Method and system for improving resilience of power grid information-physical fusion system

Abstract

The present invention relates to a method and system for improving the resilience of a power grid information-physical fusion system, the method comprising: pre-setting multiple influencing factors for improving resilience for the power grid information-physical fusion system; using an expert method to score the mutual influence relationship between the influencing factors, and obtaining a pairwise comparison matrix; obtaining papers related to improving system resilience, constructing a knowledge graph of resilience-improving factors, using the TransH algorithm to calculate the distance between the influencing factors based on the knowledge graph, and forming a correction matrix; fusing the pairwise comparison matrix and the correction matrix to obtain a final pairwise comparison matrix; calculating the centrality and causality of the final pairwise comparison matrix, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system. Compared with the prior art, the present invention can greatly improve the accuracy of the analysis results, reduce the analysis cost, and fully improve the resilience of the UGCPS.

Description

Technical Field

[0001] The present invention relates to the field of power grid technology, and in particular to a method and system for improving the resilience of a power grid information-physical fusion system. Background Art

[0002] Building a new power system with large-scale access to new energy is a realistic need for clean and low-carbon energy transformation. In addition to the clean and low-carbon transformation of the power system, energy security is also one of the important tasks of the power system. However, there is a certain conflict between the two goals. The access of large-scale new energy has brought volatility and intermittency to energy supply, affecting the goal of energy security. And security requires setting a certain capacity of thermal power units as backup, which will increase carbon dioxide emissions and hinder low-carbon transformation.

[0003] Digital technology can accurately predict the power generation of new energy and the power load on the user side, and accurately calculate the power gap to reduce the capacity of the standby unit, so as to reduce the carbon emissions caused by the standby after the access of new energy, and thus solve the contradiction between low-carbon transformation and safe supply. The effective implementation of digital technology depends on the information interaction between intelligent electrical equipment and computing power hubs in the power system. The communication network within the new power system is responsible for this kind of information interaction. The communication network and the power grid together constitute the cyber-physical system (CPS). If the power CPS fails, the digital technology cannot be effectively applied, which will cause problems in the safe supply of the power grid or hinder the clean and low-carbon transformation of the power grid.

[0004] The power grid is the key link of low carbon and supply guarantee of the new power system. The Urbangrid cyber-physical system (UGCPS) of the power grid is related to the application of digital technology in the power grid, and also affects the safe supply guarantee and low-carbon transformation of the power grid. UGCPS urgently needs to improve its resilience to ensure the intelligent operation of the power grid and complete the safe supply guarantee and low-carbon transformation. The invention with the authorization announcement number CN118333410A and the patent number ZL202410765123 discloses the power grid risk transmission analysis, determination, prevention and control methods, systems, equipment and media, which fully integrates the risk transmission information, realizes the effective characterization and analysis of the dynamic evolution process of the target power grid system security risk, and identifies key nodes and key parameters. However, for the identified key parameters, the current power grid operators are not clear about how to further focus on which more important factors to improve the resilience of UGCPS. Summary of the invention

[0005] The purpose of the present invention is to provide a method and system for improving the resilience of a power grid information-physical fusion system in order to overcome the defect of the above-mentioned prior art that it is unclear which factors should be focused on to improve the resilience of UGCPS.

[0006] The purpose of the present invention can be achieved by the following technical solutions:

[0007] A method for improving the resilience of a power grid information-physical fusion system comprises the following steps:

[0008] Set up multiple factors that influence resilience improvement for the power grid cyber-physical fusion system;

[0009] The expert method is used to score the mutual influence relationship between two influencing factors to obtain a pairwise comparison matrix;

[0010] Obtain papers related to improving the resilience of power grid cyber-physical fusion systems, construct a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between influencing factors based on the knowledge graph to form a correction matrix;

[0011] The pairwise comparison matrix and the correction matrix are combined to obtain a final pairwise comparison matrix;

[0012] The centrality and causality of the final pairwise comparison matrix are calculated, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system.

[0013] Furthermore, the power grid cyber-physical fusion system is divided into a multi-layer logical architecture, including an information network, a power grid and a coupling part;

[0014] The influencing factors corresponding to the information network include the accuracy of information network operation and maintenance, the reliability of information transmission channels, the reliability of information collection systems and the defense capability against network attacks;

[0015] The influencing factors corresponding to the power grid include volatility regulation of high-proportion new energy, multi-energy complementary control, electric vehicle access regulation, user-side management, control of renewable energy power generation and control of energy-consuming terminals;

[0016] The coupling part includes the ability of power electronic equipment to resist extreme events, the grid structure of the power grid, the stability of the interaction between the power grid and the information network, and the monitoring and control of chain failures of the energy system, transportation network and power grid.

[0017] Furthermore, the expression of the pairwise comparison matrix obtained by expert scoring is:

[0018] ;

[0019] In the formula, is the pairwise comparison matrix given by the k-th decision expert for all influencing factors, is the influencing factor of the k-th decision expert and The score of pairwise comparison is n, where n is the total number of influencing factors.

[0020] Furthermore, the knowledge graph is in the form of a triple (h, t, r), where: Represents an entity, To express a relationship;

[0021] The objective function of the TransH algorithm in calculating the distance between two influencing factors based on the knowledge graph includes:

[0022] ;

[0023] ;

[0024] ;

[0025] In the formula, is the objective function of the TransH algorithm, and is the score function of the TransH model, is a constant used to represent the spacing between positive and negative samples. is h in the hyperplane The projection of for The transposed matrix of For t in the hyperplane The projection of is the set of positive triples, is the set of negative triples;

[0026] The expression of the constraint condition satisfied by the objective function of the TransH algorithm is:

[0027] ;

[0028] In the formula, is the set of real numbers.

[0029] Furthermore, the fusion process of the pairwise comparison matrix and the correction matrix includes the following steps:

[0030] The correction matrix is ​​combined with the pairwise comparison matrix to construct an undirected weighted complex network G, wherein the complex network G has m nodes, which represent decision experts; The weight of each edge represents the consistency among experts;

[0031] Perform expert clustering on the complex network G and divide the decision experts in the complex network G into c Clusters of classes;

[0032] Calculating the network cohesion according to the complex network G after clustering, and calculating the cluster consistency according to the correction matrix;

[0033] Calculating expert importance based on network cohesion, and calculating cluster importance based on the network cohesion and cluster consistency;

[0034] Expert weights are calculated based on expert importance, and cluster weights are calculated based on cluster importance and cluster consistency;

[0035] According to the expert weights and cluster weights, the NDWAA operator value of the complex network G is calculated to obtain the final pairwise comparison matrix.

[0036] Furthermore, the process of performing expert clustering on the complex network G includes:

[0037] S101: constructing an original correlation coefficient matrix according to the complex network G ;

[0038] S102: Calculate a new correlation coefficient matrix based on CONCOR transformation , and repeat this step to get a new correlation coefficient matrix ; Until the correlation coefficient matrix is ​​obtained ,and ;

[0039] S103: Correlation coefficient matrix The row elements in are divided into two clusters, where For 1 cluster, for another cluster;

[0040] S104: According to the cluster division result, the decision experts are divided into two categories, and steps S101 to S104 are repeated until the number of elements in each subclass is satisfy So far, the decision experts are finally divided into c clusters.

[0041] Furthermore, the calculation expression of the network cohesion is:

[0042] ;

[0043] In the formula, is the network cohesion, is the total number of influencing factors, is the scoring result between the i-th influencing factor and the j-th influencing factor;

[0044] The calculation expression of the cluster consistency is:

[0045] ;

[0046] In the formula, is the correction matrix, For cluster consistency, is the total number of elements in the correction matrix, is the correction result between the i-th element and the j-th element in the correction matrix;

[0047] The calculation expression of the expert importance is:

[0048] ;

[0049] In the formula, For Node The importance of experts Delete nodes from complex network G The network cohesion afterwards;

[0050] The calculation expression of the expert weight is:

[0051] ;

[0052] In the formula, For Node Expert weights;

[0053] The calculation expression of the cluster importance is:

[0054] ;

[0055] In the formula, For cluster The importance of the cluster, Delete a cluster for a complex network G Network consistency after all experts in;

[0056] The calculation expression of the cluster weight is:

[0057] ;

[0058] In the formula, For cluster The cluster weight of .

[0059] Furthermore, the calculation expression of the NDWAA operator value is:

[0060] ;

[0061] In the formula, is the NDWAA operator value, For NDWAA synthesis calculation, is the value after standardization.

[0062] Furthermore, the calculation expression of the centrality is:

[0063] ;

[0064] ;

[0065] ;

[0066] ;

[0067] ;

[0068] ;

[0069] ;

[0070] ;

[0071] In the formula, is the standardized distance matrix, is the distance value between influencing factor i and influencing factor k; is the total number of influencing factors, is the overall relationship matrix, is the identity matrix, the superscript i is the power of the matrix, is the influence degree of influencing factor i, is the degree of influence of influencing factor i, is the relationship between influencing factor i and influencing factor j, is the centrality of influencing factor i, is the centrality;

[0072] The calculation expression of the cause degree is:

[0073] ;

[0074] ;

[0075] In the formula, For the reason, is the causal degree of influencing factor i.

[0076] The present invention also provides a resilience enhancement system for a power grid information-physical fusion system, comprising:

[0077] An influencing factor setting module is used to set multiple influencing factors for improving resilience of the power grid cyber-physical fusion system;

[0078] The expert method module is used to score the mutual influence relationship between two influencing factors using the expert method to obtain a pairwise comparison matrix;

[0079] A correction matrix calculation module is used to obtain papers related to improving the resilience of power grid cyber-physical fusion systems, construct a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between influencing factors based on the knowledge graph to form a correction matrix.

[0080] A fusion module, used for fusing the pairwise comparison matrix and the correction matrix to obtain a final pairwise comparison matrix;

[0081] The importance calculation module is used to calculate the centrality and causality of the final pairwise comparison matrix, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and to determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system.

[0082] Furthermore, the power grid cyber-physical fusion system is divided into a multi-layer logical architecture, including an information network, a power grid and a coupling part;

[0083] The influencing factors corresponding to the information network include the accuracy of information network operation and maintenance, the reliability of information transmission channels, the reliability of information collection systems and the defense capability against network attacks;

[0084] The influencing factors corresponding to the power grid include volatility regulation of high-proportion new energy, multi-energy complementary control, electric vehicle access regulation, user-side management, control of renewable energy power generation and control of energy-consuming terminals;

[0085] The coupling part includes the ability of power electronic equipment to resist extreme events, the grid structure of the power grid, the stability of the interaction between the power grid and the information network, and the monitoring and control of chain failures of the energy system, transportation network and power grid.

[0086] Furthermore, the expression of the pairwise comparison matrix obtained by expert scoring is:

[0087] ;

[0088] In the formula, is the pairwise comparison matrix given by the k-th decision expert for all influencing factors, is the influencing factor of the k-th decision expert and The score of pairwise comparison is n, where n is the total number of influencing factors.

[0089] Furthermore, the knowledge graph is in the form of a triple (h, t, r), where: Represents an entity, To express a relationship;

[0090] The objective function of the TransH algorithm in calculating the distance between two influencing factors based on the knowledge graph includes:

[0091] ;

[0092] ;

[0093] ;

[0094] In the formula, is the objective function of the TransH algorithm, and is the score function of the TransH model, is a constant used to represent the spacing between positive and negative samples. is h in the hyperplane The projection of for The transposed matrix of For t in the hyperplane The projection of is the set of positive triples, is the set of negative triples;

[0095] The expression of the constraint condition satisfied by the objective function of the TransH algorithm is:

[0096] ;

[0097] In the formula, is the set of real numbers.

[0098] Furthermore, in the fusion module, the fusion process of the pairwise comparison matrix and the correction matrix includes the following steps:

[0099] The correction matrix is ​​combined with the pairwise comparison matrix to construct an undirected weighted complex network G, wherein the complex network G has m nodes, each node representing a decision expert; the complex network G has The weight of each edge represents the consistency among experts;

[0100] Perform expert clustering on the complex network G and divide the decision experts in the complex network G into c Clusters of classes;

[0101] Calculating the network cohesion according to the complex network G after clustering, and calculating the cluster consistency according to the correction matrix;

[0102] Calculating expert importance based on network cohesion, and calculating cluster importance based on the network cohesion and cluster consistency;

[0103] Expert weights are calculated based on expert importance, and cluster weights are calculated based on cluster importance and cluster consistency;

[0104] According to the expert weights and cluster weights, the NDWAA operator value of the complex network G is calculated to obtain the final pairwise comparison matrix.

[0105] Furthermore, the process of performing expert clustering on the complex network G includes:

[0106] S101: constructing an original correlation coefficient matrix according to the complex network G ;

[0107] S102: Calculate a new correlation coefficient matrix based on CONCOR transformation , and repeat this step to get a new correlation coefficient matrix ; Until the correlation coefficient matrix is ​​obtained ,and ;

[0108] S103: Correlation coefficient matrix The row elements in are divided into two clusters, where For 1 cluster, for another cluster;

[0109] S104: According to the cluster division result, the decision experts are divided into two categories, and steps S101 to S104 are repeated until the number of elements in each subclass is satisfy So far, the decision experts are finally divided into c clusters.

[0110] Furthermore, the calculation expression of the network cohesion is:

[0111] ;

[0112] In the formula, is the network cohesion, is the total number of influencing factors, is the scoring result between the i-th influencing factor and the j-th influencing factor;

[0113] The calculation expression of the cluster consistency is:

[0114] ;

[0115] In the formula, is the correction matrix, For cluster consistency, is the total number of elements in the correction matrix, is the correction result between the i-th element and the j-th element in the correction matrix;

[0116] The calculation expression of the expert importance is:

[0117] ;

[0118] In the formula, For Node The importance of experts Delete nodes from complex network G The network cohesion afterwards;

[0119] The calculation expression of the expert weight is:

[0120] ;

[0121] In the formula, For Node Expert weights;

[0122] The calculation expression of the cluster importance is:

[0123] ;

[0124] In the formula, For cluster The importance of the cluster, Delete a cluster for a complex network G Network consistency after all experts in;

[0125] The calculation expression of the cluster weight is:

[0126] ;

[0127] In the formula, For cluster The cluster weight of .

[0128] Furthermore, the calculation expression of the NDWAA operator value is:

[0129] ;

[0130] In the formula, is the NDWAA operator value, For NDWAA synthesis calculation, is the value after standardization.

[0131] Furthermore, the calculation expression of the centrality is:

[0132] ;

[0133] ;

[0134] ;

[0135] ;

[0136] ;

[0137] ;

[0138] ;

[0139] ;

[0140] In the formula, is the standardized distance matrix, is the distance value between influencing factor i and influencing factor k; is the total number of influencing factors, is the overall relationship matrix, is the identity matrix, the superscript i is the power of the matrix, is the influence degree of influencing factor i, is the degree of influence of influencing factor i, is the relationship between influencing factor i and influencing factor j, is the centrality of influencing factor i, is the centrality;

[0141] The calculation expression of the cause degree is:

[0142] ;

[0143] ;

[0144] In the formula, For the reason, is the causal degree of influencing factor i.

[0145] The present invention also provides a device for improving the resilience of a power grid information-physical fusion system, comprising a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of the method described above.

[0146] The present invention also provides a computer-readable storage medium, on which a computer program is stored, and the computer program is executed by a processor to perform the above method.

[0147] Compared with the prior art, the present invention has the following advantages:

[0148] (1) This paper constructs a set of index systems that can accurately describe the improvement of UGCPS resilience. This index system can effectively improve the resilience of UGCPS and has 14 resilience improvement factors at three levels. On the one hand, the mutual influence relationship between factors is scored pairwise by the expert method. On the other hand, the DEMATEL method of text embedding is implemented with the help of TransH to obtain the correction matrix. Finally, the importance and causal relationship of the UGCPS resilience improvement factors are analyzed to obtain the relevant resilience improvement factors.

[0149] The above scheme of the present invention can greatly improve the accuracy of the analysis results and reduce the analysis cost; it is helpful to improve the construction and operation efficiency of UGCPS and reduce the cost investment and risk probability of UGCPS.

[0150] (2) Compared with other methods, the proposed method fully considers the subjective judgment of 9 experts and the knowledge information of 299 papers with the help of complex network theory, and can more accurately characterize the importance ranking and mutual influence relationship of 14 resilience improvements. BRIEF DESCRIPTION OF THE DRAWINGS

[0151] Figure 1 A schematic flow chart of a method for improving the resilience of a power grid information-physical fusion system provided in an embodiment of the present invention;

[0152] Figure 2 A schematic diagram of a triple entity provided in an embodiment of the present invention;

[0153] Figure 3 A schematic diagram of a TransH provided in an embodiment of the present invention;

[0154] Figure 4 A pairwise comparison matrix fusion flow chart provided in an embodiment of the present invention;

[0155] Figure 5 A two-dimensional vectorized display of toughness improvement factors provided in an embodiment of the present invention. DETAILED DESCRIPTION

[0156] In order to make the purpose, technical solutions and advantages of the embodiments of the present invention clearer, the technical solutions in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Generally, the components of the embodiments of the present invention described and shown in the drawings here can be arranged and designed in various different configurations.

[0157] Therefore, the following detailed description of the embodiments of the present invention provided in the accompanying drawings is not intended to limit the scope of the invention claimed for protection, but merely represents selected embodiments of the present invention. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work are within the scope of protection of the present invention.

[0158] It should be noted that similar reference numerals and letters denote similar items in the following drawings, and therefore, once an item is defined in one drawing, further definition and explanation thereof is not required in subsequent drawings.

[0159] Example 1

[0160] like Figure 1 As shown, this embodiment provides a method for improving the resilience of a power grid information-physical fusion system, including the following steps:

[0161] Pre-set multiple factors affecting resilience improvement for the power grid cyber-physical fusion system;

[0162] S1: Use the expert method to score the mutual influence relationship between the two influencing factors and obtain a pairwise comparison matrix;

[0163] S2: Obtain papers related to improving the resilience of power grid cyber-physical fusion systems, construct a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between influencing factors based on the knowledge graph to form a correction matrix.

[0164] S3: Fuse the pairwise comparison matrix and the correction matrix to obtain the final pairwise comparison matrix;

[0165] S4: Calculate the centrality and causality of the final pairwise comparison matrix, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system.

[0166] The following is a detailed description of each step:

[0167] 1. Mechanism for improving the resilience of power grid cyber-physical fusion systems

[0168] UGCPS is a comprehensive system that integrates information communication and Internet technologies to achieve optimized dispatch of clean energy within the city. The deep integration of information technology and power grid is a major feature of UGCPS. Information technology has spawned various data applications and services, and has also built the power grid into an energy information network platform with strong multi-device interactivity and smooth information flow. Based on this feature, UGCPS can be divided into a three-layer logical architecture, as shown in the information network, power grid and coupling part.

[0169] The power grid is a dispatching network for power supply and consumption, and is the core and foundation of UGCPS. At present, the power grid is constantly integrating renewable energy and clean energy such as electric vehicles. The power grid is no longer a single physical carrier for power transmission and distribution. The power grid is gradually expanding to run multiple tasks in parallel, including optimizing energy allocation, ensuring energy supply, and operating the power market.

[0170] The information network is the information center of UGCPS. Its core business is to ensure the real-time stability of power supply and to reduce the carbon emissions of the power grid as much as possible with the help of its own dispatching services.

[0171] The coupling department is the place where the UGCPS information network and the power grid are deeply coupled, including three parts: data collection, data analysis, and data decision-making. Data collection is a process that obtains real-time data through sensors. After collection, the data is transmitted to the data center through the communication network. Data analysis is a process that mines and visualizes the value contained in the data through a variety of algorithms. Data decision-making is a process that relies on the results of data analysis to provide suggestions for the next step of operation.

[0172] Based on the above analysis of the functions and characteristics of UGCPS, this embodiment constructs a set of influencing factors for improving UGCPS resilience, which includes 14 indicators, for three important subjects. The specific factors are shown in Table 1.

[0173] Table 1 Factors that improve the toughness of UGCPS

[0174]

[0175] The improvement factors of the information network include the accuracy of information network operation and maintenance, the reliability of information transmission channels, the reliability of information collection systems, and the defense capabilities against network attacks. This level mainly improves the resilience of UGCPS from the information collection and transmission links. The power grid mainly includes the volatility regulation of a high proportion of new energy, multi-energy complementary control, electric vehicle access control, user-side management, renewable energy power generation control, and energy terminal control. This level mainly improves the resilience of UGCPS by reducing the uncertainty on both sides of the power grid. The improvement factors of the coupling part include the monitoring and control of multi-network chain failures, the resistance of power electronic equipment, the grid structure of the power grid, and the interactive stability of the power grid and the information network. This level mainly improves the resilience of UGCPS at the coupling level of the power grid and the information network.

[0176] 2. Construction of pairwise comparison matrix in step S1

[0177] In order to reduce the subjectivity of the DEMATEL analysis process, this embodiment invites multiple experts to score the indicators. For the constructed indicator system, the experts evaluate their mutual influence. Therefore, a group DEMATEL evaluation process is formed. In this process, it is assumed that there are m decision experts, marked as At the same time, the decision process involves n promotion factors, which are marked as During the scoring process, the kth decision expert Give a pairwise comparison matrix for all improvement factors As shown in formula (1), represents the kth decision expert Factors for improvement and Ratings for pairwise comparisons.

[0178] (1)

[0179] 3. Knowledge correction matrix construction in step S2

[0180] In order to increase the objectivity of the DEMATEL analysis process, a knowledge graph of UGCPS resilience improvement factors was established based on relevant paper data. After embedding the knowledge graph with the help of TransH, this embodiment corrected the pairwise comparison matrix and implemented knowledge correction in the following three steps.

[0181] (1) Triplet establishment

[0182] A positive sample triple (h, t, r) is constructed to represent the structure of the knowledge graph. Represents an entity, Represents a relationship. The triples of entities and relationships are composed as follows Figure 2The entity types involved in this embodiment are: papers, fields, and promotion factors.

[0183] (2) Graph Embedding

[0184] The TransH model can increase the accuracy of "one-to-many", "many-to-one" and "many-to-many" entity relationships, as shown in Figure 3. Corresponding to a hyperplane ; For h,t in the hyperplane The projection satisfies ; Negative samples and distance as large as possible. As calculated in formula (2), and The distance is calculated as shown in formula (4). The objective function of TransH is shown in formula (3) and formula (4). The objective function should satisfy the vector scale constraint (5) and the orthogonality constraint (6).

[0185] (2)

[0186] (3)

[0187] (4)

[0188] (5)

[0189] (6)

[0190] 4. Matrix fusion in step S3

[0191] Through the TransH algorithm, the distances between the lifting factors are calculated to form a distance matrix Formula (7) shows The calculation process.

[0192] (7)

[0193] The matrix obtained in step S2 This can be done as in step S1 Let m=m+1, so that these two parts can be unified into It is necessary to use complex network theory to merge them into a unified pairwise comparison matrix X. This process is relatively complicated, and the process is as follows Figure 4 shown.

[0194] 4.1. Expert Network Construction

[0195] According to the consistency among decision-making experts, an undirected weighted complex network G can be constructed. G has m nodes, which represent decision-making experts. The weights of the edges represent the consistency among experts. When the consistency of all experts is the same, the weighted complex network degenerates into an unweighted complex network.

[0196] 4.2 Expert Clustering

[0197] Drawing on the idea of ​​CONCOR clustering and combining the characteristics of group DEMATEL, a node clustering method suitable for complex networks in this paper is proposed. The specific process is as follows.

[0198] (1) Construct the original correlation coefficient matrix .

[0199] (2) CONCOR transformation: Calculate the new correlation coefficient matrix Repeat this step to get a new correlation coefficient matrix .

[0200] (3) From the CONCOR clustering algorithm, we can see that after multiple CONCOR transformations, we will get ,and .

[0201] (4) The row elements in are divided into 2 clusters. For 1 cluster, for another cluster.

[0202] (5) Divide the decision experts into two categories based on the initial clustering results, and repeat (1)-(5) to divide the two subcategories separately until the number of elements in each subcategory reaches satisfy After clustering, the experts are divided into c clusters, with the properties of formula (9).

[0203] (9)

[0204] 4.3 Expert consensus and weight calculation

[0205] (1) Network and cluster consistency calculation

[0206] In the group DEMATEL evaluation problem, group satisfaction can reflect the degree of consensus. The cohesion in a complex network can measure the satisfaction of the group. The higher the cohesion of the complex network, the closer the connection between the nodes, and the higher the group satisfaction. According to the derivation of the literature, formula (10) calculates the network cohesion. Formula (11) calculates the cluster cohesion.

[0207] (10)

[0208] (11)

[0209] (2) Node and cluster weight calculation

[0210] Node weight and cluster weight are very important for complex network problems. Node weight and cluster weight are calculated with the help of network and cluster consistency. Formula (12) calculates the importance of the node as the basis of the node weight.

[0211] (12)

[0212] In the formula, Represents a complex network G where nodes are deleted The network consistency after that. As shown in formula (13), the normalized result of the importance of the node is the node weight.

[0213] (13)

[0214] Similarly, we give the concept of subgroup importance. Formula (14) gives the calculation method of subgroup importance.

[0215] (14)

[0216] In the formula, Represents a complex network G where clusters are deleted The network consistency after all experts in .

[0217] Different from the node weight, when determining the subgroup weight, not only the subgroup importance but also the subgroup cohesion should be considered. Formula (15) shows the subgroup weight.

[0218] (15)

[0219] (3) Density fusion operator calculation

[0220] Network density can be expressed by the network density weighted arithmetic averaging (NDWAA). Formula (16) shows the expert group evaluation results synthesized by NDWAA.

[0221] (16)

[0222] 5. Calculation of the importance of the lifting factors in step S4

[0223] 5.1 Matrix Normalization

[0224] Formula (17) shows the standardized distance matrix .

[0225] (17)

[0226] 5.2 Relationship Matrix Generation

[0227] According to formula (18), the overall relationship matrix can be calculated . T represents the direct or implicit relationship between the promotion factors. Represents the influence between factor i and factor j.

[0228] (18)

[0229] Where I is the identity matrix and i represents the power of the matrix.

[0230] 5.3 Causal Relationship Calculation

[0231] According to the matrix T, the influence can be calculated and the degree of influence Formula (19) and formula (20) show how to calculate P and Q. Based on P and Q, the centrality can be calculated. and cause degree , as shown in formula (21) and formula (22).

[0232] Based on C and R, a "cause and effect diagram" can be drawn for each indicator. C is the horizontal axis, indicating the importance of each element. R is the vertical axis, indicating the degree of influence of each element on other elements.

[0233] (19)

[0234] (20)

[0235] (twenty one)

[0236] (twenty two)

[0237] 6. Example Analysis

[0238] 6.1 Basic Data

[0239] This embodiment selects 10 professors and experts in the fields of power operation and power management, and all of them come from research institutes, power grid companies and universities. Half of the experts have more than 10 years of work experience, while the others have more than 5 years of work experience. Limited by the hardware level of the system, this article selects 300 papers, and after processing, 299 valid papers (with DOI and abstracts) are retained. The search terms of the article are: "TS=(power grid) AND TS=(information-physical fusion system) AND TS=(risk) AND(TS=(power grid) OR TS=(information network) OR TS=(coupling))". In the text classification process, 239 articles are training sets and 60 articles are test sets. In TransH embedding, this embodiment constructs 2665 training triples and 260 test triples.

[0240] 6.2. Results of toughness improvement mechanism analysis

[0241] In order to verify the reliability of the factor analysis, this example uses Crombach α reliability analysis and obtains Based on this, it can be proved that the results of the improvement factor evaluation have high credibility. Risk analysis can be carried out based on this data.

[0242] In this embodiment, a questionnaire was designed and distributed to the above 10 experts. The content of the questionnaire is the comparison of the influence of each pair of improvement factors. The influence level is divided into 5 levels, represented by (0, 1, 2, 3, 4). For example, Table 2 shows the scores of the influence of four (R1, R2, R3, R4) improvement factors by two experts.

[0243] Table 2 Pairwise comparison matrix between experts 1 and 2

[0244]

[0245] After TransH, each improvement factor becomes a vector representation. Figure 5 As shown in the figure, different areas on the coordinate axis correspond to different impact levels, impact factors and related papers. The papers in the middle part belong to these areas at the same time. Figure 5 The results show the reliability of the embedding results.

[0246] According to formula (8) and formula (18), this embodiment synthesizes the pairwise comparison matrix of 9 experts and 1 knowledge correction matrix to form the results in Table 4.

[0247] Table 3 Knowledge correction matrix

[0248]

[0249] Table 4 Influence relationship matrix

[0250]

[0251] The influence and influence of each secondary influencing factor can be obtained by adding the results of Table 4. According to formula (19) and formula (22), the centrality and result degree of each secondary factor can be obtained, and the results are shown in Table 5. The last two columns of Table 5 are the centrality and cause degree, respectively. The centrality indicates the importance of the promotion factor, and the cause degree indicates the strength of the influence of the promotion factor on other factors. From the centrality of Table 5, it can be seen that the construction of the power grid is of great importance to the resilience improvement of UGCPS. Among them, the highest are user-side management, multi-energy complementary control and control of renewable energy generation. The above three factors will be connected to many intelligent devices to realize the real-time response of urban user load to price, or the interaction between energy supply and energy demand of wind, light, water and fire. These intelligent devices need to be connected to the power grid through the information network, but the connection process often has no unified network standard, and the equipment often pays more attention to business functions rather than information security, which makes the equipment easy to be invaded and causes the risk of UGCPS. From the cause degree, we can see that the three factors that have the greatest impact on other factors belong to different levels, namely, the accuracy of information network operation and maintenance, the volatility regulation of high proportion of new energy, and the ability of power electronic equipment to resist extreme events. The accuracy of information network operation and maintenance is the basis for the stable operation of the information network. If there is a problem in the operation and maintenance process, other factors will be out of control and a chain failure will occur, reducing the resilience of UGCPS. The volatility regulation of high proportion of new energy is the primary task of the power grid. If there is volatility in power supply, measures such as multi-energy complementary control, electric vehicle regulation and user-side management will be restricted, which will have a great impact on the failure or reliability of the power grid, thereby reducing the resilience of UGCPS. The ability of power electronic equipment to resist extreme events reflects the attention paid to the key nodes of the coupling department. In the face of extreme events, once the power electronic equipment fails, it will not only cause a disconnection between power supply and demand, but also the information collection and equipment regulation of UGCPS will cause corresponding risks, reducing the resilience of UGCPS.

[0252] Table 5 The influence of factors on improving resilience

[0253]

[0254] After evaluation by 9 experts and information from 299 papers, the order of resilience-enhancing factors was determined based on centrality: R8> R6> R10> R7> R5> R11> R14> R1> R4> R9> R2> R13> R3> R12 (the significance of the factors is as follows Figure 2According to the causal degree, the cause factors are determined to be: R12> R1> R5> R4> R9; the result factors are: R6> R8> R13> R2> R11> R7> R10> R14> R3.

[0255] Finally, it is judged that the accuracy of information network operation and maintenance (R1), the fluctuation regulation of high proportion of new energy (R5) and the ability of power electronic equipment to resist extreme events (R12) are the three most important factors for improving the resilience of UGCPS.

[0256] Example 2

[0257] This embodiment also provides a resilience improvement system for a power grid cyber-physical fusion system, including:

[0258] An influencing factor setting module is used to set multiple influencing factors for improving resilience of the power grid cyber-physical fusion system;

[0259] The expert method module is used to score the mutual influence relationship between two influencing factors using the expert method to obtain a pairwise comparison matrix;

[0260] The correction matrix calculation module is used to obtain papers related to improving the resilience of the power grid information-physical fusion system, build a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between the influencing factors based on the knowledge graph to form a correction matrix;

[0261] A fusion module is used to fuse the pairwise comparison matrix and the correction matrix to obtain the final pairwise comparison matrix;

[0262] The importance calculation module is used to calculate the centrality and causality of the final pairwise comparison matrix, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and to determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system.

[0263] It should be noted that the specific content and beneficial effects of the system of the present application can be found in the above-mentioned method embodiment, which will not be repeated here.

[0264] Example 3

[0265] This embodiment provides a device for improving the resilience of a power grid information-physical fusion system, including a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of a method for improving the resilience of a power grid information-physical fusion system as described in Embodiment 1.

[0266] Example 4

[0267] This embodiment provides a computer-readable storage medium, on which a computer program is stored. The computer program is executed by a processor to perform the steps of a method for improving the resilience of a power grid information-physical fusion system as described in Embodiment 1.

[0268] The preferred specific embodiments of the present invention are described in detail above. It should be understood that a person skilled in the art can make many modifications and changes based on the concept of the present invention without creative work. Therefore, any technical solution that can be obtained by a person skilled in the art through logical analysis, reasoning or limited experiments based on the concept of the present invention on the basis of the prior art should be within the scope of protection determined by the claims.

Claims

1. A method for improving the resilience of a power grid information-physical fusion system, characterized in that: The following steps are involved: Set up multiple factors that influence resilience improvement for the power grid cyber-physical fusion system; The expert method is used to score the mutual influence relationship between two influencing factors to obtain a pairwise comparison matrix; Obtain papers related to improving the resilience of power grid cyber-physical fusion systems, construct a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between influencing factors based on the knowledge graph to form a correction matrix; The pairwise comparison matrix and the correction matrix are combined to obtain a final pairwise comparison matrix; The centrality and causality of the final pairwise comparison matrix are calculated, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system; The knowledge graph is in the form of a triple (h, t, r), where: Represents an entity, To express a relationship; The objective function of the TransH algorithm in calculating the distance between two influencing factors based on the knowledge graph includes: ； ； ； In the formula, is the objective function of the TransH algorithm, and is the score function of the TransH model, is a constant used to represent the spacing between positive and negative samples. is h in the hyperplane The projection of for The transposed matrix of For t in the hyperplane The projection of is the set of positive triples, is the set of negative triples; The expression of the constraint condition satisfied by the objective function of the TransH algorithm is: ； In the formula, is the set of real numbers; The fusion process of the pairwise comparison matrix and the correction matrix comprises the following steps: The correction matrix is ​​combined with the pairwise comparison matrix to construct an undirected weighted complex network G, wherein the complex network G has m nodes, each node representing a decision expert; the complex network G has The weight of each edge represents the consistency among experts; Expert clustering is performed on the complex network G, and the decision experts in the complex network G are divided into c Clusters of classes; Calculating the network cohesion according to the complex network G after clustering, and calculating the cluster consistency according to the correction matrix; Calculating expert importance based on network cohesion, and calculating cluster importance based on the network cohesion and cluster consistency; Expert weights are calculated based on expert importance, and cluster weights are calculated based on cluster importance and cluster consistency; According to the expert weights and cluster weights, the NDWAA operator value of the complex network G is calculated to obtain the final pairwise comparison matrix.

2. The method for improving the resilience of a power grid information-physical fusion system according to claim 1, characterized in that: The power grid information-physical fusion system is divided into a multi-layer logical architecture, including an information network, a power grid and a coupling part; The influencing factors corresponding to the information network include the accuracy of information network operation and maintenance, the reliability of information transmission channels, the reliability of information collection systems and the defense capability against network attacks; The influencing factors corresponding to the power grid include volatility regulation of high-proportion new energy, multi-energy complementary control, electric vehicle access regulation, user-side management, control of renewable energy power generation and control of energy-consuming terminals; The coupling part includes the ability of power electronic equipment to resist extreme events, the grid structure of the power grid, the stability of the interaction between the power grid and the information network, and the monitoring and control of chain failures of the energy system, transportation network and power grid.

3. The method for improving the resilience of a power grid information-physical fusion system according to claim 1, characterized in that: The expression of the pairwise comparison matrix obtained by expert scoring is: ； In the formula, is the pairwise comparison matrix given by the k-th decision expert for all influencing factors, is the influencing factor of the k-th decision expert and The score of pairwise comparison is n, where n is the total number of influencing factors.

4. The method for improving the resilience of a power grid information-physical fusion system according to claim 1, characterized in that: The process of expert clustering of the complex network G includes: S101: constructing an original correlation coefficient matrix according to the complex network G ; S102: Calculate a new correlation coefficient matrix based on CONCOR transformation , and repeat this step to get a new correlation coefficient matrix ; Until the correlation coefficient matrix is ​​obtained ,and ; S103: Correlation coefficient matrix The row elements in are divided into two clusters, where For 1 cluster, for another cluster; S104: According to the cluster division result, the decision experts are divided into two categories, and steps S101 to S104 are repeated until the number of elements in each subclass is satisfy So far, the decision experts are finally divided into c clusters.

5. The method for improving the resilience of a power grid information-physical fusion system according to claim 1, characterized in that: The calculation expression of the network cohesion is: ； In the formula, is the network cohesion, is the total number of influencing factors, is the scoring result between the i-th influencing factor and the j-th influencing factor; The calculation expression of the cluster consistency is: ； In the formula, is the correction matrix, For cluster consistency, is the total number of elements in the correction matrix, is the correction result between the i-th element and the j-th element in the correction matrix; The calculation expression of the expert importance is: ； In the formula, For Node The importance of experts Delete nodes from complex network G The network cohesion afterwards; The calculation expression of the expert weight is: ； In the formula, For Node Expert weights; The calculation expression of the cluster importance is: ； In the formula, For cluster The importance of clusters, Delete clusters for complex network G Network consistency after all experts in; The calculation expression of the cluster weight is: ； In the formula, For cluster The cluster weight of .

6. A method for improving resilience of a power grid information-physical fusion system according to claim 5, characterized in that: The calculation expression of the NDWAA operator value is: ； In the formula, is the NDWAA operator value, For NDWAA synthesis calculation, is the value after standardization.

7. The method for improving resilience of a power grid information-physical fusion system according to claim 1, characterized in that: The calculation expression of the centrality is: ； ； ； ； ； ； ； ； In the formula, is the standardized distance matrix, is the distance value between influencing factor i and influencing factor k; is the total number of influencing factors, is the overall relationship matrix, is the identity matrix, the superscript i is the power of the matrix, is the influence degree of influencing factor i, is the degree of influence of influencing factor i, is the relationship between influencing factor i and influencing factor j, is the centrality of influencing factor i, is the centrality; The calculation expression of the cause degree is: ； ； In the formula, For the reason, is the causal degree of influencing factor i.

8. A resilience enhancement system for a power grid information-physical fusion system, characterized in that: include: An influencing factor setting module is used to set multiple influencing factors for improving resilience of the power grid cyber-physical fusion system; The expert method module is used to score the mutual influence relationship between two influencing factors using the expert method to obtain a pairwise comparison matrix; A correction matrix calculation module is used to obtain papers related to improving the resilience of power grid cyber-physical fusion systems, construct a knowledge graph of resilience improvement factors, and use the TransH algorithm to calculate the distance between influencing factors based on the knowledge graph to form a correction matrix. A fusion module, used for fusing the pairwise comparison matrix and the correction matrix to obtain a final pairwise comparison matrix; The importance calculation module is used to calculate the centrality and causality of the final pairwise comparison matrix, so as to determine the resilience improvement ranking results of each influencing factor according to the centrality, and determine the causal factors and result factors in each influencing factor and their ranking according to the causality, so as to select the corresponding influencing factors to improve the resilience of the power grid information-physical fusion system; The knowledge graph is in the form of a triple (h, t, r), where: Represents an entity, To express a relationship; The objective function of the TransH algorithm in calculating the distance between two influencing factors based on the knowledge graph includes: ； ； ； In the formula, is the objective function of the TransH algorithm, and is the score function of the TransH model, is a constant used to represent the spacing between positive and negative samples. is h in the hyperplane The projection of for The transposed matrix of For t in the hyperplane The projection of is the set of positive triples, is the set of negative triples; The expression of the constraint condition satisfied by the objective function of the TransH algorithm is: ； In the formula, is the set of real numbers; In the fusion module, the fusion process of the pairwise comparison matrix and the correction matrix includes the following steps: The correction matrix is ​​combined with the pairwise comparison matrix to construct an undirected weighted complex network G, wherein the complex network G has m nodes, each node representing a decision expert; the complex network G has The weight of each edge represents the consistency among experts; Expert clustering is performed on the complex network G, and the decision experts in the complex network G are divided into c Clusters of classes; Calculating the network cohesion according to the complex network G after clustering, and calculating the cluster consistency according to the correction matrix; Calculating expert importance based on network cohesion, and calculating cluster importance based on the network cohesion and cluster consistency; Expert weights are calculated based on expert importance, and cluster weights are calculated based on cluster importance and cluster consistency; According to the expert weights and cluster weights, the NDWAA operator value of the complex network G is calculated to obtain the final pairwise comparison matrix.

9. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 8, characterized in that: The power grid information-physical fusion system is divided into a multi-layer logical architecture, including an information network, a power grid and a coupling part; The influencing factors corresponding to the information network include the accuracy of information network operation and maintenance, the reliability of information transmission channels, the reliability of information collection systems and the defense capability against network attacks; The influencing factors corresponding to the power grid include volatility regulation of high-proportion new energy, multi-energy complementary control, electric vehicle access regulation, user-side management, control of renewable energy power generation and control of energy-consuming terminals; The coupling part includes the ability of power electronic equipment to resist extreme events, the grid structure of the power grid, the stability of the interaction between the power grid and the information network, and the monitoring and control of chain failures of the energy system, transportation network and power grid.

10. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 8, characterized in that: The expression of the pairwise comparison matrix obtained by expert scoring is: ； In the formula, is the pairwise comparison matrix given by the k-th decision expert for all influencing factors, is the influencing factor of the k-th decision expert and The score of pairwise comparison is n, where n is the total number of influencing factors.

11. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 8, characterized in that: The process of expert clustering of the complex network G includes: S101: constructing an original correlation coefficient matrix according to the complex network G ; S102: Calculate a new correlation coefficient matrix based on CONCOR transformation , and repeat this step to get a new correlation coefficient matrix ; Until the correlation coefficient matrix is ​​obtained ,and ; S103: Correlation coefficient matrix The row elements in are divided into two clusters, where For 1 cluster, for another cluster; S104: According to the cluster division result, the decision experts are divided into two categories, and steps S101 to S104 are repeated until the number of elements in each subclass is satisfy So far, the decision experts are finally divided into c clusters.

12. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 8, characterized in that: The calculation expression of the network cohesion is: ； In the formula, is the network cohesion, is the total number of influencing factors, is the scoring result between the i-th influencing factor and the j-th influencing factor; The calculation expression of the cluster consistency is: ； In the formula, is the correction matrix, For cluster consistency, is the total number of elements in the correction matrix, is the correction result between the i-th element and the j-th element in the correction matrix; The calculation expression of the expert importance is: ； In the formula, For Node The importance of experts Delete nodes from complex network G The network cohesion afterwards; The calculation expression of the expert weight is: ； In the formula, For Node Expert weights; The calculation expression of the cluster importance is: ； In the formula, For cluster The importance of clusters, Delete clusters for complex network G Network consistency after all experts in; The calculation expression of the cluster weight is: ； In the formula, For cluster The cluster weight of .

13. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 12, characterized in that: The calculation expression of the NDWAA operator value is: ； In the formula, is the NDWAA operator value, For NDWAA synthesis calculation, is the value after standardization.

14. The resilience enhancement system of a power grid cyber-physical fusion system according to claim 8, characterized in that: The calculation expression of the centrality is: ； ； ； ； ； ； ； ； In the formula, is the standardized distance matrix, is the distance value between influencing factor i and influencing factor k; is the total number of influencing factors, is the overall relationship matrix, is the identity matrix, the superscript i is the power of the matrix, is the influence degree of influencing factor i, is the degree of influence of influencing factor i, is the relationship between influencing factor i and influencing factor j, is the centrality of influencing factor i, is the centrality; The calculation expression of the cause degree is: ； ； In the formula, For the reason, is the causal degree of influencing factor i.

15. A device for improving the resilience of a power grid information-physical fusion system, characterized in that: The method comprises a memory and a processor, wherein the memory stores a computer program, and the processor calls the computer program to execute the steps of any one of the methods described in claims 1 to 7.

16. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a computer program, and the computer program is used by a processor to execute the method according to any one of claims 1 to 7.

Images