Weak point identification method and medium for power optical transmission network based on node influence

Abstract

The present invention relates to a method for identifying weak points in an electric power optical transmission network based on node influence, comprising the following steps: obtaining the electric power optical transmission network to be identified, abstracting the electric power optical transmission network to be identified as an undirected and unweighted network; determining an index A considered by site influence, and calculating the weight T1 of index A; constructing a comprehensive site influence index based on index A and weight T1; determining an index B considered by equipment influence, and calculating the weight T2 of index B; constructing a comprehensive equipment influence index based on index B and weight T2; calculating the values of the comprehensive site influence index and the comprehensive equipment influence index based on the actual parameters of the undirected and unweighted network and the electric power optical transmission network; and identifying weak points in the electric power optical transmission network based on the values. The present invention relies on the comprehensive site influence index and the comprehensive equipment influence index to identify weak points in the electric power optical transmission network, making the evaluation results more objective and comprehensive.

Description

Technical Field

[0001] The present invention relates to a method and medium for identifying weak points in an electric power optical transmission network based on node influence, and belongs to the technical field of electric power optical transmission network maintenance. Background Art

[0002] Node influence represents the degree of impact a node has on network resilience. The failure of a few high-impact nodes can severely disrupt or even paralyze the entire network. To ensure that the optical transmission network maintains its connectivity and backbone communication capabilities during large-scale natural disasters such as typhoons, earthquakes, and tsunamis, the dispatch center must maintain sufficient dispatch capabilities even in harsh environments. However, the topology of power communication networks is complex, and the influence of nodes is affected by a variety of factors. Relying solely on expert experience has limited accuracy. In reality, deterioration in equipment status is often due to changes in the network structure, which increases their influence and dramatically increases operational pressure. Therefore, the influence of these nodes is often relatively large and difficult to detect. Summary of the Invention

[0003] In order to overcome the above problems, the present invention provides a method and medium for identifying weak points in an electric power optical transmission network based on node influence. The method relies on comprehensive site influence indicators and comprehensive equipment influence indicators to identify weak points in the electric power optical transmission network, making the evaluation results more objective and comprehensive.

[0004] The technical solutions of the present invention are as follows:

[0005] First aspect

[0006] A method for identifying weak points in a power optical transmission network based on node influence comprises the following steps:

[0007] Acquire a power optical transmission network to be identified, and abstract the power optical transmission network to be identified into an undirected and unweighted complex network;

[0008] Determine the indicator A that is considered for site influence and calculate the weight of indicator A;

[0009] Construct a comprehensive index of site influence based on indicator A and its weight;

[0010] Determine the indicator B that is considered for the equipment impact and calculate the weight of indicator B;

[0011] Construct a comprehensive index of equipment influence based on index B and its weight;

[0012] Calculating values of the site influence comprehensive index and the equipment influence comprehensive index based on actual parameters of the undirected and unweighted complex network and the power optical transmission network;

[0013] The weak points of the power optical transmission network are identified according to the values of the site influence comprehensive index and the equipment influence comprehensive index.

[0014] Furthermore, the indicator A includes the node degree, clustering coefficient and voltage level of each node.

[0015] Furthermore, the comprehensive index q of site influence of node i i for:

[0016]

[0017] Among them, w z 、w g and w l are the node degree, clustering coefficient and voltage level weights respectively; N is the total number of nodes in the power optical transmission network; i and j are nodes in the power optical transmission network;

[0018] z i 、z j The node degrees k of nodes i and j are i 、k j Normalized results;

[0019] The node degree k of node n n After normalization, we get z n , the formula is as follows:

[0020]

[0021] k n =∑ m∈G δ nm ;

[0022]

[0023] Where G is the set of all nodes in the power optical transmission network, n and m are nodes in the power optical transmission network;

[0024] g i 、g j are the average clustering coefficients of nodes i and j respectively;

[0025] The average clustering coefficient g of node n n as follows:

[0026]

[0027] Clustering coefficient c of node n n as follows;

[0028]

[0029] Among them, en is the number of edges between node n’s neighbor nodes;

[0030] l i ′、l j ′ are the average voltage coefficients of nodes i and j respectively;

[0031] The average voltage coefficient l at node n n 'as follows:

[0032]

[0033] Voltage coefficient l at node n n Determined by the voltage level of node n.

[0034] Furthermore, the voltage coefficients corresponding to node voltage levels of 500kV, 220kV, 110kV, and below 35kV are 4, 3, 2, and 1, respectively.

[0035] Furthermore, the indicator B includes the aging degree, number of failures and bandwidth pressure of each node.

[0036] Furthermore, the comprehensive index of device influence p of device i i for:

[0037]

[0038] Among them, w s 、w f and w n are the weights of bandwidth pressure, number of failures and aging degree respectively; N is the total number of devices in the node; i and j are the devices in the node;

[0039] s i ′、s j ′ is the bandwidth pressure of devices i and j respectively;

[0040] Bandwidth pressure s of device n n 'as follows:

[0041]

[0042] Among them, l n 、l m Bandwidth usage of the device;

[0043] f i ′、f j ′ are the number of failures f of equipment i and j respectively i 、f j Normalized results;

[0044] The normalized number of failures of device n is f n ′, the formula is as follows

[0045]

[0046] n i ′、n j ′ are the aging degrees of equipment i and j respectively;

[0047] Aging degree of equipment n n 'as follows:

[0048]

[0049] Among them, t n is the number of years of operation of equipment n, T n is the aging age of device n, r n is the redundancy of device n.

[0050] Furthermore, the bandwidth usage of the device is the ratio of the used bandwidth of the device to the maximum bandwidth.

[0051] Furthermore, the redundancy r n The value of is the number of homogeneous devices of device n.

[0052] Furthermore, the weights between different indicators are determined as follows:

[0053] The weights between different indicators are determined by the multi-attribute decision-making method. The algorithm is as follows:

[0054]

[0055] Among them, r ij is the i-th attribute value of the j-th indicator, m is the number of attribute values, n is the number of indicators, w j is the weight of the j-th indicator.

[0056] Second aspect

[0057] A storage medium stores a computer program, which outputs the site influence comprehensive index and the device influence comprehensive index described in the first aspect when the program is executed.

[0058] The present invention has the following beneficial effects:

[0059] 1. This identification method identifies power optical transmission network points through comprehensive site influence indicators and comprehensive equipment influence indicators, making the evaluation results more objective and comprehensive.

[0060] 2. The comprehensive site influence index of this identification method takes into account the voltage level, clustering coefficient and node degree of the node, which solves the limitation of the existing technology that only evaluates the node influence based on the topological location attributes of the communication network node.

[0061] 3. The comprehensive device influence index of this identification method takes into account the device's aging, number of failures, and bandwidth pressure, which solves the limitation of existing technologies that only evaluate node influence based on device performance attributes. BRIEF DESCRIPTION OF THE DRAWINGS

[0062] FIG1 shows the change of the network efficiency reduction rate after a node is deleted from the network according to an embodiment of the present invention. DETAILED DESCRIPTION

[0063] The present invention will be described in detail below with reference to the accompanying drawings and specific embodiments.

[0064] First aspect

[0065] Example 1

[0066] A method for identifying weak points in a power optical transmission network based on node influence comprises the following steps:

[0067] Obtaining a power optical transmission network to be identified, and abstracting the power optical transmission network to be identified into an undirected and unweighted complex network; in one embodiment of the present invention, the network G = (V, E) is an undirected and unweighted network consisting of |V| = N nodes and |E| = M edges, through which the connection relationship between the nodes can be quickly obtained;

[0068] Determine the indicator A that is considered for site influence and calculate the weight of indicator A;

[0069] Construct a comprehensive index of site influence based on indicator A and its weight;

[0070] Determine the indicator B that is considered for the equipment impact and calculate the weight of indicator B;

[0071] Construct a comprehensive index of equipment influence based on index B and its weight;

[0072] Calculating values of the site influence comprehensive index and the equipment influence comprehensive index based on actual parameters of the undirected and unweighted complex network and the power optical transmission network;

[0073] The weak points of the power optical transmission network are identified according to the values of the site influence comprehensive index and the equipment influence comprehensive index.

[0074] Example 2

[0075] A method for identifying weak points in an electric power optical transmission network based on node influence. Based on the first embodiment, the indicator A includes the node degree, clustering coefficient and voltage level of each node.

[0076] The calculation of node influence is the premise of network topology diagnosis. Each node in the network has different effects on the network influence according to its influence. Prioritizing the maintenance of nodes with great influence in the network is an effective way to improve the network influence. In the process of calculating node influence, the node degree is first considered as an important indicator for judging node influence. It reflects the scale of the node's neighbors and emphasizes the influence of the node's own attributes. Since the power communication network is a sparse network and has scale-free properties, most of its node degrees are 1 or 2. Only a small number of nodes have a larger node degree and are mostly star-shaped. Therefore, the invention uses the node degree indicator as a condition for judging node influence.

[0077] While the clustering coefficient doesn't reflect the size of neighboring nodes, it describes the presence of triangular structures around a node and, to a certain extent, reflects the closeness of node neighbors. In power communication networks, whether a node can find a path to continue service transmission after a node fails is also an important criterion for determining a node's influence.

[0078] In a hierarchical power grid, high voltage nodes are very important and should have a higher protection priority.

[0079] In a specific embodiment of the present invention, the comprehensive site influence index q of node i is i for:

[0080]

[0081] Among them, w z 、w g and w l are the node degree, clustering coefficient and voltage level weights respectively; N is the total number of nodes in the power optical transmission network; i and j are nodes in the power optical transmission network;

[0082] z i 、z j The node degrees k of nodes i and j are i 、k j Normalized results;

[0083] The node degree k of node n n After normalization, we get z n , the formula is as follows:

[0084]

[0085] k n =∑ m∈G δ nm ;

[0086]

[0087] Where G is the set of all nodes in the power optical transmission network, n and m are nodes in the power optical transmission network;

[0088] g i 、g j are the average clustering coefficients of nodes i and j respectively;

[0089] The average clustering coefficient g of node n n as follows:

[0090]

[0091] Clustering coefficient c of node n n as follows;

[0092]

[0093] Among them, e n is the number of edges between node n’s neighbor nodes;

[0094] l i ′、l j ′ are the average voltage coefficients of nodes i and j respectively;

[0095] The average voltage coefficient l at node n n 'as follows:

[0096]

[0097] Voltage coefficient l at node n n Determined by the voltage level of node n.

[0098] In a specific embodiment of the present invention, the voltage coefficients corresponding to the node voltage levels of 500kV, 220kV, 110kV, and 35kV or below are 4, 3, 2, and 1, respectively.

[0099] Example 3

[0100] A method for identifying weak points in a power optical transmission network based on node influence, based on Example 1, wherein the indicator B includes the aging degree, number of failures, and bandwidth pressure of each node. This method overcomes the limitation of existing technologies that evaluate node influence based solely on device performance attributes.

[0101] In a specific embodiment of the present invention, the comprehensive index of device influence p of device i is i for:

[0102]

[0103] Among them, w s 、w f and w nare the weights of bandwidth pressure, number of failures and aging degree respectively; N is the total number of devices in the node; i and j are the devices in the node;

[0104] s i ′、s j ′ is the bandwidth pressure of devices i and j respectively;

[0105] Bandwidth pressure s of device n n 'as follows:

[0106]

[0107] Among them, k n 、k m Bandwidth usage of the device;

[0108] f i ′、f j ′ are the number of failures f of equipment i and j respectively i 、f j Normalized results;

[0109] The normalized number of failures of device n is f n ′, the formula is as follows

[0110]

[0111] n i ′、n j ′ are the aging degrees of equipment i and j respectively;

[0112] Aging degree of equipment n n 'as follows:

[0113]

[0114] Among them, t n is the number of years of operation of equipment n, T n is the aging age of device n, r n is the redundancy of device n.

[0115] When the equipment's operating years exceed the aging age limit, it is determined to be in an aging state. The ratio of the two can intuitively quantify the degree of equipment aging.

[0116] In a specific implementation of the present invention, the bandwidth usage of the device is the ratio of the used bandwidth of the device to the maximum bandwidth.

[0117] In a specific embodiment of the present invention, the redundancy r n The value of is the number of homogeneous devices in device n. For example, if a device uses the "1+1" backup mode, the redundancy of the device is 2. The greater the redundancy, the lower the vulnerability of the device.

[0118] Example 4

[0119] A method for identifying weak points in a power optical transmission network based on node influence, based on embodiments 1 to 3, uses a multi-attribute decision-making method to determine the weights between different indicators. The algorithm is as follows:

[0120]

[0121] Among them, r ij is the i-th attribute value of the j-th indicator, m is the number of attribute values, n is the number of indicators, w j is the weight of the j-th indicator.

[0122] Since different indicators contribute differently to the influence of nodes, and according to the multi-attribute decision-making theory, the higher the similarity of attribute values, the smaller the influence of the attribute on the decision result, the algorithm in this paper determines the contribution of different indicator attribute values to the influence of nodes based on their similarity.

[0123] The following is a specific embodiment of the present invention.

[0124] In this embodiment, network topology diagnosis simulation is performed in an ideal network and a real network respectively. The basic characteristics of the selected simulated network topology are shown in Table 1.

[0125] Table 1 Characteristic parameters of different simulation networks

[0126]

[0127]

[0128] Where N represents the number of network nodes, M represents the number of edges, <k>represents the average degree of the node, <d>represents the average shortest distance of the network, <c>It represents the average clustering coefficient of the nodes, where the basic characteristics of the ideal network are taken as the average value of all networks of the same type.

[0129] This embodiment uses scale-free networks and small-world networks, which have similar characteristics to the power communication network, as representatives to perform diagnostic analysis on the ideal network.

[0130] In conducting network invulnerability diagnostic analysis, this paper simulates nine scale-free networks with different parameters, each with the number of new edges m ranging from 2 to 10, and seven small-world networks with different parameters, each with an average degree set from 4 to 10. The total number of nodes in all generated networks is 1000. The average shortest distances of the nine different scale-free networks range from [2.5628, 4.3123], with an average of 3.0725, and the average clustering coefficients range from [0.0127, 0.0603], with an average of 0.0385. The average shortest distances of the seven different small-world networks range from [3.6046, 6.2177], with an average of 4.3790, and the average clustering coefficients range from [0.1715, 0.2329], with an average of 0.2170.

[0131] When conducting network robustness diagnostic analysis, simulations were performed on a scale-free network with a new edge count of 2 and a small-world network with an average degree parameter of 4. The total number of nodes in the resulting network was set to 100. This is because the number of edges between nodes in an ideal network is proportional to the number of nodes and is much larger than in a real power communication network. Adding more nodes would reduce the experimental effect. The average shortest distance of the resulting scale-free network was 3.1082 and the average clustering coefficient was 0.0982. The average shortest distance of the resulting scale-free network was 3.8006 and the average clustering coefficient was 0.2030.

[0132] 1) Invulnerability diagnosis and analysis

[0133] When conducting invulnerability diagnosis and analysis on an ideal network, the comprehensive index of site influence (q i ), comprehensive index of equipment influence (p i ), node degree index (k i ) and secondary degree index (f i ) were compared. In the same network, the top 10% of nodes ranked by influence were deleted based on rankings based on different metrics. The network efficiency drop after removing nodes was observed for scale-free and small-world networks of different states, thereby analyzing the changes in network invulnerability. Figure 1 shows the changes in network efficiency drop after removing nodes for networks with different parameters.

[0134] As shown in Figure 1(a), for scale-free networks with different m values, after deleting the top 10% of nodes, as the m value increases, the network efficiency decline rate corresponding to all indicators decreases, indicating that as the m value increases, the network's invulnerability becomes stronger and the impact of deleting some nodes on network performance becomes smaller and smaller. Moreover, although all indicators show a downward trend, the comprehensive site influence indicator q i After sorting and deleting nodes, the network efficiency decrease rate is significantly higher than other indicators, indicating that q i It can more accurately judge the influence of nodes on network invulnerability than other indicators. As can be observed from Figure 1(b), each indicator is still on a downward trend, which means that with the increase of average degree, network invulnerability becomes stronger and stronger, but the comprehensive index of device influence p i In most cases, it is better than other indicators, indicating that q i The indicators can still ensure good effectiveness.

[0135] 2) Robustness diagnostic analysis

[0136] When conducting robustness diagnostic analysis on an ideal network, the simulation also uses the comprehensive index of site influence (q i ), comprehensive index of equipment influence (p i ), degree index (k i ) and secondary degree index (f i ) Under the same conditions, the top-ranked nodes, as calculated by different metrics, were successively deleted from the same network. The changes in network connectivity after node deletion were observed, thereby analyzing the impact of nodes calculated by different metrics on network robustness. For the resulting scale-free network, after deleting the first five nodes, divergences emerged between the different metrics. The top-ranked nodes, as ranked by different metrics, and the network robustness after successive node deletions are shown in Tables 2 through 6.

[0137] Table 2 According to k i Sorting of top-ranked nodes

[0138] Node number <![CDATA[k i ]]> <![CDATA[f i ]]> <![CDATA[e i ]]> <![CDATA[p i ]]> <![CDATA[q i ]]> 2 22 144 12 0.4998 0.5180 3 16 116 7 0.4225 0.3649 7 11 78 2 0.3178 0.2373 11 11 84 5 0.3337 0.2373 13 10 74 3 0.3064 0.2118 4 9 79 3 0.3200 0.1862

[0139] Table 3 According to f i Sorting of top-ranked nodes

[0140] Node number <![CDATA[k i ]]> <![CDATA[f i ]]> <![CDATA[e i ]]> <![CDATA[p i ]]> <![CDATA[q i ]]> 2 22 144 12 0.4998 0.5180 3 16 116 7 0.4225 0.3649 11 11 84 5 0.3337 0.2373 4 9 79 3 0.3200 0.1862 7 11 78 2 0.3178 0.2373 13 10 74 3 0.3064 0.2118

[0141] Table 4 According to p i Sorting of top-ranked nodes

[0142]

[0143]

[0144] Table 5 According to q i Sorting of top-ranked nodes

[0145] Node number <![CDATA[k i ]]> <![CDATA[f i ]]> <![CDATA[e i ]]> <![CDATA[p i ]]> <![CDATA[q i ]]> 2 22 144 12 0.4998 0.5180 3 16 116 7 0.4225 0.3649 7 11 78 2 0.3178 0.2373 11 11 84 5 0.3337 0.2373 13 10 74 3 0.3064 0.2118 4 9 79 3 0.3200 0.1862

[0146] Table 6 Network robustness after continuous node deletion

[0147] Number of deleted nodes 1 2 3 4 5 6 <![CDATA[η k ]]> 0.980 0.960 0.941 0.921 0.883 0.864 <![CDATA[η f ]]> 0.980 0.960 0.941 0.921 0.902 0.864 <![CDATA[η p ]]> 0.980 0.960 0.941 0.921 0.902 0.864 <![CDATA[η q ]]> 0.980 0.960 0.941 0.921 0.883 0.828

[0148] From Tables 2 to 6, we can see that although the first four nodes selected by different indicators are different, their impact on network robustness is the same. This is because the number of edges in an ideal network is much larger than that in a real network. When nodes are first deleted, almost all nodes in the network are connected by more than two paths. Deleting nodes has almost no effect on the connectivity between the remaining nodes. After deleting 5 nodes, the secondary degree indicator f i And equipment influence comprehensive index p i The 4th node is considered to be very important and ranked third. The comprehensive site influence index considers the 13th node to be more important because the node degree itself is more important than the neighbor node degree in a scale-free network, and the experimental results also prove this point. When the 6th node is deleted, the node degree index k i Node 4 is considered important, but the indicator in this paper selects node 40. This is because although the degrees of nodes 4 and 40 are the same, the neighbor connections of node 40 are significantly less than those of node 4, so its impact on network robustness is greater. For the generated small-world network, since its network connectivity is greater than that of the scale-free network, the network connectivity does not show a significant difference until the first 16 nodes are deleted. According to the node degree indicator k i , Second level index f i , equipment influence comprehensive index q i After deleting the nodes, the network connectivity was 0.6875, 0.6875 and 0.7042 respectively, while the comprehensive index of site influence p i After deleting 16 nodes, the network connectivity is 0.6711, which shows that the nodes deleted according to the comprehensive site influence index have a greater impact on the network robustness. Therefore, the comprehensive site influence index is more representative of the node influence.

[0149] Second aspect

[0150] A storage medium stores a computer program, which outputs the site influence comprehensive index and the device influence comprehensive index described in the first aspect when the program is executed.

[0151] The above descriptions are merely embodiments of the present invention and are not intended to limit the patent scope of the present invention. Any equivalent structures made using the contents of the present invention's description and drawings, or directly or indirectly applied in other related technical fields, are also included in the patent protection scope of the present invention.< / c> < / d> < / k>

Claims

1. A method for identifying weak points in a power optical transmission network based on node influence, characterized in that: The following steps are involved: Acquire a power optical transmission network to be identified, and abstract the power optical transmission network to be identified into an undirected and unweighted complex network; Determine the indicator A that is considered for site influence and calculate the weight of indicator A; Construct a comprehensive index of site influence based on indicator A and its weight; Determine the indicator B that is considered for the equipment impact and calculate the weight of indicator B; Construct a comprehensive index of equipment influence based on index B and its weight; Calculating values of the site influence comprehensive index and the equipment influence comprehensive index based on actual parameters of the undirected and unweighted complex network and the power optical transmission network; Identifying weak points of the power optical transmission network based on the values of the site influence comprehensive index and the equipment influence comprehensive index; The index A includes the node degree, clustering coefficient and voltage level of each node; node Comprehensive index of site influence for: ； in, 、 and are the weights of node degree, clustering coefficient and voltage level respectively; is the total number of nodes in the power optical transmission network; 、 It is a node in the power optical transmission network; 、 Node 、 Node degree 、 Normalized results; 、 Node 、 The average clustering coefficient of 、 Node 、 The average voltage coefficient of The indicator B includes the equipment aging degree, number of failures and bandwidth pressure of each node; equipment Comprehensive index of equipment influence for: . ； in, 、 and are the weights of bandwidth pressure, number of failures and aging degree respectively; is the total number of devices on the node; 、 The device for the node; 、 Equipment 、 Bandwidth pressure; 、 Equipment 、 Number of failures 、 Normalized results; 、 Equipment 、 degree of aging; Determine the weights between different indicators, specifically: The weights between different indicators are determined by the multi-attribute decision-making method. The algorithm is as follows: ； ； in, For the The first indicator attribute values, is the number of attribute values, is the number of indicators, For the The weight of an indicator.

2. The method for identifying weak points in a power optical transmission network based on node influence according to claim 1, characterized in that: node Node degree After normalization, we get , the formula is as follows: ； ； ； ； in, is the set of all nodes in the power optical transmission network, 、 It is a node in the power optical transmission network; node The average clustering coefficient as follows: ； point The clustering coefficient as follows: ； in, For nodes The number of edges between neighboring nodes; node The average voltage coefficient as follows: ； node Voltage coefficient Determined according to the voltage level of the node.

3. The method for identifying weak points in a power optical transmission network based on node influence according to claim 2, characterized in that: The voltage coefficients corresponding to node voltage levels of 500kV, 220kV, 110kV, and below 35kV are 4, 3, 2, and 1, respectively.

4. The method for identifying weak points in a power optical transmission network based on node influence according to claim 3 is characterized in that: equipment Bandwidth pressure as follows: ； in, 、 Bandwidth usage of the device; equipment After normalizing the number of failures, we get , the formula is as follows ； equipment Degree of aging as follows: ； in, For devices Years of operation, For equipment The aging years, For devices redundancy.

5. The method for identifying weak points in a power optical transmission network based on node influence according to claim 4 is characterized in that: The bandwidth usage of the device is the ratio of the used bandwidth of the device to the maximum bandwidth.

6. The method for identifying weak points in a power optical transmission network based on node influence according to claim 4, characterized in that: Redundancy The value of the device The number of homogeneous devices.

7. A storage medium, characterized in that: A computer program is stored, and when the program is executed, the method for identifying weak points in a power optical transmission network based on node influence as described in any one of claims 1 to 6 is implemented.

Images