure caused.

The risk of the accident chain in step 9 is calculated as follows:

step 9.1, obtaining the ith link S in the S accident chain by using the formula (13)_iResult of Sev (S)_i)：

In formula (13): v'_cl(S_i)、V′_pl(S_i) The ith link S of the S accident chain in the information network and the power grid respectively_iOf the existing load, V_cl、V_plInitial loads of an information network and a power grid respectively;

step 9.2, obtaining the risk R (S) of the S accident chain by using the formula (14):

in the formula (14), p_iFor the ith link S of the S-th accident chain_iThe probability of occurrence.

Compared with the prior art, the invention has the beneficial effects that:

1. the method improves the conditions of the network dependency relationship, and adds the consideration to the transmission characteristics of the power grid and the information grid on the basis of researching the network topology dependency relationship, namely, the conditions of isolated failure of the power nodes (caused by the disconnection of the circuit in the power grid due to serious overload) and overload failure of the information nodes are considered in the dependency process, the cross-space propagation process of the fault between the dependent networks is further researched, and the accuracy of the accident chain prediction is improved.

2. According to the invention, the accident chain search is carried out by combining two factors of network attack and self fault of the information node, the failure of the information node caused by network attack, topological structure and initial data packet distribution is considered, and the failure probability caused by betweenness, degree and information occupancy rate of the information node is integrated, so that the information node which has a large influence on the electric power information physical system can be screened more comprehensively, the key node is effectively identified, and the prevention and control are carried out in time.

3. According to the invention, a risk element theory is applied in the cascading failure prediction process, an accident chain which is greatly influenced by risk factors is found out, a novel risk calculation method is provided, the node load is represented by an intermediate number, the accident chain consequence integrates the power grid load loss rate and the information grid load loss rate, the accuracy of risk prediction is improved, and the method is suitable for the research of cascading failure risk evaluation of the power information physical system.

Drawings

Fig. 1 is a diagram illustrating an IEEE39 node system according to the present invention.

FIG. 2 is a schematic diagram of a physical system for power information according to the present invention;

FIG. 3 is a diagram of an information network architecture according to the present invention;

FIG. 4 is a flowchart illustrating the network dependency phase incident chain prediction process of the present invention;

FIG. 5 is a flow chart of the accident chain prediction at a subsequent failure stage of the present invention;

FIG. 6 is a comparison chart of different risk calculations after the initial node fails according to the present invention;

FIG. 7 is a comparison graph of the average risk in the network dependency phase of the initial fault set based on different factors.

Detailed Description

In this embodiment, a cascading failure risk assessment method based on a network dependency relationship is performed according to the following steps:

step 1, constructing a power information physical system comprising a power grid and an information grid, and improving a dependency relationship condition between the power grid and the information grid;

applying a complex network theory, and respectively representing a power grid and an information grid by using graphs Gp (Vp, Ep) and Gc (Vc, Ec), wherein Vp is power grid physical equipment, Ep is a line, and Ec is a communication link; the information node Vc is responsible for data transmission and processing of the power node Vp, and the power node Vp provides power support for the information node Vc.

The information network consists of three parts: core layer, bone stem layer, access layer.

The core layer is responsible for economic dispatching and safe operation of a power grid, is generally provided with a main dispatching center and a standby dispatching center, and uses high-end routing equipment to ensure uninterrupted energy supply so as to ensure reliability;

the backbone layer is responsible for regional scheduling, is provided with a standby power supply, the number of nodes of the standby power supply is related to the number of partitions of a power grid, the convergence and the forwarding of access layer information in a region are dependent on the backbone layer of the region, and in order to improve the reliability of information transmission, the nodes of the backbone layer are interconnected in a ring;

the access layer nodes are responsible for collecting power grid information and transmitting control commands, wherein the nodes connected with the transformer substation and the power plant in the access layer are directly connected with the backbone layer, and part of important power plant nodes are connected with the backbone layer and the core layer.

In this embodiment, an IEEE39 node power system is used to simulate a cascading failure cross-space propagation process of a power information physical system, a structure of an IEEE39 node system is shown in fig. 1, a structure of a power information physical system is shown in fig. 2, and a structure of an information network is shown in fig. 3.

The dependency relationship condition between the power grid and the information grid is improved according to the following process:

step 1.1, except for the autonomous nodes, other information nodes in the information network are regarded as effective nodes and reserved if the dependent nodes supply power, or else, regarded as invalid and deleted;

except for the autonomous nodes, other power nodes in the power grid are regarded as effective nodes and reserved if the dependent nodes of the other power nodes are controlled, and otherwise, the other power nodes are regarded as invalid and deleted;

step 1.2, when the line in the power grid is heavily overloaded, the overload protection device disconnects the corresponding line;

when lines connected with power nodes in a power grid are disconnected, the power nodes become isolated nodes and become invalid, and the invalid power nodes are deleted;

the node load is represented by the node intermediaries in the information network, and the constraint conditions of the information nodes in the information network are constructed by using the formula (1):

in the formula (1), P_cFor the tolerance factor of the jth information node in the information network,for the current load of the jth information node in the information network,the initial load of the jth information node in the information network;

and if the load of the jth information node does not meet the constraint condition, indicating that the corresponding jth information node is overloaded and invalid and deleting the information node.

Step 1.3, when the power grid is decomposed into a plurality of subsets, if at least one generator and one load exist in the subsets, the corresponding subsets can continue to work, the power nodes in the subsets are still effective, and otherwise, the power nodes in the subsets are deleted;

when the information network is split into a plurality of subsets, if the access layer nodes in the subsets can still upload information to the core layer through the communication links, the corresponding subsets can continue to work, the information nodes in the subsets are still effective, otherwise, the information nodes in the subsets are deleted.

Step 2, constructing a risk transfer function of the fault;

the risk transfer function in cascading failures of the electric power information physical system comprises the following steps: the risk transfer function of the initial failure of the information node, the risk transfer function of the failure of the power node, the risk transfer function of the subsequent failure of the information node, the risk transfer function of the initial line fault of the power grid and the risk transfer function of the subsequent line fault of the power grid;

obtaining the jth information node cause topological structure x by using the formula (2)₁Probability of failure caused

In the formula (2), k_jThe number of the jth information node is Nc, and Nc is the total number of the information network nodes;

obtaining the information occupancy rate x of the jth information node by using the formula (3)₂Probability of failure caused

In the formula (3), n_jThe number of data packets of the jth information node;

obtaining the network attack x of the jth information node by using the formula (4)₃Probability of failure caused

In the formula (4), g_iThe betweenness of the jth information node;

obtaining the risk transfer function of the initial failure of the jth information node by using the formula (5)

Obtaining the risk transfer function of the failure of the jth power node by using the formula (6)

In the formula (6), the reaction mixture is,representing the risk factor x of the jth power node depending on the information node Vc due to the power node Vp₄Probability of failure caused;representing the risk factor x of the jth power node due to the isolation of the power node₅Probability of failure caused.

Obtaining the risk transfer function of the subsequent failure of the jth information node by using the formula (7)

In the formula (7), the reaction mixture is,the risk factor x representing that the jth information node depends on the power node Vp due to the information node Vc₆Probability of failure caused;representing the risk factor x of the jth information node due to node overload₇Probability of failure caused.

The initial line fault L of the power grid can be obtained by using the formula (8)_kRisk transfer function ofComprises the following steps:

in formula (8), the risk factor x₈Is a line aging factor, and x₉In order to be a factor of the trend,the method is characterized in that when the power system normally operates, the line fault probability in the same region and the same period is as follows:

in formula (9), Np is the power grid bus line l_k、λ_okRespectively, is a line L_kLength and failure rate per unit length.

When the power grid is not in fault, L_kProbability of failure due to tidal current factors:

in the formula (10), eta_kIndicates the line load factor index, omega_kRepresenting a power flow ratio index; e.g. of the type_kThe power flow change index is expressed, and the following indexes are included:

in the formulae (11) to (13), F_k,0Is L_kInitial tidal current value of, F_k,maxIs L_kCurrent limit value of, Δ F_iIs L_kL after operation quit_iA tidal current variation of;

using the power flow fluctuation index A_kmTidal current coupling index B_kmAnd a line load rate index H_kmTo represent L_kImpact on other lines after disconnection:

in the formulae (14) to (16), F_m、F_m,maxAre respectively a line L_mTidal current value and tidal current ofLimit value, F'_mIs L_kAfter breaking L_mPower flow value of F_kIs a line L_kTidal current value of, H_kmIs a line L_mThe load factor of (c).

When the line tide exceeds the limit value, if the dispatching center receives the line overload information, overload control is adopted; and if the overload protection circuit cannot receive the overload protection circuit, the circuit with the highest overload degree is protected and disconnected. Otherwise, according to the relevance indexSelecting a lower-level cut-off line:

the greater the correlation index is, the greater L is_kBreaking pair L_mThe larger the influence of the fault probability is, when the relay protection misoperation rejection factor of the power grid is considered, if the circuit is seriously overloaded, the subsequent circuit fault probability is the probability that the circuit overload protection and the circuit breaker both correctly act; if the line is generally overloaded, the probability of a subsequent line fault comprises three parts:

(1) probability of failure P caused by power flow transfer₁：

(2) Probability of failure P caused by protection malfunction and breaker malfunction₂：

P₂＝P_{mis_r}×(1-P_{in_c})+P_{mis_c} (19)

In formula (19): p_{mis_r}、P_{in_c}、P_{mis_c}Respectively, protection malfunction, circuit breaker failure, and circuit breaker malfunction probability.

(3) Fault probability P caused by accidental factors such as error of dispatcher, bad weather, aging fault of element₃。

The risk transfer function for a subsequent line fault of the power grid can be obtained by using equation (20) as follows:

in the formula (20), P_{in_r}To protect the probability of a refusal action, a risk factor x₁₀Is a power flow transfer factor.

Step 3, forming an initial fault set of the nodes in the information network according to the risk transfer function;

defining a variable S and initializing S to be 1;

the initial failure probability of the nodes in the information network is calculated, so that an initial failure set of the nodes in the information network is formed according to the set failure probability threshold value as shown in table 1.

TABLE 1 initial failure set

Vc16	0.1851
		Vc2	0.1430
Vc6	0.1044
		Vc8	0.1008
Vc26	0.0100
		Vc14	0.0992
Vc17	0.0967
		Vc25	0.0957
Vc27	0.0776
		Vc15	0.0757

Step 4, setting the S-th node in the initial fault set as a failure node, and enabling the S-th failure node to be positioned in the first link of the S-th accident chain;

in the S accident chain, obtaining a cross-space propagation process of the fault between the power grid and the information network according to the risk transfer function; this phase is called the web dependent phase, and the process of performing incident chain prediction is shown in fig. 4.

Step 5, the power grid carries out regional power rebalance control:

if the power grid is disconnected, firstly, performing active power output adjustment on the generator in the disconnection area, and then performing load flow calculation on the power grid;

if the power grid is not split, directly carrying out load flow calculation on the power grid;

step 6, selecting an initial fault line of the power grid after power rebalancing according to the risk transfer function, and selecting a subsequent fault line according to the risk transfer function, thereby forming an accident chain path;

if the on-off information of the fault line is successfully uploaded to the dispatching center, the dispatching center issues corresponding control measures; otherwise, the dispatching center has no control; this phase is called the subsequent failure phase and the fault chain prediction process is implemented as shown in fig. 5.

Step 7, judging whether the accident chain searching meets the ending condition, if so, executing step 9, otherwise, executing step 8;

the ending conditions of the accident chain search are as follows: the load loss rate of the power grid reaches 30% or the power grid is divided into three subsets;

taking Vc8 and Vc27 failures as examples, the corresponding cascading failure evolution processes are shown in table 2 and table 3, and the development of the network dependency stage and the subsequent failure stage is described in detail respectively.

TABLE 2 Vc8 cascading failure evolution process (Unit: MW) for failure

In table 2, the information node Vc8 fails due to the dual effects of self-failure and network attack, the network dependency relationship Vp8 fails, the load of the calculated information node and the power grid load are known to be known, Vc6 and Vc4 fail due to overload, no serious overload line exists, the network dependency relationship Vp6 and Vp4 fails, the load of the information node and the power grid load are calculated again, the L1 and L2 are seriously overloaded, the protection action is performed by disconnecting L1 and L2, so that Vp1 is isolated and fails, the network dependency relationship Vp1 fails, the load of the information node and the power grid load are calculated again, no information node or power node fails, the transient state of the system is stable, the system is classified into three parts at this time, regional power balance is performed, and the network dependency stage ends. Since the end condition is reached, there is no subsequent fault phase and the cascading faults stop.

TABLE 3 cascading failure evolution process (unit: MW) for Vc27 failure

In table 3, the information node Vc27 fails due to the dual effects of the self-fault and the network attack, and the network dependency phase ends because the power node Vp27 fails, the information node load and the power grid load are calculated, no information node or power node fails, and the network dependency phase ends. According to a risk transfer function of the initial line fault of the power grid, finding out that the initial fault of a subsequent fault stage is L12, after L12 is disconnected, the overload is generally eliminated by L17 and L18, a dispatching center issues an overload control command, a core-layer node controls a dependent power node to execute the control command, a generator 3 cuts 44MW, and a load 15 cuts 44 MW; selecting a lower-level fault according to a risk transfer function of a subsequent line fault of the power grid, disconnecting L19 and overloading a wireless path; selecting a lower-level fault L6 again, wherein the system is unstable at the moment and is split into two parts for emergency control; subsequently, sequentially disconnecting the L24 and the L9 to overload a wireless path; after the L13 is disconnected, the system is decomposed into three parts, the regional power balance is performed, the end condition is reached, and the cascading failure stops.

Step 8, assigning the S +1 to the S and returning to the step 4;

and 9, recording and outputting the relevant data of the whole accident chain, wherein the data comprises the following steps: and calculating the risk of the accident chain by adopting the control measures and control quantity of each link of the accident chain and the accident occurrence probability.

Obtaining the ith link S in the S accident chain by using the formula (21)_iResult of Sev (S)_i)：

In formula (21): v'_cl(S_i)、V′_pl(S_i) The ith link S of the S accident chain in the information network and the power grid respectively_iOf the existing load, V_cl、V_plInitial loads of an information network and a power grid respectively;

the risk of the S accident chain r (S) is obtained using equation (14):

in the formula (22), p_iFor the ith link S of the S-th accident chain_iThe probability of occurrence.

In the embodiment, the tolerance coefficient pc of the information network node takes a value of 0.2, the overload control adopts sensitivity-based tripping load shedding, and the probability of protection refusal is P_{in_r}0.0013 outline of circuit breaker failureA rate of P_{in_c}0.0005, the probability of circuit breaker malfunction is P_{mis_c}0.0001, probability of accidental factor is P₃0.0002. The risk 1 is used for representing the cascading failure risk calculated by the method, the risk 2 is used for representing the risk calculated only by considering the power grid consequence, and the risk after the initial failure centralized information node of the table 2 is calculated by using the risk calculation methods is shown in fig. 6.

Comparing risks 1 and 2, there is a clear difference in the trend of the two. After the failure of Vc17, Vp17 is failed due to network dependence, L25 is connected with Vp17 and Vp18 and is disconnected, L6 is disconnected in a subsequent fault stage, the disconnection of L25 and L6 causes the failure of Vp18 due to isolation, Vc18 is failed due to network dependence, no information node or power node is failed, and L20 and L30 are disconnected subsequently. Because the information network loss is large when the subsequent fault stage comprises a network dependence stage, the risk of Vc17 failure in risk 1 is large, and risk 2 only considers the loss load rate of the power network, ignores the information network and causes lower risk. Compared with risk 2, the method takes the loss of the information network into consideration, and considers the electric power information physical system as a whole, so that the cascading failure risk can be comprehensively evaluated, and certain practical significance is achieved.

The Vc8 connects the areas 2 and 3, and after the failure, the other nodes Vp8, Vc6, Vc4, Vp6, Vp4, Vp1 and Vc1 in the areas 2 and 3 sequentially fail due to the network dependency relationship, so that the number of affected nodes is large, the range is wide, and the risk of cascading failure is large. After the Vc16 fails, no other node fails due to the network dependency relationship, but the risk of cascading failures is still large, because the Vc16 is located in the area 4 and connected with the areas 2 and 3, after the Vc16 fails, the three areas are affected, the influence range is large, and after the Vc16 fails, the system is immediately solved into three parts, which shows that the Vc16 is critical in the network, so the probability of the failure is the largest, and the risk is also large. The partition result is obtained by modifying the community partition algorithm through the Prim algorithm, the community partition algorithm is adopted to enable the internal connection of the community to be tight, and the Prim algorithm modification enables the electrical connection between the communities to be greatly enhanced. The Vc16 and Vc8 are located at the junction of the areas and are connected with at least two areas, so that the failure of the Vc16 and the Vc8 affects the areas connected with the Vc16 and the Vc8, and nodes in the areas fail.

Fig. 7 is a comparison of average risks at network dependency stages of initial fault sets with different factors, and it can be seen that the average risk is highest when a network attack is considered in combination with a node self-failure, and the average risk is lowest when only a self-failure is considered, because the information node self-failure is related to a network topology and information occupancy, and the network topology is related to the degree of the node, and a large degree indicates that more nodes are connected to the node, but the importance of the connected node is not considered, and global consideration is lacked, and the node importance is not considered in the failure caused by the information occupancy, and a certain contingency exists. The high betweenness indicates that the relationship between any two nodes in the network is strongly influenced by the node, the importance degree of the node in the whole network is high, and the node is a measurement based on a global path, so that the influence caused by considering betweenness in network attack is larger than that caused by considering degree and information occupancy rate in a failure scene. The method combines network attack with node self failure, comprehensively considers failure probability caused by betweenness, degree and information occupancy rate, screens out information nodes which have large influence on an information network, makes prevention and control measures for the information network in time and avoids loss caused by further spread of cascading failure.

TABLE 4 influence of network dependencies on cascading failures (Unit: MW)

Table 4 takes Vc27 as an example to compare and analyze the influence of the network dependency stage on cascading failures, if the dependency relationship between the power grid and the information grid is not considered, that is, Vp27 failure cannot be caused by Vc27 failure, the topology structure of the power grid is not changed, and after the subsequent failure stages L3, L4, and L30 are disconnected, the information can still be uploaded to the scheduling center, and the scheduling center formulates and issues a corresponding control strategy according to the received information. And the Vc27 uploads the on-off information of the L31, after the on-off information of the L31 is failed, the dispatching center cannot receive the on-off information of the L31, no control command is issued, the actual power grid is decomposed into three parts, and the safety and stability device acts.

If the dependency relationship between the power grid and the information network is considered, namely after Vc27 fails, Vp27 also fails, the topology structure of the power grid changes, the initial fault line changes in the subsequent fault stage, and the spread of the subsequent accident chain is further influenced. Comparing the risk of the two accident chains, and considering the chain fault risk when considering the network dependency relationship is larger than the situation which is not considered. In the subsequent fault stage, only 4 lines are disconnected when the network dependence relationship is not considered, and 6 lines are disconnected when the network dependence relationship is considered, however, due to the network dependence relationship, Vp27 fails, all the loads 27 are cut off, the L26 and the L31 are disconnected, the power grid topology is changed, at this time, although the wireless path is overloaded, the power grid is damaged, the reliability is reduced, and the disconnection probability of the subsequent lines is increased. The network dependence phenomenon generally exists in the electric power information physical system, and if the dependence relationship between a power grid and an information grid can be reduced through decoupling, the reliability of the electric power information physical system can be greatly improved, and the cascading failure risk is reduced.