Single-phase grounding fault section location method of active distribution network based on multi-point synchronization information under 5G communication

Abstract

The present invention discloses a method for locating a single-phase grounding fault section of an active distribution network based on multi-point synchronization information under 5G communication. The method targets a distributed new energy access distribution network system in which a single-phase grounding fault occurs. According to the 5G ultra-high reliability and low-latency communication architecture, the method obtains the multi-point synchronization direction information of the distribution network, and constructs a complex fault section locating technology based on the processes of positioning objective function construction, linearization, and solution. The positioning technology realizes transient zero-sequence direction discrimination according to the difference in the main resonant frequency bands of the upstream and downstream when the system has a single-phase grounding fault, proposes a new definition of the switching function and a section positioning objective function, and further uses the CPLEX solver to solve the linear programming model to realize fault identification and section positioning. The method of the patent is novel, and the linearized logical relationship positioning model is not only applicable to single faults, but also to multiple faults. Its application in the new energy access distribution network system can accelerate the rapid identification and recovery of faults in new power systems.

Description

Technical Field

[0001] The present invention relates to the technical field of a section positioning method, and more specifically to the technical field of a section positioning method for a single-phase grounding fault in an active power distribution network. Background Art

[0002] After a large number of distributed power sources are connected, the distribution network changes from a single-power radial network to a multi-source multi-directional network. Accurately and quickly locating line faults is an important part of improving the power supply reliability of active distribution networks. According to the grounding method and fault type of the distribution system, the section positioning method is generally divided into a small current section positioning method and a large current section positioning method. Under a small current fault, due to the weak change of the three-phase power frequency fault current, most of these section positioning methods use zero-sequence current, zero-sequence voltage or high-frequency harmonics as the fault signal source for positioning. According to the fault signal source and positioning principle, this type of section positioning method can be divided into a method based on FTU telemetry data and a method based on FTU telesignaling data; the methods based on telemetry data include steady-state zero-sequence current comparison method, transient zero-sequence current similarity method, 5th harmonic method, zero-sequence admittance method, centroid frequency method, etc.; the methods based on telesignaling data include: steady-state zero-sequence current direction method, transient zero-sequence current direction method. The small current section positioning method mostly adopts a distributed positioning structure. Although this structure is simple and reliable, in terms of positioning essence, its fault section identification only uses the fault data of adjacent nodes, and the positioning fault tolerance is poor.

[0003] In recent years, the segment positioning optimization methods based on telesignaling data have made great progress, but the linear segment positioning optimization methods based on algebraic relationships still have the following shortcomings: 1) It has high solution efficiency and numerical stability, meeting the requirements of online positioning, but ignores the influence of fault signal transmission time, underreporting and false alarm in the communication process, and cannot achieve multi-point online synchronous positioning; 2) The optimization model is constructed according to the distribution of three-phase fault current direction. Although it has good fault tolerance, it is only suitable for large current faults in distribution networks, and is not applicable to small current faults with the lowest probability of occurrence in distribution networks; 3) It only solves the problem of single faults and cannot accurately locate under multiple faults. Summary of the invention

[0004] In order to solve the problem of single-phase grounding fault when new energy is connected to the distribution network, the fault current is small, the voltage of the whole network is consistent, and the line voltage is symmetrical. It is difficult to achieve high-precision positioning by only using the zero-sequence fundamental component to realize fault location, which is affected by the grounding of the arc suppression coil. The access of new energy to the distribution network further changes the unidirectional radiation of the distribution network power supply, making it difficult for traditional section positioning technology to adapt. Based on this, the present invention proposes a single-phase grounding fault section positioning method for an active distribution network based on multi-point synchronous information under 5G communication. The method uses the transient zero-sequence direction of the system and the expected value calculated based on the switching function to construct an objective function, and introduces linearization processing and solution to realize fault identification and section positioning. The method is novel, and the linearized logical relationship positioning model is not only applicable to single faults, but also to multiple faults.

[0005] The technical solution adopted by the present invention is:

[0006] A method for locating single-phase grounding fault sections in active distribution networks based on multi-point synchronization information under 5G communication includes 5G communication network architecture, transient zero-sequence direction discrimination frequency band selection, switching function definition and linearization, objective function construction and linearization, and linearization model solution. The specific steps are as follows:

[0007] Step 1: Design a 5G-based segment positioning communication architecture based on ultra-high reliability and low latency communication services and segment positioning network slicing technology;

[0008] Step 2: Taking the main resonant frequencies of the upstream and downstream nodes of the fault point as the lower and upper boundary frequencies of the characteristic frequency band, and retaining a certain margin, the interactive frequency band is selected as the characteristic frequency band for determining the direction of transient zero-sequence current;

[0009] Step 3: According to the distribution characteristics of the transient zero-sequence current direction, a switching function suitable for single-phase grounding fault is constructed according to the new definition of node state, and the switching function is linearized by the left and right side approximation method;

[0010] Step 4: Construct the section positioning objective function based on the measured value and expected value of the transient zero-sequence current direction, and introduce auxiliary variables to linearize the objective function;

[0011] Step 5: Use CPLEX solver to solve the linear programming model to achieve section location and fault isolation.

[0012] The beneficial effects of the present invention are:

[0013] (1) Based on 5G communication, regional measurement information interaction is realized, and it has obvious advantages in protecting delay and reliability;

[0014] (2) The objective function model constructed according to the characteristics of transient zero-sequence current direction distribution can accurately locate the fault section under single-phase grounding fault;

[0015] (3) The switching function is linearized by the left-right approximation method, making the entire linearized positioning model applicable not only to single faults but also to multiple faults. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] Figure 1 The 5G communication network architecture of the present invention;

[0017] Figure 2 It is a timing diagram of the protection action in the present invention;

[0018] Figure 3 This is a structural diagram of an active distribution network containing 4 feeders in the present invention;

[0019] Figure 4 It is a characteristic diagram of transient zero-sequence current direction distribution in the present invention;

[0020] Figure 5 is the topological structure of the feeder 4 in the present invention;

[0021] Figure 6 It is the process of determining the direction of transient zero-sequence current of nodes S2 and S3 in the present invention;

[0022] Figure 7 Selection of a fault frequency band for determining the direction of transient zero-sequence current in the present invention;

[0023] Figure 8 It is the T-type distribution network topology structure in the present invention;

[0024] Fig. 9 This is a multi-power source multi-T-coupled node distribution network in the present invention. DETAILED DESCRIPTION

[0025] The method for locating a single-phase grounding fault section of an active distribution network based on multi-point synchronization information under 5G communication provided by the present invention specifically includes the following steps:

[0026] Step 1: 5G communication network architecture

[0027] 5G communication generally has three application scenarios, namely ultra-reliable and low latency communications (ultra-Reliable and Low Latency Communications, uRLLC), enhanced mobile broadband, and massive machine-type communications. The synchronous protection of the distribution network belongs to the uRLLC service in the three major application scenarios of 5G, which has strict requirements on transmission delay and reliability. According to the mobile standard definition, a centralized communication network architecture is mainly composed of terminals, base stations, core networks, and master station servers. Except for the terminals and base stations, which belong to air interface transmission, the others are in the form of optical fiber aggregation. In order to meet the ultra-low latency requirements of the uRLLC service, it is necessary to introduce network slicing technology in the 5G architecture. Therefore, the communication network architecture based on multi-point synchronous protection of active distribution networks for 5G communications is as follows Figure 1 shown.

[0028] From the occurrence of a fault to the time when the section switch or interconnecting switch receives the action signal and isolates the fault, the entire protection process based on 5G communication includes the following links: 1) Fault detection. The delay of this link includes the set fault signal sampling cycle length and the inherent time of FTU startup and calculation; 2) Remote signaling data upload. The delay of this link includes the transmission delay from FTU to 5G base station, the transmission delay from 5G base station to core network, and the transmission delay from core network to main station; 3) Fault section identification. The delay of this link mainly depends on the linear state of the positioning model, variable dimension, solution method and other factors; 4) Remote control data downlink. The delay of this link is the same as the remote signaling data upload time; 5) Switch action. This link is mainly the inherent action time of the section switch and interconnecting switch after receiving the action signal. The action timing diagram of the entire protection process is shown as follows. Figure 2 shown.

[0029] Step 2: Transient zero-sequence direction determination frequency band selection

[0030] For the distribution network connected to new energy, the main resonant component of the node upstream of the fault point and the main resonant component of the node downstream of the fault point are different. Whether the main resonant component of the upstream node or the main resonant component of the downstream node is selected, the reliability of the transient zero-sequence current direction determination will be reduced. In order to improve the reliability of the transient zero-sequence current direction determination, this patent selects the lower boundary frequency based on the main resonant frequency of the node upstream of the fault point as the characteristic frequency band, selects the upper boundary frequency based on the main resonant frequency of the node downstream of the fault point as the characteristic frequency band, and retains a certain margin, and selects the interactive frequency band as the characteristic frequency band for transient zero-sequence current direction determination.

[0031] Step 3: Distribution characteristics of transient zero-sequence current direction

[0032] An active distribution network with four feeders is used as an example to analyze the distribution characteristics of the transient zero-sequence current direction under a single-phase fault. Distributed power sources are connected to the middle of feeder 3 and the end of feeder 4. The neutral point of the main power source is grounded through an arc suppression coil, and the neutral point of the distributed power source is not grounded. The specific structure Figure 3 shown.

[0033] In a system with an ungrounded neutral point, the compensation current i of the arc suppression coil L = 0, therefore, all transient zero-sequence currents in the line are capacitive. In the neutral point grounding system through the arc suppression coil, in the characteristic frequency band from the third harmonic to the first series resonance, the overcompensation effect of the arc suppression coil can be ignored, and all transient zero-sequence currents are also capacitive. If a single-phase grounding fault is set in section 3 of feeder 4, the distribution of the transient zero-sequence current direction of the active distribution network is as follows: Figure 4 shown.

[0034] If the main power supply flows to the load in the positive direction of transient zero-sequence current, then Figure 4 The distribution characteristics of the transient zero-sequence current direction can be found in:

[0035] 1) Reverse transient zero-sequence current flows through nodes on the path from the fault point to the bus, such as nodes S1, S2, and S3;

[0036] 2) Forward current flows through nodes on the healthy branch and non-fault line upstream of the fault point, such as nodes S5 and S6;

[0037] 3) Forward current flows through all nodes downstream of the fault point, such as node S4.

[0038] Step 4: Switching function definition and linearization

[0039] (1) Switch function definition

[0040] According to the distribution characteristics of the transient zero-sequence current direction, the node state I i A new definition is given as shown in formula (1):

[0041]

[0042] According to the definition of formula (1), a switching function I suitable for single-phase grounding fault is constructed. j (X):

[0043]

[0044] Where [j, d] represents the set of sections downstream of node j, ∪ represents the logical “or”, and X is the expected set of line section faults X = [x1, x2,, x N ]. This switching function is not only applicable to single-power distribution networks, but also to distributed power distribution networks.

[0045] …

[0046] The power distribution network is connected to the power grid.

[0047] (2) Linearization of the switching function

[0048] In order to make the linearized switching function also applicable to multiple faults, the present invention performs linearization approximation on it from the left and right sides, and the approximation method is shown in formula (3).

[0049]

[0050] Wherein, K represents the maximum fault multiplicity limit, and this patent sets K=3; Σ represents the algebraic “+”.

[0051] Step 5: Objective function construction and linearization

[0052] (1) Objective function

[0053] According to the minimum difference idea, the objective function constructed using the transient zero-sequence current direction is shown in formula (4):

[0054]

[0055] Where F(X) represents the objective function value, N is the number of FTUs at the switch, i.e., the number of nodes; j The FTU node status actually received by the master station. According to the segment state assumption X = [x1, x2, ..., x N ] and the FTU node state expectation calculated by the switching function I(X), ω is the weight coefficient, usually 0.5. i It is the segment status, and its encoding rules are:

[0056]

[0057] (2) Linearization of the objective function

[0058] In order to transform the nonlinear programming model into a linear integer programming model, the absolute value operation in the objective function (4) is removed. Two auxiliary variables P = [p1…p N ] and Q=[q1…q N ], transform the objective function (3) into a linear objective function, as shown in (6).

[0059]

[0060] Step 6: Linearized model solution

[0061] The CPLEX solver is used to solve the linear programming model by finding the minimum value of the objective function, and the corresponding decision variable X = [x1, x2, ... x j ,…,x N ]=[00,…,1,…,0]; then x j Recorded as the fault section.

[0062] Calculation example

[0063] (1) Determination of the direction of transient zero-sequence current

[0064] Will Figure 3 In the active distribution network, feeder 4 is set as an overhead line and divided into 5 nodes and 5 sections, where the length of each section is set to 5 km. The feeder topology is as follows: Figure 5 shown.

[0065] In section 2, a single-phase ground fault is set, and the grounding resistance R f =50Ω, fault angle φ=300, FTU collects zero-sequence current and line voltage of 5 nodes, and determines the direction of transient zero-sequence current of each node according to the method in the literature (Zhang Linli, Xu Bingyin, Xue Yongduan, et al. Transient location method of small current grounding fault based on line voltage and zero-mode current [J]. Proceedings of the CSEE, 2012, 32(13):110-115). The results are shown in Table 1, where Figure 6 The process of distinguishing the direction of transient zero-sequence current at the upstream node S2 and the downstream node S3 of the fault point is given. Regarding the extraction of the transient zero-sequence current characteristic frequency band in direction discrimination, this document selects the main resonant component of the zero-sequence current at the downstream node of the fault point as the transient zero-sequence current characteristic frequency band. Since the main resonant component of the upstream node of the fault point and the main resonant component of the downstream node of the fault point are different, whether the main resonant component of the upstream node or the main resonant component of the downstream node is selected, the reliability of transient zero-sequence current direction discrimination will be reduced. For example, Figure 7 In the figure, the main resonant component of the node upstream of the fault point is around 200HZ, while the main resonant component of the node downstream of the fault point is around 1400HZ. If the frequency band near 1400HZ is used as the characteristic frequency band, the amplitude of the upstream of the fault point is less than 0.724A; if the frequency band near 200HZ is used as the characteristic frequency band, the amplitude of the downstream of the fault is less than 0.153A; in order to improve the reliability of the transient zero-sequence current direction determination, this patent uses the main resonant frequency of the node upstream of the fault point as the lower boundary frequency of the characteristic frequency band, and the main resonant frequency of the node downstream of the fault point as the upper boundary frequency of the characteristic frequency band, and retains a certain margin, and selects 150HZ~2000HZ as the characteristic frequency band for transient zero-sequence current direction determination.

[0066] Table 1 Transient zero-sequence current direction and state coding of each node

[0067]

[0068] In addition, the influence of different grounding resistance and fault angle on the direction determination of transient zero-sequence current is considered, and the results are shown in Table 2 and Table 3.

[0069] Table 2 Direction determination results under different grounding resistances (φ=900)

[0070]

[0071] Table 3 Direction determination results under different fault angles (R f =50Ω)

[0072]

[0073] It can be seen from Table 3 and Table 4 that the proposed improved transient zero-sequence current direction determination method can accurately determine the fault current direction in all cases except for the misjudgment of the S3 node when the grounding resistance is 500 ohms; compared with the direction determination method using the frequency band near the main resonant component, the proposed transient zero-sequence current direction determination method has higher determination reliability.

[0074] (2) T-type distribution network verification

[0075] Will Figure 3 The feeder 4 of the active distribution network is set to a T-type structure, such as Figure 8 As shown, it includes 7 nodes and 7 sections, and a distributed power source is connected at the end of section 5.

[0076] The linear integer programming model for segment positioning is established as follows:

[0077]

[0078] A single-phase grounding fault is set in section 6. The FTU collects the transient zero-sequence current direction of 7 nodes and obtains its state code according to the coding rules as shown in Table 4.

[0079] Table 4 Zero-sequence current direction and state coding of each node in T-type distribution network

[0080]

[0081] The collected steady-state and transient state codes of each node are brought into the linear integer programming model and solved by CPLEX to obtain the decision variable X = [x1, x2, x3, x4, x5, x6, x7] = [0000010]; thus, it is determined that the fault occurs in section 6. From the positioning results, it can be seen that since distributed power sources usually adopt the neutral point ungrounded mode, the access of distributed power sources will not affect the distribution characteristics of the zero-sequence current direction. Therefore, the proposed method can also accurately locate the fault section in the active distribution network connected to the distributed power source.

[0082] In order to verify the adaptability of the proposed method to small current faults, single-phase grounding faults are set at different locations, and the proposed method is compared with the section location method based on the three-phase fault current direction (logic model and linear model). The comparative location results are shown in Table 5.

[0083] Table 5 Location results under single-phase ground fault

[0084]

[0085]

[0086] It can be seen from the positioning results in Table 5 that the proposed positioning method can accurately locate the fault section under single-phase grounding faults at different locations. However, the section positioning method based on the direction of three-phase fault current cannot locate the fault section under single-phase grounding faults. Therefore, the proposed method has good adaptability to small current faults.

[0087] (3) Verification of multi-source multi-T-coupled node distribution network

[0088] Build a multi-power multi-T-coupled node distribution network, such as Fig. 9 The distribution network includes a circuit breaker, 27 section switches and tie switches, corresponding to 28 feeder sections respectively; distributed power sources are connected at the ends of sections 10, 20 and 28 respectively.

[0089] The linear integer programming model of the multi-source multi-T-coupled node distribution network is established:

[0090]

[0091] In order to verify the adaptability of the proposed positioning method to multiple faults, single-phase double faults and single-phase triple faults are set in different sections of the distribution network, and the influence of node status false alarms on the positioning results is considered. The positioning results under different conditions are shown in Table 6.

[0092] Table 6 Location results under multiple faults

[0093]

[0094]

[0095] It can be seen from Table 6 that the proposed method can accurately locate the fault section when multiple single-phase grounding faults occur in the distribution network. This is mainly due to the fact that the switch function in the model is linearized by the left-right clamping method, which gets rid of the limitation of single faults.

[0096] (4) Protection delay and reliability analysis

[0097] Take the example to verify the advantages of the proposed method in protecting delay. Fig. 9 In the distribution network shown, the positioning time of the nonlinear model and the linear model are respectively counted, and with reference to the relevant delay data and the communication architecture proposed in this patent, a delay comparison of various protection schemes is obtained, as shown in Table 7.

[0098] Table 7 Delay comparison of various protection schemes

[0099]

[0100] It can be seen from Table 7 that the delay of the protection scheme using 5G communication is significantly shorter than that of the protection scheme using other communication methods; the solution of the nonlinear model requires the use of an intelligent algorithm, and the positioning delay is significantly longer than that of the linear model; compared with other protection schemes, the proposed protection scheme "5G+linear positioning model" has the smallest protection delay, can achieve online positioning and fault isolation, and better meet the requirements of multi-point synchronous protection of active distribution networks.

[0101] Taking the example to verify the reliability advantage of the proposed method, the number of missed and false alarms of 5G communication is limited to a single point, and only involves missed alarms, and the number of missed and false alarms of GPRS and 4G communication is limited to 2 points, involving missed alarms and false alarms. Define "reliability = number of correct positioning times / number of all missed and false alarm scenarios", in Fig. 9 A single fault is set in section 7 of the distribution network shown in the figure. All positioning models adopt the proposed linear positioning model. The protection reliability under different communication modes is statistically calculated as shown in Table 8.

[0102] Table 8 Reliability comparison of various protection schemes

[0103]

[0104]

[0105] It can be seen from Table 8 that the reliability of the protection scheme using 5G communication is higher than that of the protection scheme using other communication methods; due to the absence of immature convergence in the linear model, the reliability of the linear model is slightly higher than that of the nonlinear model under the same communication mode; the combined reliability of the "5G+linear model" can reach 96.42%, and the reason for not accurately locating the fault is that when the underreporting location occurs on both sides of the fault point, the two adjacent sections are combined into one section. In this scenario, it is difficult to solve the problem through the fault tolerance of the model, and protection action verification is required to accurately determine the fault location.

Claims

A method for locating single-phase grounding fault sections in an active distribution network based on multi-point synchronization information under 1.5G communication is characterized by: The method includes 5G communication network architecture, transient zero-sequence direction discrimination frequency band selection, switching function definition and linearization, objective function construction and linearization, and linearization model solution. The transient zero-sequence direction discrimination frequency band selection is specifically as follows: considering that the main resonant component of the node upstream of the fault point and the main resonant component of the node downstream of the fault point are different, a single frequency will reduce the reliability of transient zero-sequence current direction discrimination, it is proposed to construct a characteristic frequency band using the main resonant frequency of the node upstream of the fault point as the main resonant frequency of the node downstream of the fault point as the main resonant frequency of the node to select the upper boundary frequency, and retain a certain margin to select the interactive frequency band as the characteristic frequency band for transient zero-sequence current direction discrimination. The switching function definition and linearization are as follows: Step 1: Switch function definition According to the distribution characteristics of the transient zero-sequence current direction, a switching function I suitable for single-phase grounding fault is constructed. j (X): Where [j, d] represents the set of sections downstream of node j, ∪ represents logical "or", and X is the set of line section faults X = [x1, x2,, x N ], where 1 is the positive direction and 0 is the negative direction. This switching function is not only applicable to single-power distribution networks, but also to distribution networks with distributed power sources connected; Step 2: Linearization of the switching function In order to make the linearized switching function also applicable to multiple faults, this method linearizes and approximates it from the left and right sides. The approximation method is shown in formula (2): Where K represents the maximum fault multiplicity, and K=3; ∑ represents the algebraic "+", The objective function construction and linearization are specifically as follows: Step 1: Objective function construction According to the minimum difference idea, the objective function constructed using the transient zero-sequence current direction is shown in formula (3): Where F(X) represents the objective function value, N is the number of FTUs at the switch, i.e., the number of nodes; j The FTU node status actually received by the master station. According to the segment state assumption X = [x1, x2, ..., x N ] and the FTU node state expectation calculated by the switching function I(X), ω is the weight coefficient, usually 0.5, x i is the section status, 1 means fault in the section, 0 means no fault in the section; Step 2: Linearization of the objective function In order to transform the nonlinear programming model into a linear integer programming model, the absolute value operation in the objective function formula (4) is removed, and two auxiliary variables P = [p1…pN] and Q = [q1…qN] are introduced to transform the objective function formula (2) into a linear objective function, as shown in formula (4): Based on the absolute value operation being greater than 0, construct Equation, where x i =0 / 1,X=[x1,x2,...x N ], p i =0 / 1,P=[p1,p2,...p N ], q i =0 / 1,Q=[q1,q2,...q N ].

2. According to the method for locating single-phase grounding fault section of active distribution network based on multi-point synchronization information under 5G communication in claim 1, it is characterized in that: The 5G communication network architecture is specifically as follows: based on ultra-high reliability and low-latency communication services, it meets the delay requirements from the occurrence of distribution network faults to the section switch or interconnection switch receiving the action signal and isolating the fault, introduces section positioning network slicing technology into the 5G architecture, and realizes air interface transmission from the base station to the core network.

3. According to the method for locating single-phase grounding fault section of active distribution network based on multi-point synchronization information under 5G communication in claim 1, it is characterized in that: The linear model is solved as follows: Based on the objective function and the switch function linear programming model, the CPLEX solver is used to solve and obtain the minimum value of the objective function, and the corresponding decision variable X = [x1, x2, ... x j ,…,x N ]=[00,…,1,…,0]; then x j Recorded as the fault section.

Images