Power distribution network fault positioning method and system based on load impedance

Abstract

The invention relates to the technical field of power systems, in particular to a power distribution network fault positioning method and system based on load impedance, and the method comprises the following steps: classifying faults of a distributed power supply power distribution network; fault positioning is carried out in the completely observable power distribution network according to current and voltage data recorded by the miniature phasor measurement unit; in the partial observable power distribution network, a fast reverse or forward load flow calculation method is adopted to obtain initial load impedance of each node, and based on the initial load impedance of each node, a particle swarm optimization algorithm is adopted to optimize complex power of each node to obtain a load value of each node; in the partial observable power distribution network, the fault distance of each section is calculated by adopting a least square method for the fault position. Aiming at different characteristics of a completely observable power grid and a partially observable power grid, technical means such as a fast reverse or forward load flow calculation method, a particle swarm algorithm and a least square method are combined, and the limitation that a traditional method is difficult to apply to the partially observable power grid is solved.

Description

Technical Field

[0001] The present invention relates to the technical field of power systems, and in particular to a fault location method for a distribution network with distributed generation based on load impedance. Background Art

[0002] With the wide application of distributed generation (DG) in the distribution network (PDN), the problems of fault complexity and vulnerability in the distribution network have become increasingly prominent. Accurately and quickly locating the fault location has become the key to ensuring the stability and security of the power system.

[0003] Traditional fault location methods, such as differential equation methods, impedance-based methods, traveling wave-based methods, and phasor measurement unit (PMU)-based methods, although each has its own advantages, still face many challenges in practical applications. The impedance-based method locates faults by calculating the equivalent impedance of current and voltage, but its main disadvantage is that it may generate multiple fault points, resulting in inaccurate location. The state estimation method based on PMU relies on sufficient measurement data to ensure the observability of the power grid and cannot accurately locate faults in a partially observable power grid. In addition, existing methods are difficult to be applicable to both fully observable power grids and partially observable power grids in the presence of DG, which limits their practical applications.

[0004] Therefore, there is an urgent need for a fault location method to locate faults in these two types of distribution networks, namely fully observable and partially observable distribution networks in distributed generation. Summary of the Invention

[0005] In view of at least one of the above technical problems, the present invention provides a fault location method for a distribution network with distributed generation based on load impedance, and uses the improvement of the method to locate faults in these two types of distribution networks, namely fully observable and partially observable distribution networks in distributed generation.

[0006] According to the first aspect of the present invention, there is provided a fault location method for a distribution network based on load impedance, characterized by comprising the following steps:

[0007] Classify faults in the distribution network with distributed generation;

[0008] In a fully observable distribution network, locate faults according to the current and voltage data recorded by the micro phasor measurement unit;

[0009] In a partially observable distribution network, use the fast backward or forward power flow calculation method to obtain the initial load impedance of each node. Based on the initial load impedance of each node, use the particle swarm optimization algorithm to optimize the complex power of each node to obtain the load value of each node, and then participate in the calculation of the injection current of the selected section nodes and the voltages of the upstream and downstream nodes;

[0010] In a partially observable distribution network, when the network cannot be fully observed by micro - phasor measurement units, there are multiple fault locations during positioning. For the fault locations, the least - squares method is used to calculate the fault distance of each section.

[0011] Among them, the least - squares method outputs multiple results, and the fault distance that is less than the selected section length and is positive is selected as the correct result.

[0012] In some embodiments of the present invention, the faults are classified into ground faults and phase - to - phase faults according to whether the fault is grounded. The ground - fault equation is used to determine the distance of different ground faults, and the ground - fault equation is as follows:

[0013]

[0014] The phase - to - phase fault equation is used to determine the distance of phase - to - phase faults, and the phase - to - phase fault equation is as follows:

[0015]

[0016] In the formula, k0 to k5 are the general terms of k 0m to k 5m 、k 0a to k 5a 、k 0b to k 5b and are defined vectors; k 0m to k 5m correspond to ground - fault parameters, m is an element of the set p, k 0a to k 5a and k 0b to k 5b correspond to a - b phase - to - phase fault parameters; a, b, c are the three phases of the distribution - network line, p is the set of a, b, c three phases; x is the fault distance; is the current at the fault point, corresponds to the ground fault, corresponds to the phase - to - phase fault; V S is the input voltage of each section calculated without micro - phasor measurement units; I S is the input current of each section calculated without micro - phasor measurement units, which is recorded by the section with micro - phasor measurement units at the upstream node.

[0017] In some embodiments of the present invention, the following equations are used to update the voltage and current at the fault point:

[0018] I F =I D -I u

[0019] V F =k0V S+k1xI S +k2x 2 V S +k1x 3 I S +k4x 4 V S +k5x 5 I S

[0020]

[0021] I D =k I0 .I s +k I1 x.V s +k I2 x 2 .I s +k I3 x 3 .V s +k I4 x 4 .I s +k I5 x 5 .V s

[0022] In the formula, I F is the fault point current, and I D is the input current at the fault point; V F is the fault point voltage; I u is the load current; l is the total length of the section; Y' is the branch admittance; Z L -1 is the reciprocal of the load impedance; Z' is the equivalent impedance of the branch; k I0 to k I5 are defined vectors calculated from the distributed line parameters.

[0023] In some embodiments of the present invention, when a fault occurs among the substation, the load, and the distributed power source, the following steps are adopted:

[0024] S110: Calculate the upstream and downstream voltages of the fault section by using the distribution line parameters and the recorded information of the relevant micro phasor measurement units;

[0025] S120: Calculate the voltage at the fault point from both sides of the substation and the distributed power source according to the distance of the fault point;

[0026] S130: Make the ground fault equation and the phase-to-phase fault equation equal, recalculate, and obtain a new fault distance.

[0027] In some embodiments of the present invention, when estimating the load impedance of each node in a distribution network, if it is a fully observable distribution network, the following steps are adopted:

[0028] S210: Assume that the load impedance of each node remains unchanged within one cycle before and after a fault occurs;

[0029] S220: Use the pre-fault current and voltage data recorded by the micro phasor measurement unit to calculate the load impedance of each node;

[0030] S230: Determine the equivalent load impedance at the end of the fault section by using the load impedance of each node.

[0031] In some embodiments of the present invention, when estimating the load impedance of each node in a distribution network, if it is a partially observable distribution network, first estimate the initial load impedance, and then calculate the complex power value of each node;

[0032] When estimating the initial load impedance, the complex power of each node is recorded by a data recorder with a fixed time constant T. If a fault occurs between time t and t + T, and the fault occurrence time is α, use the recorded data to estimate the complex power value of each node at the fault occurrence time α. For the nearest node equipped with a micro phasor measurement unit, its complex power value at the fault occurrence time α can be determined by the following formula:

[0033]

[0034] In the formula, S j is the complex power of node j measured by the micro phasor measurement unit; S m is the complex power measured by the micro phasor measurement unit at the nearest node equipped with a micro phasor measurement unit; S i is the complex power recorded by the data recorder; N is the set of nodes with data recorders; M is the set of nodes with micro phasor measurement units; γ is the loss of complex power on the line;

[0035] Use the complex power data recorded by the data recorder to calculate the complex power of each node in set N at time t + T by the following formula:

[0036]

[0037] In the formula, m k is the weight factor; x′ i(t+T) is the complex power of the i-th node at time t + T; is the complex power of the i-th node recorded for the k-th time at time t + T; K is the total number of complex power data. In some embodiments of the present invention, the following formula is used to simplify the estimation:

[0038]

[0039] Wherein, NN is the number of all data in set K, which is used to estimate the amount of information; is the complex power recorded for the kk-th time at the i-th node at time t + T. k ∈ Lyear refers to the data of several days before and after the fault last year, and kk ∈ Nyear refers to the data of several days before the fault this year;

[0040] The data recorder records the accurate value of the complex power at time t and determines the estimated value of the complex power at time t + T. Therefore, the initial complex power value of each node is estimated using the following formula at time α:

[0041]

[0042] Wherein, x′ i(α) is the estimated value of the initial complex power of the i-th node in set N at time α; x it is the recorded value of the complex power of the i-th node at time t; x′ i(t+T) is the estimated value of the complex power of the i-th node at time t + T;

[0043] The following formula is the square difference between the current of the node with the nearest micro - phasor measurement unit and the current recorded by the micro - phasor measurement unit and the current calculated by the fast backward or forward power flow calculation method:

[0044]

[0045] Wherein, F(x) is ……, is the m - phase current before the fault recorded by the nearest micro - phasor detection unit; is the m - phase current of each node calculated using the fast backward or forward power flow calculation method and complex power calculation.

[0046] In some embodiments of the present invention, when calculating the complex power value of each node, the following steps are included:

[0047] S310: Take the initial estimated value of the complex power as the initial optimal solution, generate multiple particles, set a random solution for each particle, and set the initial velocity to zero;

[0048] S320: Calculate the fitness of each particle and determine the best position and global optimal solution among all solutions;

[0049] S330: Update the velocity of each particle;

[0050] S340: Determine the updated position of each particle;

[0051] S350: If the above conditions are met, the calculation stops; otherwise, loop in S330. In some embodiments of the present invention, if the power grid is fully observable by the micro - phasor measurement units, fault location is performed by recording the currents and voltages of all micro - phasor measurement units, and the transient current and voltage data of all nodes at the time of the fault are used to determine the fault section. The algorithm calculates the fault distance based on the assumption that the fault occurs in each section and outputs the result.

[0052] In some embodiments of the present invention, if the power grid is partially observable, the least - squares method provides multiple fault locations for the distribution network with branches. The fault information is recorded by the closest micro - phasor measurement unit. All possible fault points are simulated, and the actual fault voltage is compared with the simulated voltage. The matching value is defined as the least - squares error between the recorded voltage and the simulated voltage, and the formula is as follows:

[0053]

[0054] In the formula, takes the simulated voltage of the k - th possible fault section as a sample; takes the actual voltage as a sample; N is the number of samples and the recorded data; G K is the matching value of the k - th possible fault section; m is the set of phases, G K is the matching value, G K The smaller the value, the closer the simulated voltage is to the recorded voltage; the true fault section is the subscript corresponding to the lowest value of G K ;

[0055] The formula for the true fault section is as follows:

[0056] min{G K}K = N f

[0057] In the formula, N f is the number of fault points;

[0058] When a fault occurs in the power grid, all power sources inject current into the fault point.

[0059] According to the second aspect of the present invention, a distribution network fault location system based on load impedance is further provided, including:

[0060] A classification unit for classifying faults in the distributed - power - source distribution network;

[0061] A micro - phasor measurement unit for fault location based on the current and voltage data recorded by the micro - phasor measurement unit;

[0062] A calculation unit is used to obtain the initial load impedance of each node in a partially observable distribution network by using a fast backward or forward power flow calculation method. Based on the initial load impedance of each node, a particle swarm algorithm is used to optimize the complex power of each node to obtain the load value of each node, and then participate in the calculation of the injection current of the selected section nodes and the voltages of the upstream and downstream nodes.

[0063] A fault location unit, in a partially observable distribution network, when the network cannot be fully observed by micro - phasor measurement units, there are multiple fault locations during location. For the fault locations, the least - squares method is used to calculate the fault distance of each section, and the fault distance that is less than the length of the selected section and is positive is selected as the correct result from multiple results. The beneficial effects of the present invention are as follows: First, in view of the different characteristics of fully observable and partially observable power grids, the present invention designs corresponding fault location calculation processes, and combines technical means such as fast backward or forward power flow calculation methods, particle swarm algorithms, and least - squares methods to solve the limitations of traditional methods that are difficult to apply in partially observable power grids. Compared with the prior art, through flexible fault classification and targeted calculations, this method can be applied to both fully observable and partially observable power grids, and solves the limitations of traditional methods that are difficult to apply in partially observable power grids. Description of the Drawings

[0064] In order to more clearly illustrate the technical solutions in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments recorded in the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0065] Figure 1 It is a schematic diagram of the step - by - step process of the distribution network fault location method based on load impedance in the embodiment of the present invention;

[0066] Figure 2 It is a schematic diagram of the step - by - step process of making corresponding fault location calculations when a fault occurs among a substation, a load, and a distributed power source in the embodiment of the present invention;

[0067] Figure 3 It is a schematic diagram of the step - by - step process of estimating the load impedance of each node in a fully observable distribution network in the embodiment of the present invention;

[0068] Figure 4 It is a schematic diagram of the step - by - step process of calculating the complex power value of each node in the embodiment of the present invention. Detailed Embodiments

[0069] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings in the embodiments of the present invention. Obviously, the described embodiments are only a part of the embodiments of the present invention, rather than all the embodiments.

[0070] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those of ordinary skill in the technical field to which this invention belongs. The terms used in the description of the present invention are for the purpose of describing specific embodiments only and are not intended to limit the present invention. The term "and / or" as used herein includes any and all combinations of one or more of the associated listed items.

[0071] As Figures 1 to 4 shown, the method for fault location in a distribution network based on load impedance includes the following steps:

[0072] Classify the faults in the distributed power generation distribution network;

[0073] In a fully observable distribution network, perform fault location according to the current and voltage data recorded by the micro - phasor measurement unit;

[0074] In a partially observable distribution network, use the fast backward or forward power flow calculation method to obtain the initial load impedance of each node. Based on the initial load impedance of each node, use the particle swarm optimization algorithm to optimize the complex power of each node to obtain the load value of each node, and then participate in the calculation of the injection current of the selected section nodes and the voltages of the upstream and downstream nodes;

[0075] In a partially observable distribution network, when the network cannot be fully observed by the micro - phasor measurement unit, there are multiple fault locations during location. For the fault locations, use the least - squares method to calculate the fault distance of each section;

[0076] Among them, the least - squares method outputs multiple results, and select the fault distance that is less than the length of the selected section and is positive as the correct result.

[0077] As Figure 1As shown, first, this step is used to identify different types of faults in the distribution network. The distribution network may face different types of faults, such as short - circuit faults, grounding faults, etc. Each type of fault may have different manifestations and impacts. By classifying the fault situations into multiple categories and designing corresponding fault - location calculation methods for each situation, it can help the system more accurately locate the position and nature of the fault. Subsequently, the micro - phasor measurement unit, namely μPMU. In the case where the distribution network is fully observable, that is, accurate current and voltage data can be obtained for each key node in the network through the micro - phasor measurement unit, these data can be directly used for fault location to infer the location where the fault occurs. Optionally, FBSLF, that is, the fast backward / forward power - flow calculation method, is used to quickly calculate the power - grid state, and PSO, that is, the particle - swarm optimization algorithm, improves the accuracy of impedance estimation through intelligent optimization, especially suitable for the case of incomplete data in a partially observable power grid. In a partially observable distribution network, due to incomplete monitoring data, first, a fast backward or forward power - flow calculation method needs to be adopted to estimate the initial load impedance of each node based on the known measurement data. Subsequently, the particle - swarm optimization algorithm is used to optimize the complex power of each node, thereby improving the accuracy of fault location. For the determined potential fault locations, the least - squares method is used to calculate the fault distance of each section. Since the least - squares method may output multiple solutions, screening is required, and finally, the solution that is less than the selected section length and is positive is selected as the correct fault distance. This screening criterion ensures the rationality of the fault location and avoids physically infeasible solutions. Here, "deducing each section" means inferring the specific position range of the fault area through the least - squares method based on data such as current, voltage, and load that are already known.

[0078] In the above - mentioned embodiment, the present invention first designs corresponding fault - location calculation processes according to the different characteristics of fully observable and partially observable power grids, and combines technical means such as the fast backward or forward power - flow calculation method, the particle - swarm algorithm, and the least - squares method to solve the limitation that traditional methods are difficult to apply in a partially observable power grid. Compared with the prior art, this method can be applied to both fully observable and partially observable power grids through flexible fault classification and targeted calculations, and solves the limitation that traditional methods are difficult to apply in a partially observable power grid.

[0079] In the embodiment of the present invention, faults are classified into grounding faults and phase - to - phase faults according to whether the fault is grounded. The grounding - fault equation is used to determine the distance of different grounding faults, and the grounding - fault equation is shown as follows:

[0080]

[0081] The phase - to - phase fault equation is used to determine the distance of phase - to - phase faults, and the phase - to - phase fault equation is shown as follows:

[0082]

[0083] Wherein, k0 to k5 are collectively referred to as k 0m ~k 5m , k 0a ~k 5a , k 0b ~k 5b , which are defined vectors; k 0m to k 5m correspond to the ground fault parameters, m is an element of the set p, and k 0a to k 5a and k 0b to k 5b correspond to the phase a - phase b fault parameters; a, b, c are the three phases of the distribution network line, p is the set of the three phases a, b, c; x is the fault distance; is the current at the fault point, corresponds to the ground fault, corresponds to the phase - to - phase fault; V S is the input voltage of each section calculated without a micro - phasor measurement unit; I S is the input current of each section calculated without a micro - phasor measurement unit, which is recorded by the section with a micro - phasor measurement unit at the upstream node.

[0084] A ground fault is a fault that occurs when a certain phase in the distribution network contacts or shorts to the ground. The ground fault equation uses a set of mathematical relationships composed of parameters such as current and voltage to calculate the specific location where the fault occurs. These equations may involve parameters such as the current at the fault point and the line impedance of the distribution network. A phase - to - phase fault is a fault that occurs when a short circuit occurs between two or more phases in the distribution network. For a phase - to - phase fault, another set of equations specifically for phase - to - phase short circuits is used to estimate the fault distance. In the distribution network, a micro - phasor measurement unit can also be set up to improve the accuracy of fault location. For example, an electrical quantity measurement device can be added to accurately capture the changes in current and voltage. In this embodiment, by defining a series of current, voltage, and impedance parameters and combining the data of the fault point current and the phasor measurement unit, the fault distance is calculated, and fast and accurate fault location can be achieved.

[0085] In the embodiment of the present invention, the above - mentioned equations are applicable to faults that occur only between the substation and the load. In this case, the solution may not converge to a reasonable range (a positive value and less than the fault section). Therefore, the following equations are used to update the voltage and current at the fault point:

[0086] I F =I D -I u

[0087] V F =k0VS +k1xI S +k2x 2 V S +k1x 3 I S +k4x 4 V S +k5x 5 I S

[0088]

[0089] I D =k I0 .I s +k I1 x.V s +k I2 x 2 .I s +k I3 x 3 .V s +k I4 x 4 .I s +k I5 x 5 .V s

[0090] In the formula, I F is the fault point current, and I D is the input current at the fault point; V F is the fault point voltage; I u is the load current; l is the total length of the section; Y' is the branch admittance; is the reciprocal of the load impedance; Z' is the branch equivalent impedance; k I0 to k I5 are defined vectors, calculated from distributed line parameters.

[0091] According to the given equations, the voltage and current at the fault point will be calculated through a set of known electrical parameters. Through a reasonable mathematical model, combined with the line parameters and load conditions of the distribution network, etc., the voltage and current values at the fault point are updated in real time. By introducing various factors such as load current, branch admittance, and load impedance, and combining with distributed line parameters for calculation, the accuracy of fault location can be effectively improved. Relying on an accurate network model and real-time power measurement data, it is ensured that the fault location system can quickly respond in a complex distribution network environment.

[0092] In the embodiment of the present invention, as Figure 2 shown, when a fault occurs between the substation, the load, and the distributed power source, the following steps are taken:

[0093] S110: Calculate the voltages upstream and downstream of the fault section using the distribution line parameters and the recorded information of relevant micro - phasor measurement units;

[0094] S120: Calculate the voltage at the fault point from both sides of the substation and distributed power sources according to the distance of the fault point;

[0095] S130: Make the ground - fault equation and the phase - to - phase fault equation equal, recalculate and obtain a new fault distance. That is: First, assume that the fault occurs in each selected section, then calculate the fault location for the selected section and determine the fault distance. If it is a non - fault section, the calculation answer is not feasible. At this time, a certain section needs to be re - selected as the fault section for calculation to obtain the fault distance (new fault distance) until the obtained answer is feasible, which indicates that this section is the fault section. In step S110, the distribution line parameters can include electrical characteristics such as the impedance and admittance of the line. Through these parameters, the distribution of current, voltage, etc. can be inferred. In step S120, the voltage at the fault point will change due to factors such as the fault type, line characteristics, and load conditions. By calculating the distance of the fault point, the voltage at the fault point can be further calculated. Then, by comparing the ground - fault equation and the phase - to - phase fault equation and making their solutions equal, the voltage at the fault point can be determined. In step S130, the calculation of the fault distance involves the comprehensive analysis of information such as the establishment of the network model, line impedance, current, and voltage distribution. Through the known fault voltage and current, as well as the relevant parameters in the network, the distance between the fault point and the substation or distributed power source can be finally obtained. In addition, the equivalent load impedance at the end of each section needs to be calculated. The key to the above steps is to locate the fault point by comprehensively using the electrical parameters of the distribution network, real - time measurement data, and mathematical models, which can effectively improve the operation stability and fault handling efficiency of the distribution network.

[0096] In the embodiment of the present invention, as Figure 3 shown, in estimating the load impedance of each node in the distribution network, if it is a fully observable distribution network, the following steps are adopted:

[0097] S210: Assume that the load impedance of each node remains unchanged within one cycle before and after the fault occurs;

[0098] S220: Use the pre - fault current and voltage data recorded by the micro - phasor measurement unit to calculate the load impedance of each node;

[0099] S230: Determine the equivalent load impedance at the end of the fault section using the load impedance of each node.

[0100] In an actual distribution network, the load impedance is usually affected by factors such as load changes and voltage fluctuations. However, in many cases, especially in a short period of time, the load impedance may not change significantly. Therefore, assuming that the load impedance remains unchanged within one cycle, this is a simplified assumption that can reduce the computational complexity and provide stability for subsequent calculations. Here, the concept of one cycle refers to a complete electrical cycle, such as a cycle of 50 Hz or 60 Hz. This assumption facilitates further calculations and analyses, especially when a fault occurs. Through the above method, the load impedance of each node can be quickly and accurately estimated when a fault occurs, and finally the equivalent load impedance of the fault section can be deduced, providing key data for subsequent fault location and power system optimization.

[0101] In an embodiment of the present invention, when estimating the load impedance of each node in a distribution network, if it is a partially observable distribution network, the initial load impedance is first estimated, and then the complex power value of each node is calculated;

[0102] When estimating the initial load impedance, the complex power of each node is recorded by a data recorder with a fixed time constant T. If a fault occurs between time t and t + T, and the fault occurrence time is α, the complex power value of each node at the fault occurrence time α is estimated using the recorded data. For the nearest node equipped with a micro - phasor measurement unit, its complex power value at the fault occurrence time α can be determined by the following formula:

[0103]

[0104] where, S j is the complex power of node j measured by the micro - phasor measurement unit; S m is the complex power measured by the micro - phasor measurement unit at the nearest node equipped with a micro - phasor measurement unit; S i is the complex power recorded by the data recorder; N is the set of nodes with data recorders; M is the set of nodes equipped with micro - phasor measurement units; γ is the loss of complex power on the line;

[0105] Using the complex power data recorded by the data recorder, the complex power of each node in set N at time t + T is calculated by the following formula:

[0106]

[0107] where, m k is the weight factor; x′ i(t+T) is the complex power of the i - th node at time t + T; is the complex power of the k-th record of the i-th node at time t+T; K is the total number of complex power data. In a distribution network, estimating the load impedance is an important step in power system analysis, especially during a fault. For a partially observable distribution network, a reasonable estimation method is needed to deduce the load impedance of each node. When estimating the initial load impedance, historical record data and the complex power values at the fault time are used; the complex power data provided by the micro phasor measurement unit and the data recorder are used to estimate the complex power at the fault time by weighted summation; finally, the historical data is adjusted using the weight factor to further deduce the complex power value of each node, providing a basis for subsequent load impedance calculation. In this embodiment, this method can effectively estimate the load impedance of each node in a partially observable distribution network. Especially during a fault, it can quickly estimate the load change through measurement data and provide key data for fault location, power flow analysis, and distribution network optimization.

[0108] In an embodiment of the present invention, the weighting factor m k is used to adjust the weight of each data (fault time, fault date, and load growth). Since the value of m k is difficult to obtain, the following formula is used to simplify the estimation:

[0109]

[0110] where NN is the number of all data in set K, which is the number of data used for estimation; is the complex power of the kk-th record of the i-th node at time t+T, k∈Lyear refers to the data of several days before and after the fault last year, and kk∈Nyear refers to the data of several days before the fault this year;

[0111] The data selected by this method can exceed two years without limitation. However, considering factors such as load growth may affect the estimation accuracy. Compared with using all data, using the data of this year and last year for preliminary estimation is more accurate. Therefore, for simplicity and to avoid errors, it is recommended to only select two years of data, including the data of several days before the fault this year and the data of several days before and after the fault last year.

[0112] The data recorder records the accurate value of the complex power at time t and determines the estimated value of the complex power at time t+T. Therefore, at time α, the following formula is used to estimate the initial complex power value of each node:

[0113]

[0114] where x′ i(α)为 is the estimated value of the initial complex power of the i-th node in set N at time α; x it is the recorded value of the complex power of the i-th node at time t; x′ i(t+T) is the estimated value of the complex power of the i-th node at time t+T;

[0115] The following formula is the square difference between the current of the nearest node equipped with a micro - phasor measurement unit and the upstream current recorded by the micro - phasor measurement unit and the current calculated by the fast backward or forward power flow calculation method:

[0116]

[0117] In the formula, F(x) is ……, is the m - phase current before the fault recorded by the nearest micro - phasor detection unit; is the m - phase current of each node calculated using the fast backward or forward power flow calculation method and complex power calculation.

[0118] The FBSLF algorithm calculates the input current (I Calculated ) closest to the micro - phasor measurement unit through an iterative method, so it is impossible to find a mathematical expression for calculating the input current. The FBSLF algorithm used in the present invention achieves high - precision complex power estimation by minimizing F(x) and obtains the optimal solution of the above formula under this constraint. It includes several constraint conditions, such as the voltage limit of each node:

[0119] 0.95 < V i < 1.05

[0120] In the formula, V i is the voltage at node i.

[0121] The input current limit of each section:

[0122] I i < I in

[0123] In the formula, I i is the calculated input current of the i - th section; I in is the rated input current of the i - th section.

[0124] The boundary of the complex power value of each section:

[0125]

[0126] In the formula, is the complex power recorded by the nearest micro - phasor measurement unit; is the estimated complex power of the i - th node; K is the set of recorded complex powers; S ik is the k th - th recorded complex power of the i th - th node.

[0127] The voltage boundary of each node is determined according to the design, historical power grid information, and the maximum and minimum load values under steady state. Incorrect voltage boundaries will cause errors in the final load impedance estimation, but have no impact on the initial load impedance estimation. Sα is a set of load impedances of known micro - phasor measurement unit nodes and unknown data recorder nodes, with the unknown parameters being the optimization variables.

[0128] In a partially observable distribution network, due to some nodes lacking sufficient real - time measurement data, the estimation of complex power becomes particularly important. Through historical data, complex power records, and algorithm optimization, the complex power of each node can be relatively accurately deduced, thus helping with subsequent load impedance estimation and fault location. FBSLF (Fault Location Method Based on Load Flow) is an algorithm for fault location in power systems, which calculates the location and impact of faults based on load flow analysis. Here, FBSLF is used to calculate the current value of each node and compare it with the actually measured current value. The goal of this comparison is to improve the accuracy of fault location and the reliability of complex power estimation by minimizing the squared difference. Micro - phasor measurement units are devices installed in the distribution network that can accurately measure the phasors of voltage and current, providing more detailed power grid status data. Due to their different installation positions on nodes, the data they provide can help optimize complex power estimation. In this embodiment, a simplified method for estimating the complex power of distribution network nodes using historical complex power data, micro - phasor measurement unit records, and the FBSLF algorithm optimization method is proposed. By taking the weighted average of complex power data and combining the squared difference of currents calculated by the micro - phasor measurement unit and the FBSLF algorithm, this method can improve the estimation accuracy, especially in a partially observable distribution network environment, providing an effective tool for subsequent load impedance estimation and fault location.

[0129] Based on the above - mentioned embodiment, when calculating the complex power value of each node, as Figure 4 shown, it includes the following steps:

[0130] S310: Take the initial estimate of the complex power as the initial optimal solution, generate multiple particles, and set a random solution for each particle, with the initial velocity set to zero;

[0131] S320: Calculate the fitness of each particle and determine the best position and global optimal solution among all solutions;

[0132] S330: Update the velocity of each particle;

[0133] S340: Determine the updated position of each particle;

[0134] S350: If the above - mentioned conditions are met, stop the calculation; otherwise, enter S330 for cycling.

[0135] In step S310, at the initial stage of particle swarm optimization, it is first necessary to generate an initial solution for each particle, representing a possible complex power estimation value. Each particle among the generated multiple particles represents a solution, and these particles are initialized with randomly generated positions and velocities. Setting the initial velocity to zero means that the particles will not suddenly deviate during the optimization process, and the initial search area only depends on the initial position. Among them, the following formula is used to calculate the fitness of each particle:

[0136] FF = 1 / F(x)

[0137] In the formula, FF is the fitness of each particle; F(x) is the squared difference between the upstream current recorded by the above-mentioned nearest micro - phasor measurement unit and the current calculated by the FBSLF algorithm.

[0138] In step S320, the fitness, that is, the quality of each particle, is measured by the fitness function. The fitness function is a criterion in the particle search process, which measures the quality of each particle solution. In this example, the fitness function can be to calculate the error between the estimated complex power of the particle and the actual complex power. The smaller the error, the better the fitness. By calculating the fitness of each particle, the historically optimal position of each particle itself, that is, the solution with the minimum error, can be found.

[0139] In step S330, the velocity of the particle determines the search direction and step size in the solution space. By updating the velocity of each particle, the particle can adjust its search direction in the solution space, thus continuously approaching the optimal solution.

[0140] In step S340, the velocity of each particle is updated using the following formula:

[0141] v i [t + 1] = wv i [t] + c1r1(x i.best [t] - x i [t]) + c2r2(x gbest [t] - x i [t])

[0142] The inertia weight coefficient is as follows:

[0143] w i+1 = w i *w damp ,w damp = 0.99

[0144] x i [t + 1] = x i [t] + v i [t + 1]

[0145] The nominal values of w, c1, and c2 are as follows:

[0146] w∈[0.4,0.9],0≤c1,c2≤2

[0147] In the formula, w i (w i+1 ) is the inertia weight of the i(i+1)th particle; wd amp is the damping factor; S ik For the i th The kth node th Second recorded complex power; x i [t+1] is the position of the i-th particle at the t+1th iteration. The position of the particle represents its current solution in the solution space. Updating the position of the particle is actually updating the complex power estimate. The updated position of the particle is determined by its current speed. By continuously updating the speed and position of the particle, the particle will gradually approach the optimal solution.

[0148] In step S350, it is determined whether to stop the iteration according to the set stop condition. If the preset optimization accuracy is met or the maximum number of iterations is reached, the calculation is stopped, and the final global optimal solution is the complex power estimation value of the node in the distribution network. If the stop condition is not met, the algorithm will return to step S330 and continue to update the speed and position of the particle until the stop condition is met.

[0149] Through the above process, the particle swarm algorithm can effectively estimate the complex power value of each node in a partially observable distribution network, providing more accurate data for load impedance estimation, fault location and grid optimization.

[0150] In an embodiment of the present invention, if the power grid is fully observed by a micro-phasor measurement unit, the fault location is performed by recording the current and voltage of all micro-phasor measurement units, and the transient current and voltage data of all nodes when the fault occurs are used to determine the fault section. The algorithm calculates the fault distance based on the assumption that the fault occurs in each section and outputs the result. It should be noted here that there are multiple output results, only one of which meets the requirements, that is, it is less than the selected section length and is a positive value. In a fully observable power grid, each node is equipped with a micro-phasor measurement unit that can record the phasor data of current and voltage in real time. With this arrangement, the operating status of the entire power grid can be fully monitored when a fault occurs, without the need for data compensation or interpolation calculations. When a fault occurs in the power grid, the fault current and voltage signals will change, especially at the initial time of the fault. Using these data collected by the micro-phasor measurement unit, the location of the fault can be accurately analyzed.

[0151] In an embodiment of the present invention, if the power grid is partially observable, the least squares method provides multiple fault locations for a distribution network with branches. The fault information is recorded by the closest micro - phasor measurement unit. All possible fault points are simulated, and the actual fault voltage is compared with the simulated voltage. The matching value is defined as the least - squares error between the recorded voltage and the simulated voltage, and the formula is as follows:

[0152]

[0153] In the formula, is the sample of the simulated voltage of the k - th possible fault section; is the sample of the actual voltage; N is the number of samples and recorded data; G K is the matching value of the k - th possible fault section; m is the set of phases, G K is the matching value, G K The smaller the value, the closer the simulated voltage is to the recorded voltage; the true fault section is the subscript corresponding to the minimum value of G K

[0154] A partially observable power grid means that through the current and voltage data of multiple nodes, we can obtain the real - time information of part of the power grid, but the state of the entire power grid cannot be fully observed. To infer the fault point from these limited data, it is usually necessary to simulate different possible fault points and compare them with the actual observed data. The matching value is measured by the least - squares error, that is, by calculating the difference between the voltage waveforms of each possible fault section and the actual recorded voltage to evaluate their similarity. The smaller the matching value, the closer the simulated voltage is to the actual voltage, which means that the assumption of the fault location is more in line with the actual situation. By calculating the matching values of all possible fault locations, the fault section with the smallest matching value will ultimately be selected as the actual fault point.

[0155] The formula for the true fault section is as follows:

[0156] min{G K} K = N f

[0157] In the formula, N f is the number of fault points;

[0158] When a fault occurs in the power grid, all power sources inject current into the fault point.

[0159] ​When a power grid fails, all power sources inject current into the fault point. The magnitude and direction of the fault current depend on the fault type, fault location, line impedance, and the operating state of the power grid. The changes in current injection will be reflected in the current and voltage measurement values of each micro - phasor measurement unit. By analyzing the current and voltage, the location of the fault point can be further confirmed. The above - mentioned method achieves relatively accurate fault location through limited observation data, especially suitable for fault diagnosis of partially observable power grids.

[0160] In an embodiment of the present invention, there is also provided a distribution network fault location system based on load impedance, including:

[0161] A classification unit for classifying faults in a distributed - power - source distribution network;

[0162] A micro - phasor measurement unit for fault location based on the current and voltage data recorded by the micro - phasor measurement unit;

[0163] A calculation unit for, in a partially observable distribution network, using the fast backward or forward power - flow calculation method to obtain the initial load impedance of each node, and based on the initial load impedance of each node, using the particle - swarm algorithm to optimize the complex power of each node to obtain the load value of each node, and then participating in the calculation of the injection current of the selected - segment nodes and the voltages of the upstream and downstream nodes;

[0164] A fault location unit for, in a partially observable distribution network, when the network cannot be fully observed by the micro - phasor measurement unit, there are multiple fault locations during location. For the fault location, the least - squares method is used to calculate the fault distance of each section, and the fault distance that is less than the length of the selected section and is positive is selected as the correct result from multiple results.

[0165] Those skilled in the art should understand that the present invention is not limited by the above - mentioned embodiments. What is described in the above - mentioned embodiments and the specification only illustrates the principle of the present invention. Without departing from the spirit and scope of the present invention, the present invention will have various changes and improvements, and these changes and improvements all fall within the scope of the present invention claimed. The scope of protection claimed by the present invention is defined by the appended claims and their equivalents.

Claims

1. A method for fault location in a distribution network based on load impedance, characterized in that, It includes the following steps: Classify the faults of the distributed power distribution network; In a fully observable distribution network, locate the faults based on the current and voltage data recorded by the micro - phasor measurement unit; In a partially observable distribution network, use the fast backward or forward power flow calculation method to obtain the initial load impedance of each node. Based on the initial load impedance of each node, use the particle swarm algorithm to optimize the complex power of each node to obtain the load value of each node, and then participate in the calculation of the injected current of the selected section nodes and the voltages of the upstream and downstream nodes; In a partially observable distribution network, when the network cannot be fully observed by the micro - phasor measurement unit, there are multiple fault locations during location. For the fault locations, use the least - squares method to calculate the fault distance of each section; Among them, the least - squares method outputs multiple results, and select the fault distance that is less than the length of the selected section and is positive as the correct result.

2. The method for fault location in a distribution network based on load impedance according to claim 1, characterized in that Classify the faults into ground faults and phase - to - phase faults according to whether the fault is grounded. The ground - fault equation is used to determine the distance of different ground faults, and the ground - fault equation is as follows: The phase - to - phase fault equation is used to determine the distance of phase - to - phase faults, and the phase - to - phase fault equation is as follows: where k0 to k5 are collectively referred to as k 0m to k 5m , k 0a to k 5a , k 0b to k 5b , which are defined vectors; k 0m to k 5m correspond to the ground fault parameters, m is an element of the set p, and k 0a to k 5a and k 0b to k 5b correspond to the phase a - b fault parameters; a, b, c are the three phases of the distribution network line, and p is the set of a, b, c three phases; x is the fault distance; is the current at the fault point, corresponds to the ground fault, corresponds to the phase - to - phase fault; V S is the input voltage calculated for each section without the micro - phasor measurement unit; I S is the input current calculated for each section without the micro - phasor measurement unit, which is recorded by the section with the micro - phasor measurement unit at the upstream node.

3. The method for fault location of a distribution network based on load impedance according to claim 2, wherein Use the following equation to update the voltage and current at the fault point: I F = I D - I u V F = k0V S + k1xI S + k2x 2 V S + k1x 3 I S + k4x 4 V S + k5x 5 I S I D = k I0 .I s + k I1 x.V s + k I2 x 2 .I s + k I3 x 3 .V s + k I4 x 4 .I s + k I5 x 5 .V s Wherein, I F is the fault point current, and I D is the input current at the fault point; V F is the fault point voltage; I u is the load current; l is the total length of the section; Y' is the branch admittance; is the reciprocal of the load impedance; Z' is the branch equivalent impedance; k I0 to k I5 are defined vectors calculated from distributed line parameters.

4. The method for locating faults in a distribution network based on load impedance according to claim 2, wherein, When a fault occurs among the substation, load, and distributed power source, adopt the following steps: S110: Calculate the upstream and downstream voltages of the fault section by using the distribution line parameters and the recorded information of the relevant micro - phasor measurement unit; S120: Calculate the voltage at the fault point from both sides of the substation and the distributed power source according to the distance of the fault point; S130: Make the ground - fault equation and the phase - to - phase fault equation equal, recalculate and obtain a new fault distance.

5. The method for fault location of a distribution network based on load impedance according to claim 1, characterized in that In estimating the load impedance of each node in the distribution network, if it is in a fully observable distribution network, adopt the following steps: S210: Assume that the load impedance of each node remains unchanged within one cycle before and after the fault occurs; S220: Use the current and voltage data before the fault recorded by the micro - phasor measurement unit to calculate the load impedance of each node; S230: Use the load impedance of each node to determine the equivalent load impedance at the end of the fault section.

6. The method for fault location of a distribution network based on load impedance according to claim 1, characterized in that, In estimating the load impedance of each node in the distribution network, if it is in a partially observable distribution network, first estimate the initial load impedance, and then calculate the complex power value of each node; When estimating the initial load impedance, the complex power of each node is recorded by the data recorder with a fixed time constant T. If a fault occurs between time t and t + T, and the fault occurrence time is α, use the recorded data to estimate the complex power value of each node at the fault occurrence time α. For the nearest node equipped with a micro - phasor measurement unit, its complex power value at the fault occurrence time α can be determined by the following formula: Wherein, S j is the complex power of node j measured by the micro - phasor measurement unit; S m is the complex power measured by the micro - phasor measurement unit at the nearest node equipped with the micro - phasor measurement unit; S i is the complex power recorded by the data recorder; N is the set of nodes with data recorders; M is the set of nodes with micro - phasor measurement units; γ is the loss of complex power on the line; Use the complex power data recorded by the data recorder to calculate the complex power of each node in the set N at time t + T by the following formula: Where, m k is the weighting factor; x’ i(t+T) is the complex power of the i-th node at the moment of t + T; is the complex power of the k-th record of the i-th node at the moment of t + T; K is the total number of complex power data.

7. The method for locating faults in a distribution network based on load impedance according to claim 6, characterized in that, Use the following formula to simplify the estimation: where NN is the number of all data in set K, which is used to estimate the amount of information; is the kk-th recorded complex power of the i-th node at the t+T moment. k∈Lyear refers to the data of several days before and after the fault last year, and kk∈Nyear refers to the data of several days before the fault this year; The data recorder records the accurate value of the complex power at time t and determines the estimated value of the complex power at time t + T. Therefore, use the following formula to estimate the initial complex power value of each node at time α: where, x' i(α) is the initial complex power estimation value of the \(i\)th node in set \(N\) at time \(\alpha\); \(x\) it is the complex power recorded value of the \(i\)th node at time \(t\); \(x'\) i(t+T) is the complex power estimation value of the \(i\)th node at time \(t + T\); The following formula represents the squared difference between the current of the nearest node equipped with a micro - phasor measurement unit and the upstream current recorded by the micro - phasor measurement unit and the current calculated by the fast backward or forward power flow calculation method: where, F(x) is... is the pre-fault m-phase current recorded by the nearest microphasor measurement unit; is the m-phase current of each node calculated using the fast backward or forward power flow calculation method and complex power calculation.

8. The method for fault location in a distribution network based on load impedance according to claim 7, characterized in that, When calculating the complex power value of each node, the following steps are included: S310: Take the initial estimate of the complex power as the initial optimal solution, generate multiple particles, set a random solution for each particle, and set the initial velocity to zero; S320: Calculate the fitness of each particle and determine the best position and the global optimal solution among all solutions; S330: Update the velocity of each particle; S340: Determine the updated position of each particle; S350: If the above conditions are met, stop the calculation; otherwise, enter S330 for cycling.

9. The method for fault location in a distribution network based on load impedance according to claim 1, wherein, If the power grid is fully observed by micro - phasor measurement units, fault location is carried out by recording the currents and voltages of all micro - phasor measurement units, and the transient current and voltage data of all nodes at the time of fault are used to determine the fault section. The algorithm calculates the fault distance based on the assumption that the fault occurs in each section and outputs the result.

10. The method for fault location of a distribution network based on load impedance according to claim 9, characterized in that, If the power grid is partially observable, the least - squares method provides multiple fault locations for the distribution network with branches. The fault information is recorded by the nearest micro - phasor measurement unit. Simulations are carried out for all possible fault points, and the actual fault voltage is compared with the simulated voltage. The matching value is defined as the least - squares error between the recorded voltage and the simulated voltage. The formula is as follows: Wherein, is the sample with the simulated voltage of the k-th possible fault segment; is the sample with the actual voltage; N is the sample quantity and the recorded data; G K is the matching value of the k-th possible fault segment; m is the phase set, G K is the matching value, G K The smaller the value, the closer the simulated voltage is to the recorded voltage; the true fault segment is G K the subscript corresponding to the lowest quantity; The formula for the true fault section is as follows: min{G K}K = N f where N f is the number of fault points; When a fault occurs in the power grid, all power sources inject current into the fault point.

11. A distribution network fault location system based on load impedance, characterized in that, Including: A classification unit for classifying faults in the distributed - power - source distribution network; A micro - phasor measurement unit for fault location based on the current and voltage data recorded by the micro - phasor measurement unit; A calculation unit for obtaining the initial load impedance of each node by using the fast backward or forward power flow calculation method in a partially observable distribution network, optimizing the complex power of each node by using the particle swarm algorithm based on the initial load impedance of each node to obtain the load value of each node, and then participating in the calculation of the injected current of the nodes in the selected section and the voltages of the upstream and downstream nodes; A fault - location unit for calculating the fault distance of each section by using the least - squares method and selecting the fault distance that is less than the length of the selected section and is positive as the correct result when there are multiple fault positions during fault location in a partially observable distribution network where the network cannot be fully observed by micro - phasor measurement units.

Images