CN107769962A - A kind of communication network failure cascade venture influence analysis method of attack resistance - Google Patents

A kind of communication network failure cascade venture influence analysis method of attack resistance Download PDF

Info

Publication number
CN107769962A
CN107769962A CN201710847598.1A CN201710847598A CN107769962A CN 107769962 A CN107769962 A CN 107769962A CN 201710847598 A CN201710847598 A CN 201710847598A CN 107769962 A CN107769962 A CN 107769962A
Authority
CN
China
Prior art keywords
node
power
failure
network
state
Prior art date
Legal status (The legal status is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the status listed.)
Pending
Application number
CN201710847598.1A
Other languages
Chinese (zh)
Inventor
刘旭
董武
汤玮
唐洁瑶
刘璐
袁汉云
刘康
刘晴
Current Assignee (The listed assignees may be inaccurate. Google has not performed a legal analysis and makes no representation or warranty as to the accuracy of the list.)
Guizhou Power Grid Co Ltd
Original Assignee
Guizhou Power Grid Co Ltd
Priority date (The priority date is an assumption and is not a legal conclusion. Google has not performed a legal analysis and makes no representation as to the accuracy of the date listed.)
Filing date
Publication date
Application filed by Guizhou Power Grid Co Ltd filed Critical Guizhou Power Grid Co Ltd
Priority to CN201710847598.1A priority Critical patent/CN107769962A/en
Publication of CN107769962A publication Critical patent/CN107769962A/en
Pending legal-status Critical Current

Links

Classifications

    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14Network analysis or design
    • H04L41/145Network analysis or design involving simulating, designing, planning or modelling of a network
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/06Management of faults, events, alarms or notifications
    • H04L41/0654Management of faults, events, alarms or notifications using network fault recovery
    • H04L41/0663Performing the actions predefined by failover planning, e.g. switching to standby network elements
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/12Discovery or management of network topologies
    • HELECTRICITY
    • H04ELECTRIC COMMUNICATION TECHNIQUE
    • H04LTRANSMISSION OF DIGITAL INFORMATION, e.g. TELEGRAPHIC COMMUNICATION
    • H04L41/00Arrangements for maintenance, administration or management of data switching networks, e.g. of packet switching networks
    • H04L41/14Network analysis or design
    • H04L41/142Network analysis or design using statistical or mathematical methods

Landscapes

  • Engineering & Computer Science (AREA)
  • Computer Networks & Wireless Communication (AREA)
  • Signal Processing (AREA)
  • Data Exchanges In Wide-Area Networks (AREA)

Abstract

The present invention relates to a kind of communication network failure of attack resistance to cascade venture influence analysis method, belongs to network technique field.This method comprises the following steps:S1:Establish coupling network model:Power network model is established, establishes overall network model, refines overall network model;S2:Analyze coupling network model:Explicitly defining for power system node failure and communication network node failure is provided respectively;The necessary condition that failure is formed is analyzed on this basis;Based on the generation for avoiding or suppressing failure and the consideration of optimization communication subnet, using the thought for destroying necessary condition, virtual backbone network concept is introduced;Finally propose important node group backup optimisation strategy.The present invention is based on Complex Networks Theory, influence of the physical attribute to power telecom network of optical cable is considered, the important node group backup optimized algorithm of power telecom network is proposed, preferably suppresses the generation of cascading failure in the case of fault attacks, is effectively improved the robustness of power telecom network.

Description

A kind of communication network failure cascade venture influence analysis method of attack resistance
Technical field
The invention belongs to network technique field, the communication network failure for being related to a kind of attack resistance cascades venture influence analysis side Method.
Background technology
Intelligent grid is the network of coupling electrical power system and communication network, and the network such as router in communication network is set It is standby to need power system power supply, and the scheduling of power system and intelligent decision need communication network to transmit and control, therefore, power network Network and communication network form isomery couple it is mutual according to mutually beneficial network.Mutually attract numerous researchers' according to the peculiar property of coupling Concern.Part researcher has inquired into the coupled relation on side in mutually beneficial and heterogeneous network, also has researcher to being equipped with smart electronicses The coupled system of the change electronic station of the advanced smart machine such as equipment I EDs and flexible AC electricity transmission system FACTS is ground Study carefully.But most of research have ignored to the interactional research of internal node at present, cause optimizing decision-making and failure During management, specific traffic assignments are excessively paid close attention to, have ignored the importance of bottom topological structure.For this problem, due to intelligence The complete job of energy power network is dependent on electric power networks and communication network mutually according to the overall network of coupling, and in data dissemination and information Node in communication process in electric power networks and communication network influences each other, therefore, it is necessary to which the power system to complexity is carried out Abstract modeling, understand influencing each other for power node and communication node in power network, process caused by data dissemination and failure.
Then, on this basis, problem and network architecture un-reasonable phenomenon are cascaded based on power telecom network failure, are based on Complex Networks Theory, consider influence of the physical attribute to power telecom network of optical cable, propose a kind of reduction power communication The important node group backup optimized algorithm of net cascading failure risk, it is effectively improved the robustness of power telecom network.
To understand the state of development of prior art, retrieval is carried out to existing patent, has compared and analyzes, has been filtered out as follows The higher technical information of degree related to the present invention:
Technical scheme 1:Patent No. CN201510919627.1 patent《A kind of risk balance method of power telecom network And system》, the invention proposes a kind of risk balance method and system of power telecom network, and methods described includes:Electric power is obtained to lead to The link information and business information between nodal information, adjacent node in letter net;According to the information of acquisition, it is determined that each industry Multiple service paths corresponding to business;A service path is arbitrarily chosen from multiple service paths respectively, forms a business sheet Set of paths;According to the information of acquisition and at least one business sheet set of paths, it is determined that the risk of each node and every The risk of individual link;According to the risk of each node and the risk of each link, network risks degree is determined;According to extremely Few a business sheet set of paths and business information, determine network load degree;According to network risks degree and network load degree, Using default service path Optimized model, the path optimizing of each business is obtained.The invention can consider power telecom network Risk and load, be effectively reduced the risk of power telecom network.
Technical scheme 2:Patent No. CN201410814118.8 patent《A kind of data network wind based on HWIL simulation Dangerous appraisal procedure》, a kind of data network methods of risk assessment based on HWIL simulation of invention proposition, this method can from connection The evaluation index to be failed by property as power telecom network;The fault tree and event tree of power telecom network failure, analysis are built first The reason for power telecom network fails;Then fault tree and event tree are directed to respectively, propose the algorithm converted to Bayesian network, should Power telecom network failure degree is judged with Bayesian network.The invention from the angle of graph theory power telecom network is failed into Row analysis, the origin cause of formation of power telecom network failure is analyzed, the evaluation failed from connected sets as power telecom network Index.
Technical scheme 3:Patent No. CN201610542404.2 patent《Power telecom network risk evaluating system》, the hair It is bright to disclose power telecom network risk evaluating system, including risk assessment device and multiple power communication devices, the power communication Device is connected to risk assessment device by network interface, and the risk assessment device generates including power telecom network evaluation index system Module, opinion rating system generation module, quantification of targets module, index weights computing module, subordinated-degree matrix structure module, mould Paste comprehensive evaluation result computing module and risk evaluation module.Power telecom network risk evaluating system of the present invention, by obscuring square The application in risk assessment of battle array and analytic hierarchy process (AHP), can to occur during risk assessment various uncertain factors, refer to Mark is analyzed, and estimating velocity is fast, precision is higher.
The path optimizing of target formation business of the technical scheme 1 based on balance risk, technical scheme 2,3 is respectively using difference Method assess power telecom network risk.Although scheme 1 can consider risk and the load of power telecom network, effectively The risk of power telecom network is reduced, but its course of work does not account for the coupled relation between powerline network and electric power networks, Research is not provided to the back mechanism of network node or link circuit resource from the angle for reducing cascading failure as far as possible;Scheme 2 is examined What is considered is the risk assessment of data network, only proposes the evaluation index to be failed from connected sets as power telecom network, does not have Provide how the method for optimizing evaluation result;The also assessment to various risks index of scheme 3, what is provided is a set of evaluation The computing system of index, the specific steps how to reduce risks and algorithm are not provided.
The content of the invention
In view of this, it is an object of the invention to provide a kind of communication network failure of attack resistance cascade venture influence analysis side Method, problem and network architecture un-reasonable phenomenon are cascaded based on power telecom network failure, analyze the high coupling of electric network composition And basic reason caused by failure, based on Complex Networks Theory, the physical attribute of optical cable has been considered to power telecom network Influence, it is proposed that the important node group backup optimized algorithm of power telecom network, can preferably suppress in the case of fault attacks The generation of cascading failure, it is effectively improved the robustness of power telecom network.
To reach above-mentioned purpose, the present invention provides following technical scheme:
A kind of communication network failure cascade venture influence analysis method of attack resistance, comprises the following steps:
S1:Establish coupling network model:Power network model is established, establishes overall network model, refines overall network mould Type;
S2:Analyze coupling network model:The clear and definite of power system node failure and communication network node failure is provided respectively Definition;The necessary condition that failure is formed is analyzed on this basis;Based on the generation and optimization communicator for avoiding or suppressing failure The consideration of net, using the thought for destroying necessary condition, introduce virtual backbone network concept;Finally propose that the backup of important node group is excellent Change strategy.
Further, the node failure is expressly defined as:
(1) power failure after optimizing propagates (Power Fault Propagation, PFP) model node failure property shown Function δ (x) is expressed as:
(2) correlation function corr is defined as: Wherein vj、viFor power node, Vc、VpFor node set, EpcFor two couplings Close the link between network, Epc=(u, v) | u ∈ Vp,v∈Vc, wherein u is power node.
Further, the necessary condition of the failure formation is:
(1) power failure treatment mechanism is when treatable failure problems occur, mutually according in coupling network, with for the moment Between a failure pair at least be present;
(2) critical path relevant with data transfer in mutually according to coupling network be present.
Further, the critical path is:Wherein SjFor electricity Stand, Ri、Ri+1For node, Ge is power plant.
Further, the critical path is solved:
S201:Solve 2 dominant sets in communication subnet by didactic mode to generate, draw intermediate result set G1, this When G1 necessarily meet constraint 1, i.e., when the not treatable failure problems of power failure treatment mechanism occur, mutually according to coupling network In, a failure pair at least be present in the same time, also certain satisfaction constraint 1 is solved on the basis of G1;
S202:Based on G1, (1,2)-CDS is calculated, is designated as G2;I.e. 1 connection 2 dominate set, minimum connected dominating set for (1, 1)-CDS;
S203:Based on G2, comprehensive physical constraint, final result G is drawnx
Final result provides two set:G2 and Gx, and the two is contrasted in terms of algorithm evaluation.
Further, the step S201 is specially:
Input G=(Vp∪Vc,Ec∪Epc), solve G1:
(1) according to EpcArrangement is ranked up, finds the routing node of all connectionsAnd according to its coupling with power node Number is closed to be ranked up;
(2) power node set is investigated, is examined respectively with the degree being connected of communication node according to power node therein Consider:
If a) degree is 1, illustrate that the power node is only connected with a communication node, will for the consideration of network-in-dialing The communication node of power node connection adds G1, avoids the power node from turning into isolated point;
If b) degree is 2, illustrate that power node is only connected with two communication nodes, two that the power node is connected are logical Letter node all adds G1;
If c) degree is more than or equal to 3, illustrate that power node is connected with multiple communication nodes, by the power node set record Get off and investigated:
I) assume to have entered G1 in the presence of two communication nodes wherein, then current results collection meets that 2 dominate, and is not required to Add G1;
Ii) assume only one or enter G1 without existing node, then node is introduced into these and is ranked up, root According to its connected power node number, the addition G1 in selection ranking forefront.
Further, the step S202 is specially:
G1 is inputted, (1,2)-CDS is solved, is designated as G2:
(1) G2=G1 is initialized;G2 concrete structure is investigated, if G2 has been connection, then G2 is result set;Such as Fruit G2 is not connected, then the number of record function T marks G2 connected subgraph, and each connected subgraph is indicated, and its T value must More than or equal to 2;
(2) target of Algorithm for Solving is that T values are equal to 1, now illustrates that G2 has been connected;Most probable is selected from communication subnet The node for declining T values adds G2;Specifically selection algorithm is:
A) initialize VcNode state be labeled as 0, neighborhood RiIt is initialized as empty set;N set expressions G2 to be added Node, be initialized as empty set;
B) node in G2 is traveled through, G2 is divided into T connected subgraph, its internal node is labeled as 1,2 respectively, 3…T;Each connected subgraph is traveled through, to k-th of connected subgraph, its internal node is labeled as k, all neighbor nodes Neighborhood adds k;
C) there will be the node of neighborhood to add N set;Node in gathering for N, arranged according to the degree of neighborhood Sequence, maximum node will be spent and add G2;Assuming that its value of the node of degree maximum of neighborhood is 1, then connected according to node logical Letter node number is ranked up.
Further, the step (1) is specially:
1) according to EcGenerate the annexation of each power node:DegC [i]=(Ci,m,N[Ci]), wherein CiFor node, M represents the connection of power node remaining power node number in G2, N [Ci] represent neighborhood, array according to m from big to small Sequence;
2) T=1, the State [i]=0 of each node are initialized;First m maximum of taking-up from DegC [] array Node Cj;According to DegC [j], by CjState function setups with N [j] are T;
3) DegC [] array is traveled through, for k-th of power node, according to CkState [k] and DegC [] handled:
If a) State [k]!=0, it has been the member of a connected set to illustrate the node, then State [N [Ck]]= State[k];
If b) State [k]==0, consider State [N [Ck]] set:
I) set is all 0, then T adds 1, State [k]=T, State [N [Ck]]=State [k];This situation is the section Front nodal point does not connect point therewith with neighbor node;
Ii) set only has a kind of value x in addition to 0, then State [k]=x, State [N [Ck]]=State [k];This feelings Condition is that the node is connected with a connected subgraph;
Iii) set has t kind values in addition to 0, and t is more than or equal to 2 herein, and it is x to take minimum in these values, then T=T-t+ 1, State [k]=x, State [N [Ck]]=State [k];Again all values in State functions are equal to the set in addition to 0 Arbitrary value is assigned to min;This situation is that the node is connected with multiple connected subgraphs.
Further, the GxSpecific algorithm is:
G2 is inputted, final result is solved, is designated as Gx
(1) judge whether the routing node being connected with power station adds G2, do not add such as, be then put into;
(2) select the maximum node of remaining node moderate to add successively, until a little all add or current collection is expired Foot-eye requirement.
Further, the important node group backup optimisation strategy is:
(1) randomly generate node failure sequence and be applied on real topological environmental, to attacking calculating each time Once by discrete dot pattern into damage curve figure;
(2) for each curve, it is once to be weighed caused by the failure sequence to obtain the maximum node of every section of tangent slope Node is wanted, and is added into set, and records an important node number of each node;
(3) using repeatedly random attack, an important node number set is formed, the random more results of number of times of attack are more smart Really, finally sorted from big to small according to random number of times of attack;
(4) the actual backup requirements of network are combined and carry out specific backup policy.
The beneficial effects of the present invention are:The present invention has carried out network modelling and topological property point to topological structure of electric Analysis, under different faults attack strategies, algorithm proposed by the present invention is on failure effect is suppressed better than degree sort algorithm.
Brief description of the drawings
In order that the purpose of the present invention, technical scheme and beneficial effect are clearer, the present invention provides drawings described below and carried out Explanation:
Fig. 1 is mutually according to coupling network model;
Fig. 2 is algorithm flow chart;
Fig. 3 is simple coupling network;
Fig. 4 is power communication backbone network;
Fig. 5 is artificial network;
Fig. 6 is optical cable topological structure;
Fig. 7 is G1 arithmetic results;
Fig. 8 is G1 topological structures;
Fig. 9 is G2 arithmetic results;
Figure 10 is G2 topological structures;
Figure 11 is GxTopological structure;
Figure 12 is recognizer;
Figure 13 is that black attack is compared;
Figure 14 is that grey attack is compared;
Figure 15 is without the topological performance results of backup.
Embodiment
Below in conjunction with accompanying drawing, the preferred embodiments of the present invention are described in detail.
1. intelligent grid is mutually according to coupling model
1.1 coupling network models
First consider the modeling of electric power networks.In order to express easily, arrange G and represent coupling network, Ge represents power station, SiTable Show power station.Given electric power networks Gp=(Vp,Ep), power node viDegree for its neighbor node number, i.e. deg (vi),vi∈ Vp
Consider further that Holistic modeling.As shown in figure 1, the icon representation power plant with upper case character G.Usually, electric power is worked as During system jam, power plant has the function of cutting machine Reduction of Students' Study Load.Square, which represents, becomes electronic station, and red line is high voltage power transmission Line.The rectangle of black color represents the load being connected with substation.This means substation institute bringing onto load is all provided by the substation Electric power.Above-mentioned all elements constitute power system.Usually, the control centre is located at the electrical management in city or cities and towns Office.Router is directly or indirectly attached by optical fiber with control centre.
Model is continued to refine.Given Gp=(Vp,Ep) represent electric power networks, and Gc=(Vc,Ec) represent communication network.Often Individual figure has respective node set and link set, i.e., respective Vp,EpAnd Vc,Ec.Link E between two coupling networkspc= {(u,v)|u∈Vp,v∈Vc}.So, as shown in figure 1, G=(V can be usedp∪Vc,Ep∪Ec∪Epc) represent coupling network model.
For the detail to go into the question, key problem is paid close attention to, it is appropriate that further coupling network model is carried out Simplification.From the node in practical problem, electric power networks and power station, node truly has physical link to be connected with node, But these links are primarily to allow the behavior of power station control sub-station.Generally, these links are without communication lines By being only used as the single-point control in power station.The routing forwarding of electric power subnet also not responsible data.In actual environment, as long as hair Power station can receive the control information that communication subnet transmits, and the row of electric power subnet can be just controlled by the link between electric power networks For this operation is in most cases unidirectional.Power substations still will pass through if it is desired to inversely communicated with power station Communication subnet carries out data transmission.So the simplification carried out here is:The still controllable substation in power station, but in topological diagram The link in power station to subnet is hidden, to simplify topological analysis.I.e. electric power networks reduction is the collection of a series of power transformation website Close, Gp=(G, S1,S2,…,Sn)。
1.2 coupling models are analyzed
Driven based on above mentioned problem, investigated generation and communication process of the network failure in coupling network in detail below. First, explicitly defining for power system node failure and communication network node failure is provided respectively.Event is analyzed on this basis Hinder the necessary condition formed.It is necessary using destroying based on the generation for avoiding or suppressing failure and the consideration of optimization communication subnet The thought of condition, introduce virtual backbone network concept.Important node group backup optimisation strategy is proposed finally.
1.3 electric network fault treatment mechanisms are analyzed
In order to clearly analyze the failure in coupling network, first node failure is explicitly defined.
The PFP model node failure indicative function δ (x) after 1 optimization are defined to be expressed as:
2 correlation function corr are defined to be defined as,Similar has,
According to electric network fault treatment mechanism, under normal conditions, the processing procedure of power failure is as follows:When a substation When breaking down, such as S3, communication network can be sent to by identifying the information of the nodes break down.When communication subnet receives mark After the information of failure, power grid control instruction can be transmitted to by power plant by route.Power station Ge according to transmission come instruction control Website S processed3, so as to avoid further failure cascade phenomenon.
The accident analysis to sum up stated, following inference can be drawn:
The power failure treatment mechanism of inference 1 is when treatable failure problems occur, mutually according in coupling network, same A failure pair at least be present in the time.
The definition according to failure pair is proved, has been only possible to two kinds of situations.
The first situation, within the same time, there occurs failure for a node only in electric power networks.In this feelings Under shape, power node warning message can timely be delivered to communication network, due to not having any failure in communication network, then According to the normal treatment mechanism of failure, power plant Ge can be timely delivered to by cutting machine or getting rid of negative information, can solve this in time Problem.
Second case, in a communication network there occurs single-point node failure, do not have any failure in electric power networks Occur.So communication network management module can normally receive communication node warning message, and timely the failure is carried out Recover, so as to solve this problem.
To sum up, inference 1 can be obtained.
The critical path relevant with data transfer in mutually according to coupling network be present in inference 2.
Prove to assume failure to for fi,j=(Si,Rj) and Si∈Vp,Rj∈Vc.Failure handling mechanisms under from normal conditions Analysis is found out, can construct one and include node RiAnd to the path in power plant Ge communication path and electric power networks:
Due to there is common node Ge, two paths can be attached as a paths.
Investigate the path after this connection:
This paths can also be expanded:
Because assume faulty website SiDirectly with there is the router node R of eventjIt is connected, then expand path and show So set up.Another situation, corr { SjAnd RjIt is not attached to.Because communication network is UNICOM, then corr { SjCan be direct Communicated with power station.Solves this failure by the failure handling mechanisms of power network itself.So with condition fault produce and power network Can not be from processing contradiction.
To sum up, it can be deduced that this expands path.
Simultaneously as power station falls within grid nodes, so the communication node being connected with power station is labeled.Most Whole path expression is:
Final path expression is analyzed, from the angle of communication network, robustness of the following subpath to communication network It is very crucial, that is, the path related to data transfer to be drawn:
The minimal solution that proposition is saved in searching is NP-hard problems.So we are solved using didactic mode.Tool Body step is:
Step 1:Solve 2 dominant sets in communication subnet by didactic mode to generate, draw intermediate result set G1. Now G1 necessarily meets constraint 1, and also certain satisfaction constraint 1 is solved on the basis of G1.
Step2:Based on G1, (1,2)-CDS is calculated, is designated as G2.
Step3:Based on G2, comprehensive physical constraint, final result G is drawnx
But the G obtained on the basis of Step3x, in extreme circumstances, it is assumed that optical cable distribution is very uneven, then draws GxIt is extremely similar with initial power node set.For the consideration of such case, final result provides two set:G2 with Gx, and the two is contrasted in terms of algorithm evaluation.
Algorithm flow is as shown in Figure 2.
2. specific algorithm
2.1Step 1 is solved
Due to MCDS and NP-hard problems, so G1 solution is also didactic.
Input G=(Vp∪Vc,Ec∪Epc), the concrete thought of G1 Algorithm for Solving is:
1) according to EpcArrangement is ranked up, finds the routing node of all connectionsAnd according to its coupling with power node Number is closed to be ranked up.
2) power node set is investigated, is examined respectively with the degree being connected of communication node according to power node therein Consider:
If a) degree is 1, in this case, illustrate that the power node is only connected with a communication node, although the electric power Node can not meet 2 dominations, but be in order at the consideration of network-in-dialing, and the communication node for still connecting the power node adds G1, not so the power node will turn into isolated point.
If b) degree is 2, in this case, illustrate that power node is only connected with two communication nodes, then the electric power section Point meets that 2 dominate just, so two communication nodes that the power node is connected all add G1.
If c) degree is more than or equal to 3, in this case, illustrate that power node is connected with multiple communication nodes, this when Need to get off this set record, consider that process may need to use in follow-up network amendment and physical attribute.Again to this Individual more communication node set are investigated:
I) assume to have entered G1 in the presence of two communication nodes wherein, then current results collection meets that 2 dominate, and is not required to Add G1.
Ii) assume only one or enter G1 without existing node, then node is introduced into these and is ranked up, root According to its connected power node number, the addition G1 in selection ranking forefront.
2.2Step2 solve
G1 is inputted, (1,2)-CDS is solved, is designated as G2.The concrete thought of algorithm is:
1) G2=G1 is initialized.G2 concrete structure being investigated, if G2 has been connection, then G2 is just result set, Obtain a result.If G2 is not connected, then the number of record function T marks G2 connected subgraph, and indicate and each connect Logical subgraph, and its T value must be more than or equal to 2.
2) target of Algorithm for Solving is that T values are equal to 1, now illustrates that G2 has been connected.Specific mode is from communication subnet The node that middle selection most probable declines T values adds G2.Specifically selection algorithm is:
A) initialize VcNode state be labeled as 0, neighborhood RiIt is initialized as empty set.N set expressions G2 to be added Node, be initialized as empty set.
B) node in G2 is traveled through, G2 is divided into T connected subgraph, its internal node is labeled as 1,2 respectively, 3,, T.Next each connected subgraph is traveled through, to k-th of connected subgraph, its internal node is labeled as k, all neighbours The neighborhood of node adds k.
C) there will be the node of neighborhood to add N set.Node in gathering for N, arranged according to the degree of neighborhood Sequence, maximum node will be spent and add G2.Assuming that its value of the node of degree maximum of neighborhood is 1, then connected according to node logical Letter node number is ranked up.
2.3 connected dominating sets solve specific algorithm
First the record function in 1 in Step 2 is illustrated, upper section only illustrates thinking, carries out detailed card below It is bright.Supernumerary segment point algorithm in 2.c in Step 2 is exactly the 3.b steps for recording function.
The purpose of record function in 2 is to calculate G2 T values, and the connection situation of figure is illustrated.Algorithm stream Journey is as follows:
1) according to EcGenerate the annexation of each power node:DegC [i]=(Ci,m,N[Ci]) represent G2 in electric power Remaining power node number of the connection of node, and neighborhood, array sort from big to small according to m.
2) T=1, the State [i]=0 of each node are initialized.First m maximum of taking-up from DegC [] array Node Cj.According to DegC [j], by CjState function setups with N [j] are T.
3) DegC [] array is traveled through, for k-th of power node, according to CkState [k] and DegC [] handled:
If a) State [k]!=0, it has been the member of a connected set to illustrate the node, then State [N [Ck]]= State[k]。
If b) State [k]==0, consider State [N [Ck]] set:
I) set is all 0, then T adds 1, State [k]=T, State [N [Ck]]=State [k].
This situation is that front nodal point does not connect the node therewith with neighbor node.
Ii) set only has a kind of value x in addition to 0, then State [k]=x, State [N [Ck]]=State [k].This feelings Condition is that the node is connected with a connected subgraph.
Iii) set has t kind values in addition to 0, and t is more than or equal to 2 herein, and it is x to take minimum in these values, then T=T-t+ 1, State [k]=x, State [N [Ck]]=State [k].Again all values in State functions are equal to the set in addition to 0 Arbitrary value is assigned to min.This situation is that the node is connected with multiple connected subgraphs.
2.4 physical constraint decision algorithms
G2 is inputted, final result is solved, is designated as Gx.The concrete thought of algorithm is:
1) judge whether the routing node being connected with power station adds G2, do not add such as, be then put into.Because Under physical environment constraint, the communication node being joined directly together with power station is to the particularly significant of network.In most cases, These nodes have been added in G2.Because the routing node being joined directly together with power station is typically all in the key portion of network Point.
2) select the maximum node of remaining node moderate to add successively, until a little all add or current collection meets Target call.
2.5 important node group recognizers
According to thinking mentioned above, specific recognizer flow is as follows:
1) randomly generate node failure sequence and be applied on real topological environmental, to attacking calculating one each time It is secondary by discrete dot pattern into damage curve figure.
2) for each curve, it is once to be weighed caused by this failure sequence to obtain the maximum node of every section of tangent slope Node is wanted, and is added into set, and records an important node number of each node.
3) 200 random attacks have been used, have ultimately formed an important node number set.The more results of number in theory It is more accurate.Finally sorted from big to small according to number.
Important node group caused by final is a collating sequence, is carried out then in conjunction with the actual backup requirements of network specific Backup policy.
Figure below is emulation powerline network, and the purpose of experiment is that checking is carried out to model and and algorithm is tested Card.Fig. 3 illustrates to the topological structure of simple coupling network.Blue Squares represent power substations, and white square represents route Node, red line represent the routing link inside communication subnet, and remaining line represents what power node was connected with communication node Link.No. 1 is communication node with No. 2 in figure, has a lightguide cable link to be connected among both.No. 3 and No. 4 are power node, point It is not connected with No. 1 and No. 2.Obvious, the solution of No. 1 the problem of with No. 2 set formed being exactly final, it meets constraints 1 and 2.
On the basis of Fig. 3, emulation experiment employs the three-class power communication backbone network topology of certain province, as shown in figure 4, Wherein green square represents power station, and red square represents the routing node being connected with power station.By 28 power nodes, 28 Communication node imports, 1 to No. 28 expression routing node, 29 to No. 56 expression power nodes.
For the consideration of convenient analysis image topology, Fig. 4 does not show most of coupling between power node and communication node Link is closed, the topological structure and No. 39 power stations that illustrate only important communication network are connected with 7,11,15 signal communication nodes Blue path.Complete topological structure is shown in Fig. 5, can therefrom find out that concrete structure is more complicated.Due to before Algorithm all considers the optical cable attribute representated by the middle different links of electric power topological network from follow-up emulation, indicates in figure 6 Fiber cable type in electric power topological network, green line sign O cables, red line represent A cables.
Fig. 7 shows the returning result of G1 algorithms and the increase process of node.Fig. 8 shows that G1 topological structure can be seen Go out in figure to have three connected components, herein can image find out that the dominant set first obtained not necessarily meets connectedness.
Fig. 9 shows the returning result of G2 algorithms and the increase process of node.It can be seen that the connected component number that T is represented Reduce successively.Figure 10 shows G2 topologies, it can be seen that G2 is connected dominating set.
Figure 10 shows G2 topologies, it can be seen that G2 is connected dominating set.And for the life of next step algorithm Into consideration, the classification of each optical cable is gone out with different colour codes.
Figure 11 shows GxAs a result.It can be seen that final result, which is the topological structure, can not find such result, i.e., Make all topological nodes and optical cable all adding set, title-based condition can not be met.In this case, other can only be selected Method solves this problem as changed network structure and increase optical cable quantity.
Due to GxAs a result physical constraint condition can not be met, according to algorithm thinking, the identification of important node group is carried out to G2 Backup.Because the requirement of emulation experiment is, it is necessary to failure situation caused by observation analysis difference attack mode.Two kinds are incorporated herein to attack Hit mode:Grey is attacked and black attack.Black attack refers to random selection node attack, can trigger power telecom network node Random fault, it is generally used to simulate the random failure under the conditions of random attack and normal operation circumstances non-attack.Grey attack refers to Attacker is after a part of network topology situation is understood, such as degree distribution and partially complete network structure, and weight is carried out to target Object of attack is selected to carry out fault attacks successively again after the judgement for the property wanted.In actual environment, both attack patterns are very It is common.
Due to the requirement of simulation comparison experiment, Figure 12 is listed according to degree sort algorithm section and important node group recognizer Arithmetic result.Performance situation in Figure 13,14 for two kinds of backup policy when black and grey fault attacks occur.Figure 15 is nothing Performance situation of the backup policy under attack.
Finally illustrate, preferred embodiment above is merely illustrative of the technical solution of the present invention and unrestricted, although logical Cross above preferred embodiment the present invention is described in detail, it is to be understood by those skilled in the art that can be Various changes are made to it in form and in details, without departing from claims of the present invention limited range.

Claims (10)

  1. A kind of 1. communication network failure cascade venture influence analysis method of attack resistance, it is characterised in that:This method includes following step Suddenly:
    S1:Establish coupling network model:Power network model is established, establishes overall network model, refines overall network model;
    S2:Analyze coupling network model:Explicitly defining for power system node failure and communication network node failure is provided respectively; The necessary condition that failure is formed is analyzed on this basis;Based on examining for the generation and optimization communication subnet for avoiding or suppressing failure Consider, using the thought for destroying necessary condition, introduce virtual backbone network concept;Finally propose important node group backup optimization plan Slightly.
  2. 2. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 1, its feature exist In:The node failure is expressly defined as:
    (1) power failure after optimizing propagates (Power Fault Propagation, PFP) model node failure indicative function δ (x) it is expressed as:
    (2) correlation function corr is defined as: Wherein vj、viFor power node, Vc、VpFor node set, EpcFor two Link between coupling network, Epc=(u, v) | u ∈ Vp,v∈Vc, wherein u is power node.
  3. 3. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 1, its feature exist In:The necessary condition that the failure is formed is:
    (1) power failure treatment mechanism is when treatable failure problems occur, mutually according in coupling network, the same time extremely A failure pair less be present;
    (2) critical path relevant with data transfer in mutually according to coupling network be present.
  4. 4. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 3, its feature exist In:The critical path is:Wherein SjFor power station, Ri、Ri+1For section Point, Ge are power plant.
  5. 5. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 4, its feature exist In:The critical path is solved:
    S201:Solve 2 dominant sets in communication subnet by didactic mode to generate, draw intermediate result set G1, now G1 It is certain to meet constraint 1, i.e., when the not treatable failure problems of power failure treatment mechanism occur, mutually according in coupling network, At least there is a failure pair in the same time, also certain satisfaction constraint 1 is solved on the basis of G1;
    S202:Based on G1, (1,2)-CDS is calculated, is designated as G2;I.e. 1 connection 2 dominate set, minimum connected dominating set be (1,1)- CDS;
    S203:Based on G2, comprehensive physical constraint, final result G is drawnx
    Final result provides two set:G2 and Gx, and the two is contrasted in terms of algorithm evaluation.
  6. 6. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 5, its feature exist In:The step S201 is specially:
    Input G=(Vp∪Vc,Ec∪Epc), solve G1:
    (1) according to EpcArrangement is ranked up, finds the routing node of all connectionsAnd according to its coupling time with power node Number is ranked up;
    (2) power node set is investigated, is accounted for respectively with the degree being connected of communication node according to power node therein:
    If a) degree is 1, illustrate that the power node is only connected with a communication node, for the consideration of network-in-dialing, by the electricity The communication node of power node connection adds G1, avoids the power node from turning into isolated point;
    If b) degree is 2, illustrate that power node is only connected with two communication nodes, two communication sections that the power node is connected Point all adds G1;
    If c) degree is more than or equal to 3, illustrates that power node is connected with multiple communication nodes, the power node set record is got off And investigated:
    I) assume to have entered G1 in the presence of two communication nodes wherein, then current results collection meets that 2 dominate, it is not necessary to again Add G1;
    Ii) assume only one or enter G1 without existing node, then node is introduced into these and is ranked up, according to it Connected power node number, select the addition G1 in ranking forefront.
  7. 7. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 6, its feature exist In:The step S202 is specially:
    G1 is inputted, (1,2)-CDS is solved, is designated as G2:
    (1) G2=G1 is initialized;G2 concrete structure is investigated, if G2 has been connection, then G2 is result set;If G2 Do not connect, then the number of record function T marks G2 connected subgraph, and each connected subgraph is indicated, and its T value must be more than Equal to 2;
    (2) target of Algorithm for Solving is that T values are equal to 1, now illustrates that G2 has been connected;Most probable is selected to make T from communication subnet The node that value declines adds G2;Specifically selection algorithm is:
    A) initialize VcNode state be labeled as 0, neighborhood RiIt is initialized as empty set;N set expressions G2 to be added section Point, is initialized as empty set;
    B) node in G2 is traveled through, G2 is divided into T connected subgraph, its internal node is labeled as 1,2,3 respectively,, T;Each connected subgraph is traveled through, to k-th of connected subgraph, its internal node is labeled as k, the neighbours of all neighbor nodes Set adds k;
    C) there will be the node of neighborhood to add N set;Node in gathering for N, sorted according to the degree of neighborhood, will The maximum node of degree adds G2;Assuming that its value of the node of degree maximum of neighborhood is 1, then the communication node connected according to node Number is ranked up.
  8. 8. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 7, its feature exist In:The step (1) is specially:
    1) according to EcGenerate the annexation of each power node:DegC [i]=(Ci,m,N[Ci]), wherein CiFor node, m is represented Remaining power node number of the connection of power node in G2, N [Ci] neighborhood is represented, array sorts from big to small according to m;
    2) T=1, the State [i]=0 of each node are initialized;The maximum nodes of first m are taken out from DegC [] array Cj;According to DegC [j], by CjState function setups with N [j] are T;
    3) DegC [] array is traveled through, for k-th of power node, according to CkState [k] and DegC [] handled:
    If a) State [k]!=0, it has been the member of a connected set to illustrate the node, then State [N [Ck]]=State [k];
    If b) State [k]==0, consider State [N [Ck]] set:
    I) set is all 0, then T adds 1, State [k]=T, State [N [Ck]]=State [k];This situation is the node and neighbour Occupying node, front nodal point does not connect therewith;
    Ii) set only has a kind of value x in addition to 0, then State [k]=x, State [N [Ck]]=State [k];This situation is The node is connected with a connected subgraph;
    Iii) set has t kind values in addition to 0, and t is more than or equal to 2 herein, and it is x to take minimum in these values, then T=T-t+1, State [k]=x, State [N [Ck]]=State [k];Again all values in State functions are equal to the set appointing in addition to 0 Meaning value is assigned to min;This situation is that the node is connected with multiple connected subgraphs.
  9. 9. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 7 or 8, its feature It is:The GxSpecific algorithm is:
    G2 is inputted, final result is solved, is designated as Gx
    (1) judge whether the routing node being connected with power station adds G2, do not add such as, be then put into;
    (2) select the maximum node of remaining node moderate to add successively, until a little all add or current collection meets mesh Mark requires.
  10. 10. a kind of communication network failure cascade venture influence analysis method of attack resistance according to claim 1, its feature exist In:The important node group backup optimisation strategy is:
    (1) randomly generate node failure sequence and be applied on real topological environmental, calculated once attacking each time By discrete dot pattern into damage curve figure;
    (2) for each curve, it is once important section caused by the failure sequence to obtain the maximum node of every section of tangent slope Point, and set is added into, and record an important node number of each node;
    (3) using repeatedly random attack, an important node number set is formed, the random more results of number of times of attack are more accurate, Finally sorted from big to small according to random number of times of attack;
    (4) the actual backup requirements of network are combined and carry out specific backup policy.
CN201710847598.1A 2017-09-19 2017-09-19 A kind of communication network failure cascade venture influence analysis method of attack resistance Pending CN107769962A (en)

Priority Applications (1)

Application Number Priority Date Filing Date Title
CN201710847598.1A CN107769962A (en) 2017-09-19 2017-09-19 A kind of communication network failure cascade venture influence analysis method of attack resistance

Applications Claiming Priority (1)

Application Number Priority Date Filing Date Title
CN201710847598.1A CN107769962A (en) 2017-09-19 2017-09-19 A kind of communication network failure cascade venture influence analysis method of attack resistance

Publications (1)

Publication Number Publication Date
CN107769962A true CN107769962A (en) 2018-03-06

Family

ID=61265989

Family Applications (1)

Application Number Title Priority Date Filing Date
CN201710847598.1A Pending CN107769962A (en) 2017-09-19 2017-09-19 A kind of communication network failure cascade venture influence analysis method of attack resistance

Country Status (1)

Country Link
CN (1) CN107769962A (en)

Cited By (8)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858771A (en) * 2019-01-02 2019-06-07 中国电力科学研究院有限公司 A kind of intelligent substation interoperability grade evaluation method and system
CN110213087A (en) * 2019-05-16 2019-09-06 北京航空航天大学 A kind of complication system Fault Locating Method based on dynamic multilayer coupling network
CN110493260A (en) * 2019-09-12 2019-11-22 贵州电网有限责任公司 A kind of network flood model attack detection method
CN111435478A (en) * 2019-01-11 2020-07-21 国网河北省电力有限公司 Graph theory-based electric power CPS structure communication vulnerability assessment method
CN113452673A (en) * 2021-05-18 2021-09-28 广西电网有限责任公司电力科学研究院 Network attack damage degree quantification method for power system
CN114039838A (en) * 2021-12-24 2022-02-11 国网浙江省电力有限公司信息通信分公司 Power communication network fault analysis method based on maximum disjoint double routes and related equipment
CN114268576A (en) * 2021-12-24 2022-04-01 国网浙江省电力有限公司信息通信分公司 Method for determining interlock fault survival parameters of power CPS and related equipment
CN115102798A (en) * 2022-06-20 2022-09-23 镇江港务集团有限公司 Emergency data backup system for equipment power failure

Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090323539A1 (en) * 2008-06-30 2009-12-31 Dazhi Wang Reliability estimation methods for large networked systems
CN103957032A (en) * 2014-04-17 2014-07-30 华北电力大学 Load redistribution method for electric power coupling network to resist cascade failure
CN104093166A (en) * 2014-07-08 2014-10-08 南京信息工程大学 Wireless sensor network connection recovery method based on minimum movement of nodes
CN105656198A (en) * 2015-12-29 2016-06-08 中国电力科学研究院 Electric power communication network redundant path strategy acquiring method

Patent Citations (4)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
US20090323539A1 (en) * 2008-06-30 2009-12-31 Dazhi Wang Reliability estimation methods for large networked systems
CN103957032A (en) * 2014-04-17 2014-07-30 华北电力大学 Load redistribution method for electric power coupling network to resist cascade failure
CN104093166A (en) * 2014-07-08 2014-10-08 南京信息工程大学 Wireless sensor network connection recovery method based on minimum movement of nodes
CN105656198A (en) * 2015-12-29 2016-06-08 中国电力科学研究院 Electric power communication network redundant path strategy acquiring method

Non-Patent Citations (4)

* Cited by examiner, † Cited by third party
Title
FAHONG YU等: "Critical node identification for complex network based on a novel minimum connected dominating set", 《SPRINGERLINK》 *
任晓龙等: "网络重要节点排序方法综述", 《软件学报》 *
张保健: "复杂大电网中基于级联失效的双网可靠性分析", 《中国优秀硕士学位论文全文数据库 工程科技Ⅱ辑》 *
郑婵等: "无线传感器网络中d-Hop 2-连通容错支配集的分布式构造算法", 《传感技术学报》 *

Cited By (13)

* Cited by examiner, † Cited by third party
Publication number Priority date Publication date Assignee Title
CN109858771A (en) * 2019-01-02 2019-06-07 中国电力科学研究院有限公司 A kind of intelligent substation interoperability grade evaluation method and system
CN111435478A (en) * 2019-01-11 2020-07-21 国网河北省电力有限公司 Graph theory-based electric power CPS structure communication vulnerability assessment method
CN110213087A (en) * 2019-05-16 2019-09-06 北京航空航天大学 A kind of complication system Fault Locating Method based on dynamic multilayer coupling network
CN110213087B (en) * 2019-05-16 2020-08-25 北京航空航天大学 Complex system fault positioning method based on dynamic multilayer coupling network
CN110493260A (en) * 2019-09-12 2019-11-22 贵州电网有限责任公司 A kind of network flood model attack detection method
CN113452673B (en) * 2021-05-18 2022-05-03 广西电网有限责任公司电力科学研究院 Network attack damage degree quantification method for power system
CN113452673A (en) * 2021-05-18 2021-09-28 广西电网有限责任公司电力科学研究院 Network attack damage degree quantification method for power system
CN114039838A (en) * 2021-12-24 2022-02-11 国网浙江省电力有限公司信息通信分公司 Power communication network fault analysis method based on maximum disjoint double routes and related equipment
CN114268576A (en) * 2021-12-24 2022-04-01 国网浙江省电力有限公司信息通信分公司 Method for determining interlock fault survival parameters of power CPS and related equipment
CN114268576B (en) * 2021-12-24 2023-10-03 国网浙江省电力有限公司信息通信分公司 Method for determining survival parameters of interlocking faults of power CPS and related equipment
CN114039838B (en) * 2021-12-24 2024-02-02 国网浙江省电力有限公司信息通信分公司 Power communication network fault analysis method and related equipment based on maximum disjoint double routes
CN115102798A (en) * 2022-06-20 2022-09-23 镇江港务集团有限公司 Emergency data backup system for equipment power failure
CN115102798B (en) * 2022-06-20 2023-08-22 镇江港务集团有限公司 Emergency data backup system for equipment outage

Similar Documents

Publication Publication Date Title
CN107769962A (en) A kind of communication network failure cascade venture influence analysis method of attack resistance
CN109191326B (en) Attack party view angle-based network attack risk assessment method for CPS (control Performance Standard) mutual existence system of power distribution network
CN106487005B (en) A kind of Electric power network planning method considering T-D tariff
CN106253270B (en) Electric system vulnerable line identifying method and system
CN110300018B (en) Object-oriented hierarchical modeling method for power grid information physical system
Shahraeini et al. Comparison between communication infrastructures of centralized and decentralized wide area measurement systems
CN103001328B (en) Fault diagnosis and assessment method of intelligent substation
CN106789190A (en) A kind of power telecom network vulnerability assessment and routing optimization method
CN107623319A (en) A kind of power network critical circuits discrimination method based on more evaluation indexes
CN113705085B (en) Intelligent power grid multi-level structure modeling and risk assessment method
CN111106999A (en) IP-optical network communication service joint distribution method and device
CN112633649A (en) Power grid multi-attribute important node evaluation and planning method
CN103151774B (en) A kind of small-world power grid cascading failure suppressing method
CN112086958B (en) Power transmission network extension planning method based on multi-step backtracking reinforcement learning algorithm
CN110503136A (en) Platform area line loss exception analysis method, computer readable storage medium and terminal device
CN107025524B (en) Equipment overload auxiliary decision-making method considering load power supply reliability requirement
CN104363104B (en) A kind of magnanimity multivariate data battle state display System and method for of Users ' Need-oriented
CN107092984A (en) A kind of network function end node propagation prediction method based on cascading failure
CN113722868B (en) Multi-index power grid node vulnerability assessment method integrating structural hole characteristics
CN107622360A (en) A kind of critical circuits recognition methods for considering subjective and objective factor
CN106067074A (en) A kind of by optimizing the method that the on off state of link promotes network system robustness
CN106779269A (en) Reliability assessment method for secondary system of intelligent substation based on VC Method
CN105933145A (en) Adjacent matrix based dependent network system and frangibility detection method thereof
CN118114989A (en) Industrial chain risk assessment method of graph neural network
CN117057623A (en) Comprehensive power grid safety optimization scheduling method, device and storage medium

Legal Events

Date Code Title Description
PB01 Publication
PB01 Publication
SE01 Entry into force of request for substantive examination
SE01 Entry into force of request for substantive examination
RJ01 Rejection of invention patent application after publication
RJ01 Rejection of invention patent application after publication

Application publication date: 20180306