A fault tree weighting-based intelligent substation communication link fault locating method

Abstract

The application belongs to the field of intelligent substation secondary system communication, and discloses a kind of intelligent substation communication link fault positioning method based on fault tree weighting, which solves the problems that massive information is difficult to identify and accurate positioning is impossible when intelligent substation communication link fails, and the method is oriented to the communication network frame of intelligent substation secondary system, establishes fault tree main branch for MMS network, SV network and GOOSE network respectively, forms fault tree nodes based on communication data type and networking mode, constructs typical node fault information model by using system communication alarm information, device alarm information and switch flow information, further combines the branch relationship of fault tree under fault link, adopts weighting method to establish fault probability model of communication link, and finally realizes fault positioning based on the comparison of fault identification results. The method has simple principle, uses less information, is suitable for multiple communication link faults, and improves the secondary operation efficiency of intelligent substation.

Description

Technical Field

[0001] This invention belongs to the field of communication technology for secondary systems of intelligent substations. It discloses a fault location method for communication links in intelligent substations based on fault tree weighting, which aims to simplify alarm information identification, achieve accurate location at the branch level in intelligent substations, intuitively reflect the fault location, and improve the efficiency of secondary operation and maintenance. Background Technology

[0002] Unlike traditional substations, smart substations use fiber optic communication instead of traditional cable media. The three-layer, two-network structure of smart substations brings convenience to information sharing within the station. However, the complex communication network connections within smart substations also lead to difficulties in locating communication link faults, resulting in generally low on-site maintenance efficiency. In fact, the logical relationships between alarm messages do not explicitly include link fault information. Therefore, the lack of intuitive and effective methods to understand the inherent logical relationships of alarm messages makes on-site troubleshooting difficult.

[0003] For the communication network of intelligent substations, analyzing and locating communication links based on different communication data types and networking methods is a basic approach. In terms of locating communication links in the secondary system of intelligent substations, some scholars have analyzed the mapping relationship between virtual and physical links in the process layer network and proposed a dynamic monitoring method for the process layer network based on process layer switches (Luo Linglu, Peng Qi, Wang Dehui, Li Chao, Shen Jian. Monitoring method for process layer network of intelligent substations [J]. Automation of Electric Power Systems, 2018, 42(11): 151-156.). The above method uses the process layer switch as the data acquisition and monitoring unit for the operation status of the process layer network and performs fault monitoring and diagnosis based on the network analyzer. However, the target is only the process layer network, and the introduction of the additional device, the network analyzer, increases the difficulty and cost of intelligent substation transformation to a certain extent. In addition, some literature has achieved visual location of fault links by parsing SCD files of smart substations (Wang Yan, Li Jin, Li Min, Guo Mingyu. Virtual and physical loop mapping and fault location method of smart substations [J]. Power Big Data, 2017, 20(12):74-79), but the above method can only locate the corresponding physical fault range. Some literature has also achieved the location of process layer network communication links by parsing SCD files and combining them with Path set (Cheng Lin, Liu Hongjun, Jiang Yi, et al. Method for fault location of process layer communication links in smart substations [P]. China: CN 107040413 B. 2020.08.04), but the location range of this method is limited to the process layer network in the three layers and two networks of smart substations, and fails to cover the entire communication network of smart substations.

[0004] However, the applicant believes that the efficiency of secondary operation and maintenance of smart substations can be further improved. Therefore, this application proposes a fault tree for the secondary system communication network interval, establishes a fault information model for MMS network nodes, a fault information model for SV network nodes, and a fault information model for GOOSE network nodes, adopts a weighted calculation method, and achieves precise location of the branch level of smart substations based on the weighted fault branch probability model and fault location criteria. Summary of the Invention

[0005] To improve the fault location capability of communication links in the secondary system of smart substations, and addressing the challenges of identifying massive amounts of information and accurately locating faults in these links, this invention proposes a fault tree-weighted fault location method for smart substation communication links. This method establishes a fault tree for the secondary system communication network intervals, and builds fault information models for MMS, SV, and GOOSE network nodes. Using a weighted calculation method, based on the constructed weighted fault branch probability model and fault location criteria, it achieves precise branch-level fault location in smart substations.

[0006] The technical solution adopted in this invention is: a fault tree-weighted method for fault location of communication links in intelligent substations, comprising the following steps:

[0007] Step 1: Establish the fault tree of the communication network bay of the intelligent substation secondary system;

[0008] Step 2: Establish an MMS network node fault information model;

[0009] Step 3: Establish an SV network node fault information model;

[0010] Step 4: Establish a GOOSE network node fault information model;

[0011] Step 5: Establish a weighted fault branch probability model;

[0012] Step 6: Fault location criteria.

[0013] The beneficial effects of this invention are as follows:

[0014] (1) This invention is aimed at communication alarms and communication network fault tree nodes. Fault location is intuitive and can better meet the operation and maintenance needs.

[0015] (2) The present invention is based on weighted calculation under multi-level nodes of the fault tree, which has high positioning accuracy and is applicable to complex fault situations;

[0016] (3) This invention does not require the addition of a communication network fault location device, which is beneficial for its application and promotion in both new and old substations. Attached Figure Description

[0017] Figure 1 This is a schematic diagram of the fault tree nodes of the intelligent substation bay in this application.

[0018] Figure 2 This is a schematic diagram of a typical bay structure of the intelligent substation described in this application.

[0019] Figure 3 This is a schematic diagram of the link L1 failure probability set calculation example in this application.

[0020] Figure 4 This is a schematic diagram of the link L5 failure probability set calculation example in this application.

[0021] Figure 5 This is a schematic diagram of the link L8 failure probability set calculation example in this application.

[0022] Figure 6 This is a schematic diagram of the link L1 and L3 complex failure probability set calculation example in this application.

[0023] Figure 7 This is a schematic diagram of the fault probability set calculation example for links L10 and L13 in this application.

[0024] Figure 8 This is a schematic diagram of the link L7 and L8 complex failure probability set calculation example in this application.

[0025] The names corresponding to the reference numerals in the above figures are as follows: Y1 General communication alarm, Y2 Station control layer network communication failure, Y3 Interval layer network communication failure, Y4 Upper layer of station control layer switch, Y5 Lower layer of station control layer switch, Y6 Interval layer device (M / S / G indicates communication network), Y7: SV network communication failure, Y8: GOOSE network communication failure, Y9: SV direct acquisition, Y10: SV network acquisition, Y11: Upper layer of SV switch, Y12: Lower layer of SV switch, Y13: GOOSE direct trip, Y14: GOOSE network trip, Y15: Upper layer of GOOSE switch, Y16: Lower layer of GOOSE switch, n: Total number of interval layer devices participating in MMS communication, m: Total number of interval layer devices participating in GOOSE communication, t: Total number of interval layer devices participating in SV communication. Detailed Implementation

[0026] Please see Figures 1 to 8 The embodiments of this application will be described in detail below with reference to the accompanying drawings. The present invention provides a fault tree-weighted intelligent substation communication link fault location method, which specifically includes the following steps:

[0027] Step 1: Establish the fault tree of the communication network bay of the intelligent substation secondary system;

[0028] To address the relative independence of protection and control in multi-bay systems of intelligent substations, a bay fault tree (hereinafter referred to as bay fault tree) is established based on the bay unit of the secondary system in intelligent substations. The nodes of the bay fault tree are established according to the communication data type and networking method. According to the communication data type, nodes are divided into MMS network faults, SV network faults, and GOOSE network faults. According to the networking method, nodes are divided into direct sampling (hop) / network sampling (hop) and upper / lower layer faults of the switch. Specifically, it includes the following parts:

[0029] Y1 General communication alarm, Y2 Station control layer network communication fault, Y3 Interval layer network communication fault, Y4 Upper layer of station control layer switch, Y5 Lower layer of station control layer switch, Y6 Interval layer device (M / S / G indicates communication network), Y7: SV network communication fault, Y8: GOOSE network communication fault, Y9: SV direct acquisition, Y10: SV network acquisition, Y11: Upper layer of SV switch, Y12: Lower layer of SV switch, Y13: GOOSE direct trip, Y14: GOOSE network trip, Y15: Upper layer of GOOSE switch, Y16: Lower layer of GOOSE switch;

[0030] Y1 to Y16 can be represented by the general formula Y i Y i Indicates the different node numbers of the fault tree of the interval. For the secondary system of the smart substation, the nodes are divided according to the communication data type and networking mode, and i takes the value of 1-16.

[0031] Based on the fault tree nodes of the secondary system interval, and according to the connection relationship and the principle of binary branching, the set of Y-shaped tree branches associated with the nodes is constructed as follows:

[0032] {Y1, Y2, Y3}, {Y2, Y4, Y5}, {Y5, YM6 i ... YM6 n}, {Y3, Y7, Y8}, {Y7, Y9, Y10}, {Y10, Y11, Y12}, {Y11, YS6 i ... YS6 m}, {Y8, Y13, Y14}, {Y14, Y15, Y16}, {Y15, YG6 i ... YG6 t}

[0033] The aforementioned YM6 i and YM6 n Essentially belonging to the same node (spacing layer device), YM6 i YM6 represents the i-th device in the corresponding interval layer during fault tree traversal. n This represents the nth device in the spacer layer; similarly, YS6 i YS6m Belonging to the same node, YS6 i YG6 represents the i-th device in the SV mesh spacer layer. i YG6 t Belonging to the same node, YG6 i This represents the i-th device in the GOOSE network interval layer; n represents the total number of interval layer devices, m represents the total number of interval layer devices participating in SV communication, and t represents the total number of interval layer devices participating in GOOSE communication.

[0034] Step 2: Establish the MMS network node fault information model; based on the alarm information and the interval fault tree established in Step 1 above, the information model of the total communication alarm node Y1 is as follows:

[0035]

[0036] F1 represents the fault information of the corresponding node Y1. Y1 is located at the top level of the fault tree of the above interval. F1 is used as the starting criterion for the weighted calculation of the fault tree. In the above formula, when the backend receives communication interruption / abnormal alarm information containing the current interval, F1 = 1. At this time, it is determined that a link fault has occurred in the secondary system, and it serves as the basis for judging downwards level by level; otherwise, F1 = 0.

[0037] The information model of MMS network fault-related nodes is based on the following F2, F4, F5, F M6i Architectural style:

[0038] Based on whether the alarm information contains MMS communication interruption information, the Y2 information model of station control layer network communication failure is established as follows: if the alarm information contains MMS communication interruption, it is considered that the link failure occurred in the station control layer MMS network, and the scope of the faulty link can be determined as the station control layer communication link in the MMS network.

[0039]

[0040] Based on the connection method of the station control layer network via the switch, the interruption of the upper layer link of the station control layer switch should correspond to the alarm of all devices in the interval layer. Based on this, the fault information model of the upper layer of the station control layer switch of node Y4 is established as follows: when the cumulative number of MMS communication alarm information of the i-th device in the interval layer generated by the monitoring system is equal to the total number of devices in the corresponding interval layer, the connection link between the upper layer link of the corresponding station control layer switch and the monitoring backend is as follows.

[0041]

[0042] Based on the connection method of the station control layer network via switches, a link failure in the lower layer of the station control layer switch corresponds to alarm information from one or more devices. To address multiple link failures, a fault information model for the lower layer of the station control layer switch at node Y5 is established as follows: the MMS communication alarm information of the i-th device in the interval layer generated by the monitoring system satisfies... At that time, the corresponding fault range belongs to the situation where the lower layer link of the station control layer switch is connected to the bay layer device;

[0043]

[0044] The station control layer switch is directly connected to the bay layer device. Therefore, an information model is directly constructed based on the alarm information of the bay layer device. If the alarm information contains I... MRi =1, which corresponds to a link failure between the i-th device in the bay layer and the station control layer switch in the MMS communication network;

[0045]

[0046] F2, F4, and F5 above represent the fault information for nodes Y2, Y4, and Y5, respectively. To distinguish faults of the interval layer device Y6 in different networks, the subscripts M, S, and G are used to distinguish interval layer devices belonging to the MMS network, SV network, and GOOSE network; for example, F... M6i This indicates that node Y6 belongs to the fault information model of the i-th device in the MMS network; where I MRi This represents the MMS communication alarm information of the i-th device in the interval layer generated by the monitoring system when MMS communication is interrupted, where n represents the total number of devices in the interval layer.

[0047] Step 3: Establish the SV network node fault information model. SV messages and GOOSE messages are used for information exchange between the process layer and the interval layer. The Y3 interval layer network communication fault model is established based on the presence of a total SV / GOOSE alarm in the alarm information, as shown in the following formula:

[0048]

[0049] The Y7 SV network communication fault is established based on whether there is an SV total alarm in the alarm information, as follows:

[0050]

[0051] This enables the differentiation between the SV and GOOSE communication links; where F3 and F7 represent the fault information of the corresponding nodes Y3 and Y7, respectively.

[0052] In a certain bay unit, the devices participating in SV communication have two connection methods: direct connection and networking. Considering that the SV direct connection link only includes the case from the protection device to the merging unit, a Y9 SV direct acquisition model is established for the case where there are alarm devices at the bay level and the merging unit is alarmed at the same time, as shown in the following formula, to achieve accurate positioning of the SV direct acquisition communication link.

[0053]

[0054] Where: F9 represents the fault information of the corresponding node Y9, I SRi I SMU These represent the communication alarm information of the i-th device or merging unit in the corresponding interval layer when SV communication is interrupted;

[0055] Combined with typical interval structures (such as) Figure 2 It is evident that the SV networking method exhibits a characteristic of having "many links at the upper layer (connected to bay layer devices) and few at the lower layer (connected to process layer devices) of the switch." Furthermore, in the event of a fault, it is impossible to determine whether the faulty link is located at the upper or lower layer of the switch based on alarm information. This invention establishes a Y10 SV network sampling model, as shown in the following formula:

[0056]

[0057] Wherein: F 10 For the fault information of the corresponding node Y10, I SSW This indicates an abnormal traffic alarm message from an SV switch based on the SNMP protocol.

[0058] Establish upper-layer models for the Y11 SV switch and lower-layer models for the Y12 SV switch, respectively F 11 F 12 This method utilizes the switch's traffic status (actively obtained by the station control and monitoring center based on the SNMP protocol) and the different numbers of devices that generate alarm information for upper and lower layer communication link failures in the switch to locate faulty links in the networking mode.

[0059]

[0060]

[0061] Wherein: F 11 F 12 These are the fault information for the corresponding nodes Y11 and Y12, respectively, where m represents the total number of devices participating in the SV network acquisition communication of this interval unit; For the upper layer of the corresponding switch, i.e., the communication link failure between the switch and the interval layer device, the number of devices that issue alarm information should be greater than or equal to 1 and less than m. In the case of a communication link failure between the lower layer of the switch and the merging unit, as can be seen from the typical interval topology, all upper-layer connection devices of the switch will issue alarm information due to the inability to receive the required data. The number of alarm devices is equal to m.

[0062] Considering that there are multiple communication links between the upper layer of the SV switch and the bay layer device, a model of the Y6 bay layer device under the SV communication network is established as follows, to further refine the fault link location range;

[0063]

[0064] Wherein: F S6i This indicates the fault information of the i-th device in the SV communication network;

[0065] Step 4: Establish a GOOSE network node fault information model; using the Y3 interval layer network communication fault model established in Step 3, combined with the Y8 GOOSE network communication fault model established based on whether there is a GOOSE total alarm in the alarm information, as shown in the following formula, the fault link range can be determined as the GOOSE communication link, and the fault tree nodes can be judged step by step downward.

[0066]

[0067] Similar to the SV communication network structure, devices participating in GOOSE communication within a certain interval unit also have two connection methods: direct connection and network connection (also known as direct jump and network jump connection methods). Based on the different alarm characteristics under the conditions of direct connection link failure and network link failure, the Y13 GOOSE direct jump model and the Y14 GOOSE network jump model are established respectively, as shown in the following two equations:

[0068]

[0069] Wherein: F 13 F 14 These are the fault information for nodes Y13 and Y14, respectively. GRi I GIT These represent the communication alarm information from the i-th device or intelligent terminal in the corresponding interval layer when GOOSE communication is interrupted; I GSW This indicates an abnormal traffic alarm message from a GOOSE switch based on the SNMP protocol.

[0070] If a communication link has been identified as faulty in a network configuration, to determine whether the fault lies at the interval layer or the process layer, establish upper-layer models for the Y15 GOOSE switch and lower-layer models for the Y16 GOOSE switch, as shown in the following two equations:

[0071]

[0072] A Y6 interval layer device model under the GOOSE communication network is established as follows, to further determine the location of the faulty link;

[0073]

[0074] Wherein: F 15 F 16 F G6i These are the fault information of the i-th device under the corresponding nodes Y15, Y16, and GOOSE communication network, respectively, and t represents the total number of devices participating in the GOOSE network jump communication of this interval unit;

[0075] Step 5: Establish a weighted fault branch probability model; considering that the range of suspected faults gradually narrows during the fault tree search, the initial judgment has the greatest impact on the location result. However, under various fault conditions, the closer to the upper level of the fault tree, the lower the distinguishability of its node information, so it is assigned a smaller proportion of weight; the information of the tree branches descending level by level is more correlated with the location of the fault link, so it is assigned a larger proportion of weight; the fault branch probability model is established by a step-by-step upward binary weighting method, and the corresponding communication link fault probability P is calculated based on the fault information of the fault tree nodes traversed by the communication link and the weights of the relevant nodes. Lx As shown in the following formula:

[0076]

[0077] Where: P Lx This represents the failure probability of the x-th communication link, where x is an indicator of the communication link number. The maximum value of x is the total number of links within the interval. For example... Figure 2 The typical interval shown has 13 communication links, so the value of x ranges from 1 to 13; s represents the total number of nodes traversed in the interval fault tree. Figure 1 The total number of all black solid nodes), in Figure 2 In the typical topology shown, there are 3 interval layer devices related to the MMS network, so n is 3; there are 2 interval layer devices related to the SV network, so m is 2; there are 3 interval layer devices related to the GOOSE network, so t is 3, therefore s is 23; p xi Let p represent the probability of the i-th node corresponding to the current x-th communication link. The set of node failure probabilities corresponding to the x-th communication link is p. x Sequence p x Each element in the array is represented as p. xi That is, p x =[p x1 p x2 …p xs ], p x The calculation is as follows:

[0078] px =FQ x

[0079] The multiplication mentioned above is dot product, which is the multiplication of corresponding elements in the array.

[0080] F represents the set of node fault information obtained based on the current fault alarm information. It is an array with 1 row and 23 columns, i.e., F = [F1 F2 F3 F4 F5 F M6i …F M6n F7 F8 F9 F 10 F 11 F 12 F S6i …F S6m F 13 F 14 F 15 F 16 F G6i …F G6t ], element F in F i F represents the fault information of the corresponding node. i Calculate according to the formulas in steps 2 to 4; when a communication link fails, obtain an F sequence as the fault information input.

[0081] Q x Q represents the set of node weighting coefficients corresponding to the x-th communication link Lx, which is also a 1x23 array. x medium element ω i Arranged as follows: Q x =[ω1ω2ω3ω4ω5ω M6i …ω M6n ω7ω8ω9ω 10 ω 11 ω 12 ω S6i …ω S6m ω 13 ω 14 ω 15 ω 16 ω G6i …ω G6t Since different links Lx only traverse a portion of the nodes during the interval fault tree traversal, Q... x Sparsity exists, Q x The non-zero elements in the formula are calculated according to the following formula:

[0082]

[0083] In the formula, j represents the number of non-zero elements in F. L represents all other possible numbers between 2 and j.

[0084] The calculated j non-zero elements are traversed according to the nodes of Lx, and ω1 to ω... j Values ​​are assigned sequentially from bottom to top to the corresponding non-zero nodes; all other nodes in the fault tree that have not been traversed are assigned 0. Since for a smart substation with a defined topology, the position and number of nodes in the fault tree traversed after a link failure are fixed, the weight Q for a link failure can be determined offline. x .

[0085] Step 6: Fault location criterion; determined by equation P Lx The failure probability P of each communication link was calculated. Lx (x=1,2,…,q) q represents the total number of links, resulting in the communication link failure probability set P=[P L1 ,P L2 ,…,P Lq By comprehensively comparing the failure probabilities of each communication link, a fault location criterion is established, as follows:

[0086]

[0087] Where: result is the result comparison output, Lx represents the xth communication link, and k represents the number of communication links in this interval unit; if there is one or more communication links (corresponding to single and multiple fault cases) with a fault probability of 1 in the calculation result, then it is output as a faulty link; if there is no communication link with a fault probability of 1 in the calculation result, then the communication link corresponding to the maximum value in the output result is a suspected faulty link and a corresponding alarm is issued.

[0088] Example Analysis:

[0089] To verify the effectiveness of the secondary system communication link fault location method proposed in this invention, this application uses a 220kV smart substation as an example to verify the effectiveness of the proposed method. The secondary system bay model used is as follows: Figure 2 As shown, the correspondence between link number Lx and the link is as follows: Figure 2 As indicated by the annotation. Corresponding Figure 2 Given n=3, m=2, t=3, F and Q mentioned in step 5 x It should be represented as a single-row array with a total of 23 elements.

[0090] According to ω j-i+1 and ω j The calculation formula yields the Q corresponding to each link Lx. x The following table lists the zero elements in the non-zero element category:

[0091]

[0092] Example 1: MMS network link failure example.

[0093] according to Figure 2 The topology of a certain bay in the intelligent substation shown is configured with faults in MMS network links L1, L2, and L4 to verify the effectiveness of the method proposed in this invention. Fault information of nodes is calculated according to the above formula, forming a node fault information set F. Following the rules described in step 5, a node weighting coefficient set Q is formed. x Calculate the failure probability P of each communication link according to the formula. Lx This forms a communication link probability set P within the interval unit. The communication link failure probability set P corresponding to a link L1 failure is as follows: Figure 3 As shown. Figure 3 Only the fault probability corresponding to L1 is 1. According to the location criterion described in step 6, the fault is located to link L1, which is the same as the assumed result, demonstrating the effectiveness of the method of the present invention under MMS. At the same time, according to the criterion for suspected faulty links described in step 6, L2, L3, and L4 are judged as suspected faulty links. This judgment locates the suspected links in other links of the MMS network besides the faulty links. The judgment range is reasonable, and relevant alarm information is issued to the suspected links. The simulation calculation results of other link faults in Example 1 are shown in the table below.

[0094]

[0095] Example 2: SV Network Link Failure Example

[0096] The SV network link failure simulation case sets up a direct-access link L5, and upper and lower layer links L6, L7, and L12 of the SV switch for case analysis and verification. In Case 2, the communication link failure probability set P corresponding to the failure of link L5 is as follows: Figure 4 As shown, Figure 4 The probability corresponding to only position L5 is 1. Based on the criterion described in step 6, the fault is located to link L5, which is the same as the assumed result. This shows the effectiveness of the method of the present invention for SV network faults. The simulation calculation results of other link faults in Example 2 are shown in the table below.

[0097]

[0098] Example 3: GOOSE Network Link Failure Example

[0099] The GOOSE network link failure simulation case sets up a direct sampling link L8 and upper and lower layer links L10 and L13 of the SV switch for case analysis and verification. In case 3, the communication link failure probability set P corresponding to the failure of link L8 is as follows: Figure 5 As shown, Figure 5 The probability of L8 position is 1. Based on the criterion described in step 6, the fault is located to link L8, which is the same as the assumed result. This shows the effectiveness of the method of the present invention for GOOSE network faults. The simulation calculation results of other link faults in example 3 are shown in the table below.

[0100]

[0101] Example 4: Complex Fault Example

[0102] The complex fault case study assesses the effectiveness of the method in handling multiple faults, primarily verifying the rationality of the fault tree node information and node weight calculations described in this invention. Complex fault scenarios include mixed MMS / GOOSE, mixed MMS / SV, mixed GOOSE / SV, and mixed scenarios across different links within the same network.

[0103] For L1 and L3 faults, L10 and L13 faults, and L7 and L8 faults, the fault probability set is plotted on... Figures 6-8 ,from Figures 6-8 As can be clearly seen, the method of this invention has a good fault identification capability when dealing with complex faults, and only the probability P under the corresponding faulty link. Lx The value of 1 indicates the reliability and effectiveness of the method. Simultaneously, according to the positioning criterion described in step 6, the maximum value P less than 1 in set P is... Lx For suspected faulty links, when links L1 and L3 fail, Max(P) L1 ,P L2 ,L,P Lk ) is P L2 and P L4 Links L2 and L3 are identified as suspected faulty links. This identification places the suspected links within the MMS network other than the faulty links. If the suspected range is reasonable, relevant alarm information is issued for the suspected links. The analysis process for other types of complex faults is similar. The calculation results for relevant complex fault examples are shown in the table below.

[0104]

[0105] The method in this invention application targets the communication network architecture of secondary systems in intelligent substations. It establishes fault tree branches for MMS, SV, and GOOSE networks, forming fault tree nodes based on communication data types and networking methods. Utilizing system communication alarm information, device alarm information, and switch traffic information, it constructs a typical node fault information model. Furthermore, combining the fault tree branch relationships under the faulty link, it establishes a fault probability model for the communication link using a weighted approach. Finally, it achieves fault location based on the comparison of fault identification results. This method is simple in principle, requires minimal information, and adapts to multiple communication link faults, improving the efficiency of secondary operation and maintenance in intelligent substations.

Claims

1. A method for fault location of communication links in intelligent substations based on fault tree weighting, characterized in that... It includes the following steps: Step 1: Establish the fault tree of the communication network bay of the intelligent substation secondary system; Step 2: Establish an MMS network node fault information model; Step 3: Establish an SV network node fault information model; Step 4: Establish a GOOSE network node fault information model; Step 5: Establish a weighted fault branch probability model; Step 6: Fault location criteria; The method for establishing a fault tree for the communication network bay of a smart substation secondary system is as follows: A fault tree for the communication network bay of a smart substation secondary system is established with nodes as follows: [Y1 General communication alarm, Y2 Station control layer network communication fault, Y3 Bay layer network communication fault, Y4 Upper layer of station control layer switch, Y5 Lower layer of station control layer switch, Y6 Bay layer device, Y7 SV network communication fault, Y8 GOOSE network communication fault, Y9 SV direct acquisition, Y10 SV network acquisition, Y11 Upper layer of SV switch, Y12 Lower layer of SV switch, Y13 GOOSE direct trip, Y14 GOOSE network trip, Y15 Upper layer of GOOSE switch, Y16 Lower layer of GOOSE switch]. This is referred to as the bay fault tree. Its associated generalized Y-shaped tree branch sets are {Y1, Y2, Y3}, {Y2, Y4, Y5}, and {Y5, YM6}, respectively. i ... YM6 n }, {Y3, Y7, Y8}, {Y7, Y9, Y10}, {Y10, Y11, Y12}, {Y11, YS6 i ... YS6 m }, {Y8, Y13, Y14}, {Y14, Y15, Y16}, {Y15, YG6 i ... YG6 t }, of which YM6 i YS6 is the i-th device in the MMS network spacer layer. i For the i-th device in the SV mesh spacer layer, YG6 i Let be the i-th device in the GOOSE network interval layer, n represent the total number of devices in the interval layer, m represent the total number of devices in the interval layer participating in SV communication, and t represent the total number of devices in the interval layer participating in GOOSE communication. The method for establishing the MMS network node fault information model is as follows: Based on alarm information and the established interval fault tree, the Y1 communication total alarm node information model is established, as shown in the following formula: F1 represents the fault information of the corresponding node Y1. Y1 is located at the top level of the fault tree of the above interval. F1 is used as the starting criterion for the weighted calculation of the fault tree. In the above formula, when the backend receives communication interruption / abnormal alarm information containing the current interval, F1 = 1. At this time, it is determined that a link fault has occurred in the secondary system, and it serves as the basis for judging downwards level by level; otherwise, F1 = 0. Establish network communication faults at the Y2 station control layer, the upper layer of the Y4 station control layer switch, the lower layer of the Y5 station control layer switch, and YM6 respectively. i The fault information model of MMS network nodes such as the interval layer device, that is, the information model of nodes related to MMS network faults, is based on the following F2, F4, F5, F M6i Architectural style: Among them, F1, F2, F4, F5, F M6i These are the fault information for nodes Y1, Y2, Y4, Y5, and Y6 belonging to the i-th device in the MMS network, respectively. MRi This represents the MMS communication alarm information of the i-th device in the interval layer generated by the monitoring system when MMS communication is interrupted, where n represents the total number of devices in the interval layer. The method for establishing the SV network node fault information model is as follows: Based on whether there is an SV / GOOSE total alarm in the alarm information, a Y3 interval layer network communication fault model is established, as shown in the following formula: The Y7 SV network communication fault is established based on whether there is an SV total alarm in the alarm information, as follows: To distinguish between SV and GOOSE communication links; For cases where alarm devices exist in the interval layer and unit alarms are merged, a Y9 SV direct acquisition model is established as follows, to achieve accurate positioning of the SV direct acquisition communication link; Establish the Y10 SV network procurement model as follows: Establish upper-layer models for the Y11 SV switch and lower-layer models for the Y12 SV switch, respectively F 11 F 12 Mode: The number of devices issuing alarm information at the upper layer of the corresponding switch should be greater than or equal to 1 and less than m. As can be seen from the typical bay topology, in the case of the lower layer of the switch, all the upper layer connected devices of the switch will issue alarm information because they cannot receive the required data. The number of alarm devices is equal to m. The Y6 interval layer device model under the SV communication network is established as follows: Among them, F3, F7, F9, F 10 F 11 F 12 F S6i These are the fault information of the i-th device in the communication network corresponding to nodes Y3, Y7, Y9, Y10, Y11, Y12, and SV, respectively. SRi I SMU These represent the communication alarm information of the i-th device or merging unit in the corresponding interval layer when SV communication is interrupted; I SSW This indicates an abnormal traffic alarm message from the SV switch based on the SNMP protocol; m represents the total number of devices participating in the SV network acquisition communication of this interval unit. The method for establishing the GOOSE network node fault information model is as follows: using the established Y3 interval layer network communication fault model, combined with the Y8 GOOSE network communication fault model established based on whether there is a GOOSE total alarm in the alarm information, as shown in the following formula; Based on the different alarm characteristics under direct communication link failure and network link failure scenarios, the Y13GOOSE direct hop model and the Y14 GOOSE network hop model are established respectively, as shown in the following two equations: If a communication link has been identified as faulty in a network configuration, to determine whether the fault lies at the interval layer or the process layer, establish upper-layer models for the Y15 GOOSE switch and lower-layer models for the Y16 GOOSE switch, as shown in the following two equations: A Y6 interval layer device model under the GOOSE communication network is established as follows, to further determine the location of the faulty link; Among them, F8, F 13 F 14 F 15 F 16 F G6i These correspond to nodes Y8 and Y respectively. 13 Y 14 Y 15 Y 16 Fault information of the i-th device in the GOOSE communication network, I GRi I GIT These represent the communication alarm information from the i-th device or smart terminal in the corresponding interval layer when GOOSE communication is interrupted; I GSW This indicates an abnormal traffic alarm message from the GOOSE switch based on the SNMP protocol; t represents the total number of devices participating in the GOOSE network jump communication within this interval unit. The method for establishing the weighted fault branch probability model is as follows: Considering that tree branch information is more relevant to fault link location during fault tree search, a step-by-step downward binary search weighted approach is used to establish the fault branch probability model as shown in the following equation: Where: P Lx Let represent the failure probability of the x-th communication link, where x is the link number and its maximum value is the total number of links within the interval; s represents the total number of nodes traversed in the interval fault tree; p xi Let p represent the probability of the i-th node corresponding to the current x-th communication link; where the set of failure probabilities of nodes corresponding to the x-th communication link is p. x Sequence p x Each element in the array is represented as p. xi That is, p x =[p x1 p x2 …p xs ], p x The calculation is as follows: p x =FQ x The multiplication described above is dot product, which is the multiplication of corresponding elements in an array; F represents the set of node fault information obtained based on the current fault alarm information, i.e., F = [F1 F2 F3 F4 F5F] M6i …F M6n F7 F8 F9 F 10 F 11 F 12 F S6i …F S6m F 13 F 14 F 15 F 16 F G6i …F G6t ], element F in F i Q represents the fault information of the corresponding node. x Let Q represent the set of node weighting coefficients corresponding to the x-th communication link Lx. x medium element ω i Arranged as follows: Q x =[ω1ω2ω3ω4ω5ω M6i …ω M6n ω7ω8ω9ω 10 ω 11 ω 12 ω S6i …ω S6m ω 13 ω 14 ω 15 ω 16 ω G6i …ω G6t ], Q x The non-zero elements in the formula are calculated according to the following formula: In the formula, j represents the number of non-zero elements in F; L represents all other possible numbers between 2 and j; The calculated j non-zero elements are traversed according to the nodes of Lx, and ω1 to ω... j Assign values ​​to the corresponding non-zero nodes from bottom to top; assign 0 to all other nodes in the fault tree that have not been traversed. The method for fault location criteria is as follows: By comprehensively comparing the fault probabilities of each communication link, a fault location criterion is established, as shown in the following formula: Where: result is the result comparison output, Lx represents the xth communication link, and k represents the number of communication links in this interval unit; if there is one or more communication links with a failure probability of 1 in the calculation result, then it is output as a faulty link; if there is no communication link with a failure probability of 1 in the calculation result, then the communication link corresponding to the maximum value in the output result is a suspected faulty link and a corresponding alarm is issued.

Images