Fault location method for distribution network with distributed power based on improved lemming algorithm

Abstract

The application provides a kind of power distribution network fault location method containing distributed power supply based on improved vole algorithm, by the network topology of power distribution network, switch state coding mode is used, switch state coding vector is defined, and switch state function is constructed;The objective function of the fault location model of the power distribution network containing distributed power supply is constructed;Actual switch state value is obtained, and switch expected state information is obtained;The objective function value of individual is calculated;Initial population is generated by using the best point set strategy, the elite pool strategy and the adaptive gaussian disturbance strategy are introduced, and the improved vole algorithm is used to iteratively optimize the candidate fault interval combination;Finally, the fault interval corresponding to the optimal individual is output;The application can efficiently realize fault location, has better ability to jump out of local optimum, and has better multiple fault location effect for containing distortion information, and experimental simulation verification can have faster positioning speed and higher positioning accuracy in different topological power distribution networks.

Description

Technical Field

[0001] This invention relates to a fault location method for distribution networks with distributed power sources based on an improved lemming algorithm, belonging to the field of distribution network fault location technology. Background Technology

[0002] With the large-scale integration of distributed generation (DG) and smart switches into the power grid, the information available to the distribution network is constantly increasing, but at the same time, fault location has become more complex. The integration of DG alters the magnitude and direction of fault currents, rendering traditional unidirectional location methods ineffective. Furthermore, the information collected and uploaded by distribution terminals may be distorted or contain false alarms, leading to discrepancies between the fault state and the actual situation. Therefore, accurately determining the fault range using limited and potentially unreliable fault information has become a critical issue in distribution network fault diagnosis.

[0003] Compared with traditional fault location methods, intelligent optimization algorithms can better adapt to complex scenarios such as multi-branch topologies and information distortion, making them significant for research. Currently, other swarm intelligent optimization algorithms, such as particle swarm optimization (PSO), sparrow optimization, whale optimization, and gray wolf optimization, have been applied to fault location problems. Among them, PSO has a simple structure and fast convergence, but it is prone to population diversity decline and local optima problems in the later stages of iteration. Newer swarm intelligent algorithms such as sparrow and whale optimization have improved search capabilities, but they may still suffer from insufficient local exploitation capabilities and low convergence stability when dealing with discrete fault interval location problems. Despite many improved intelligent optimization algorithms, they still suffer from slow convergence speed and susceptibility to local optima when handling fault location in distribution networks with multiple power sources, and their effectiveness is also weak for multi-fault location with distorted information. Summary of the Invention

[0004] The purpose of this invention is to provide a fault location method for distribution networks with distributed power sources based on an improved lemming algorithm, which solves the problems of slow convergence speed, easy getting trapped in local optima, and the need to improve the speed and accuracy of fault location for multiple faults containing distorted information in the existing technology.

[0005] The technical solution of this invention is:

[0006] A fault location method for distribution networks with distributed power sources based on an improved lemming algorithm includes the following steps:

[0007] S1. Based on the network topology of the distribution network, a switch state coding method is adopted to define the switch state coding vector R and construct the switch state function;

[0008] S2. Objective function for constructing a fault location model for a distribution network containing distributed generation sources;

[0009] S3. Encode the fault status information uploaded by the feeder terminal (FTU) of the distribution network using the switch status encoding method in step S1 to obtain the actual switch status value. The desired switch state information is obtained from the switch state function. ;

[0010] S4. Set the actual switch status value With switch expected state information Substitute the objective function into the algorithm population and calculate the objective function value for each individual.

[0011] S5. The initial population is generated using the optimal point set strategy. The elite pool strategy and the adaptive Gaussian perturbation strategy are introduced. The objective function value of the individual is calculated using step S4. The improved lemming algorithm is used to iteratively optimize the combination of candidate fault intervals to obtain the optimal individual.

[0012] S6. Using binary mapping rules, the continuous positions corresponding to the combination of fault intervals are converted into fault segment states, and finally the fault interval corresponding to the optimal individual is output.

[0013] Further, step S1 specifically involves,

[0014] S11. Suppose a distribution network has N branches, i.e., N feeder segments. Each branch of the distribution network may be in either a normal or fault state. A binary mapping rule is used to represent the fault state of the distribution network, and the fault state vector is defined as S = [s1, s2, ..., s...]. j , ..., s N ], where the operating state s of the j-th branch. j :

[0015] ;

[0016] S12. Based on the multidirectional nature of fault currents after distributed generation is connected to the distribution network, a switch state coding method is adopted. The direction from the main power source to the end of the line is defined as the positive direction, and otherwise as the negative direction. The switch state coding vector R = [R1, R2, ..., R...] is defined. j , ..., R N ], where R1, R2, ..., R j , ..., R N These represent the operating states of branches 1, 2, ..., j, ..., N in the distribution network, respectively. A switch detects a fault state R in the j-th branch. j :

[0017] ;

[0018] S13. The expression for the switching state function applicable to distribution networks containing distributed generation sources is as follows:

[0019]

[0020] in, Indicates the desired state information of the switch; and Let J represent the switch state functions upstream and downstream of switch j, respectively, with the following expressions:

[0021]

[0022]

[0023] in, Representation or logic; and These represent the total number of upstream and downstream feeder sections of switch j, respectively. and These represent the number of power sources upstream and downstream of switch j, respectively; and These indicate whether the upstream and downstream DG of switch j are connected to the grid, respectively. and These represent the state quantities of each segment between switch j and the upstream and downstream power sources, respectively. and These represent the state variables of all feeder sections upstream and downstream of switch j, respectively.

[0024] Furthermore, in step S2, the objective function The expression is as follows:

[0025]

[0026] in, The actual switch status value is given, where M represents the total number of feeder switches. is the penalty coefficient, used to adjust the impact of the fault quantity constraint on the objective function; N is the total number of feeder sections; Indicates the number of faulty sections.

[0027] Further, step S5 specifically involves,

[0028] S51. Set the population size D and the maximum number of iterations T, and initialize the population using the optimal point set strategy;

[0029] S52. Introduce an elite pool strategy: Calculate the fitness value of each individual in the t-th iteration, which is the objective function value, and construct an elite pool based on the fitness values.

[0030]

[0031] in, These are the positions of the top three elite individuals in the population at iteration t, ordered from highest to lowest fitness value, with fitness values ​​of [values ​​to be filled in]. ,satisfy ; To construct elite guiding positions; These are the weight coefficients of elite individuals, and their sum is 1.

[0032] S53. Calculate the energy factor E. If the energy factor E > 1, proceed to the next step S54; otherwise, proceed to the next step S55.

[0033] S54. Determine whether the random number Rand in the interval (0,1) is less than the first set value. If so, update the lemming position according to the long-distance migration behavior and proceed to the next step S56; otherwise, update the lemming position according to the burrowing behavior and proceed to the next step S56.

[0034] S55. Determine whether the random number Rand in the interval (0,1) is less than the second set value. If so, update the lemming position according to the foraging behavior and proceed to the next step S56; otherwise, update the lemming position according to the predator avoidance behavior and proceed to the next step S56.

[0035] S56. Calculate fitness values ​​and update lemming positions. ;

[0036] S57. Introduce an adaptive Gaussian perturbation strategy to obtain the lemming position after Gaussian perturbation. ;

[0037] S58. Compare the lemming positions after Gaussian perturbation. and lemming location The best solution is selected as the optimal solution for the t-th iteration.

[0038] S59. After updating the elite pool according to fitness, if t < T, let t = t + 1 and return to step S53; otherwise, take the optimal solution of the t-th iteration as the optimal individual.

[0039] Furthermore, in step S51, the population is initialized using a best-point-set strategy, specifically as follows:

[0040] Let G D For a unit cube in D-dimensional Euclidean space, the set of optimal points for Among them, Jia Dian Ji deviation Where n represents the population size, It is only with and positive numbers Relevant constants, Indicates the first The dimension refers to the optimal point parameter used for the j-th branch, where j = 1, 2, ..., D. Indicates the first The position of the initial individual; take , It is to satisfy The smallest prime number; spread the points of the optimal point set evenly and map them to the upper and lower bounds of the variable.

[0041] Furthermore, in step S57, the lemming position after Gaussian perturbation is obtained. :

[0042]

[0043] in, For disturbance terms; For a Gaussian perturbation random term that follows a standard normal distribution; This represents the adaptive disturbance factor, used to control the disturbance intensity. , where η min This represents the minimum disturbance value; η max t represents the maximum perturbation value; t represents the current iteration number; and T represents the maximum iteration number.

[0044] The beneficial effects of this invention are:

[0045] I. This fault location method for distribution networks with distributed power sources based on the improved lemming algorithm can efficiently locate faults and has a better ability to escape local optima. It is more effective for locating multiple faults with distorted information. Experimental simulations have verified that it can achieve fast location speed and high location accuracy in distribution networks with different topologies.

[0046] II. This fault location method for distribution networks with distributed power sources, based on an improved lemming algorithm, constructs a switch state function and an objective function for the fault location model of the distribution network with distributed power sources. It initializes the population using a best-point set initialization strategy, improving the initial search quality of the algorithm. An elite pool strategy is introduced, utilizing multiple high-performing individuals to construct a weighted guidance mechanism, which collectively guides population updates, enhancing the algorithm's ability to escape local optima and helping it find the optimal solution more quickly. An adaptive Gaussian perturbation strategy is also introduced, providing a larger perturbation in the early stages of iteration and a smaller perturbation in the later stages, balancing global exploration in the early stages with refined local search in the later stages. Attached Figure Description

[0047] Figure 1 This is a flowchart illustrating the fault location method for distribution networks with distributed power sources based on the improved lemming algorithm according to an embodiment of the present invention.

[0048] Figure 2A schematic diagram illustrating the IEEE 33-node distribution network model containing distributed generation in the embodiments;

[0049] Figure 3 This is a schematic diagram illustrating the iterative optimization of candidate fault interval combinations using the improved lemming algorithm in the embodiment.

[0050] Figure 4 This is a schematic diagram comparing the optimal point set strategy with the existing Logistic initialization of the population distribution in the embodiment, wherein (a) is a schematic diagram of the optimal point set strategy for initializing the population distribution, and (b) is a schematic diagram of the existing Logistic initialization of the population distribution.

[0051] Figure 5 This is a schematic diagram comparing the simulation results of the power grid fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a single-point fault scenario.

[0052] Figure 6 This is a schematic diagram comparing the simulation results of the distribution network fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a single-point fault scenario with distortion information.

[0053] Figure 7 This is a schematic diagram comparing the simulation results of the power grid fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a multi-point fault scenario.

[0054] Figure 8 This is a schematic diagram comparing the simulation results of the power grid fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a multi-point fault scenario with distortion information. Detailed Implementation

[0055] The preferred embodiments of the present invention will now be described in detail with reference to the accompanying drawings.

[0056] The embodiment provides a fault location method for distribution networks with distributed power sources based on an improved lemming algorithm, such as... Figure 1 This includes the following steps:

[0057] S1. Based on the network topology of the distribution network, the switch state coding method is adopted to define the switch state coding vector R and construct the switch state function.

[0058] S11. Suppose a distribution network has N branches, i.e., N feeder segments. Each branch of the distribution network may be in either a normal or fault state. A binary mapping rule is used to represent the fault state of the distribution network, and the fault state vector is defined as S = [s1, s2, ..., s...]. j , ..., s n ], where the operating state s of the j-th branch. j :

[0059] ;

[0060] S12. Based on the multidirectional nature of fault currents after distributed generation is connected to the distribution network, a switch state coding method is adopted. The direction from the main power source to the end of the line is defined as the positive direction, and otherwise as the negative direction. The switch state coding vector R = [R1, R2, ..., R...] is defined. j , ..., R N ], where R1, R2, ..., R j , ..., R N These represent the operating states of branches 1, 2, ..., j, ..., N in the distribution network, respectively. A switch detects a fault state R in the j-th branch. j :

[0061] ;

[0062] S13. In traditional single-source radial networks, only the upstream fault current needs to be considered when a fault occurs, as no fault current flows downstream. However, when distributed generation (DG) is integrated into the distribution network, the switching function used for fault location in traditional distribution networks becomes ineffective. Therefore, the expression for the switching state function suitable for distribution networks containing DG is as follows:

[0063]

[0064] in, Indicates the desired state information of the switch; and Let J represent the switch state functions upstream and downstream of switch j, respectively, with the following expressions:

[0065]

[0066]

[0067] in, Representation or logic; and These represent the total number of upstream and downstream feeder sections of switch j, respectively. and These represent the number of power sources upstream and downstream of switch j, respectively; and These indicate whether the upstream and downstream DG of switch j are connected to the grid, respectively. and These represent the state quantities of each segment between switch j and the upstream and downstream power sources, respectively. and These represent the state variables of all feeder sections upstream and downstream of switch j, respectively.

[0068] S2. Objective function for constructing a fault location model for a distribution network containing distributed power sources.

[0069] In step S2, the objective function The expression is as follows:

[0070]

[0071] in, The actual switch status value is given, where M represents the total number of feeder switches. is the penalty coefficient, used to adjust the impact of the fault quantity constraint on the objective function; N is the total number of feeder sections; Indicates the number of faulty sections.

[0072] In step S2, the constructed objective function satisfies the requirement that the calculated result is as close as possible to the actual transmitted information, avoids excessive redundancy of objective functions under the premise of detecting information matching, and also prevents missing or multiple solutions.

[0073] S3. Encode the fault status information uploaded by the feeder terminal (FTU) of the distribution network using the switch status encoding method in step S1 to obtain the actual switch status value. The desired switch state information is obtained from the switch state function. .

[0074] S4. Set the actual switch status value With switch expected state information Substitute the objective function into the target function and calculate the objective function value for each individual in the algorithm population.

[0075] S5. An initial population is generated using a best-point set strategy. An elite pool strategy and an adaptive Gaussian perturbation strategy are introduced. The objective function value of each individual is calculated using step S4. The improved lemming algorithm is then used to iteratively optimize the combination of candidate fault intervals to obtain the optimal individual. Figure 3 :

[0076] S51. Set the population size D and the maximum number of iterations T, and initialize the population using the optimal point set strategy;

[0077] In step S51, the population is initialized using the optimal point set strategy, specifically as follows:

[0078] Let GD For a unit cube in D-dimensional Euclidean space, the set of optimal points for Among them, Jia Dian Ji deviation Where n represents the population size, It is only with and positive numbers Relevant constants, Indicates the first The dimension refers to the optimal point parameter used for the j-th branch, where j = 1, 2, ..., D. Indicates the first The position of the initial individual; take , It is to satisfy The smallest prime number; spread the points of the optimal point set evenly and map them to the upper and lower bounds of the variable.

[0079] In step S51, since the quality of the initial population is very important for the intelligent optimization algorithm, and if... Figure 4 (b) The original lemming algorithm initializes the population randomly, which can lead to uneven distribution of individuals and poor traversal. This embodiment introduces a best-set initialization strategy, such as... Figure 4 (a) The optimal point set strategy has strong uniformity and traversal, and is an efficient and uniform point selection method.

[0080] S52. Introduce an elite pool strategy: Calculate the fitness value of each individual in the t-th iteration, which is the objective function value, and construct an elite pool based on the fitness values.

[0081]

[0082] in, These are the positions of the top three elite individuals in the population at iteration t, ordered from highest to lowest fitness value, with fitness values ​​of [values ​​to be filled in]. ,satisfy ; To construct elite guiding positions; These are the weight coefficients of elite individuals, and their sum is 1.

[0083] In step S52, the original lemming algorithm primarily relies on the currently globally optimal individual to guide the population during position updates. This mechanism enables the population to quickly move towards the currently better region. However, in complex optimization problems, especially in multimodal function optimization, relying solely on a single optimal individual may not find the true optimal solution and can easily lead to local optima. This embodiment introduces an elite pool strategy, selecting multiple individuals with relatively high fitness from the current population to form an elite individual set. These elite individuals then participate in the subsequent position update process. Elite individuals are used to guide the position update. Replace the current optimal solution in the original lemming algorithm's position update formula. Compared to using only Provide guidance, elite guidance location It can more comprehensively reflect the current population's best search area, thereby improving the stability of the search process.

[0084] S53. Calculate the energy factor E. If the energy factor E > 1, proceed to the next step S54; otherwise, proceed to the next step S55.

[0085] S54. Determine whether the random number Rand in the interval (0,1) is less than the first set value. If so, update the lemming position according to the long-distance migration behavior and proceed to the next step S56; otherwise, update the lemming position according to the burrowing behavior and proceed to the next step S56.

[0086] S55. Determine whether the random number Rand in the interval (0,1) is less than the second set value. If so, update the lemming position according to the foraging behavior and proceed to the next step S56; otherwise, update the lemming position according to the predator avoidance behavior and proceed to the next step S56.

[0087] S56. Calculate fitness values ​​and update lemming positions. ;

[0088] S57. Introduce an adaptive Gaussian perturbation strategy to obtain the lemming position after Gaussian perturbation. ;

[0089] In step S57, the lemming position after Gaussian perturbation is obtained. :

[0090]

[0091] in, For disturbance terms; For a Gaussian perturbation random term that follows a standard normal distribution; This represents the adaptive disturbance factor, used to control the disturbance intensity. , where η min This represents the minimum disturbance value; η max t represents the maximum perturbation value; t represents the current iteration number; and T represents the maximum iteration number.

[0092] In step S57, to further prevent the lemming algorithm from getting trapped in local optima, an adaptive Gaussian perturbation strategy is introduced. Utilizing the characteristics of Gaussian distribution perturbation, such as concentration, symmetry, and uniform variation, this strategy dynamically perturbs individuals in the algorithm, thereby enhancing the population's fine-grained search ability in local regions and further improving the algorithm's convergence accuracy. Simultaneously, an adaptive perturbation factor is introduced, enabling the algorithm to maintain strong global exploration capabilities in the early stages of iteration and gradually shift towards local exploration in the later stages. This strategy combines the perturbation factor with Gaussian perturbation to provide secondary perturbations for candidate solutions. When t is small, A larger perturbation intensity helps to expand the search range of candidate solutions; as t gradually approaches T... The perturbation intensity is relatively small, which helps to conduct a fine-grained search near the high-quality solution region. In addition, the perturbation magnitude is also affected by the elite pool and is related to the distance of the individual from the elite guidance position. Individuals far away from the elite region will receive a larger perturbation, which helps to expand the search range; when the individual is close to the elite region, the perturbation is smaller, which helps to conduct a fine-grained search around the current high-quality solution.

[0093] S58. Compare the lemming positions after Gaussian perturbation. and lemming location The best solution is selected as the optimal solution for the t-th iteration.

[0094] S59. After updating the elite pool according to fitness, if t < T, let t = t + 1 and return to step S53; otherwise, take the optimal solution of the t-th iteration as the optimal individual.

[0095] S6. Using binary mapping rules, the continuous positions corresponding to the combination of fault intervals are converted into fault segment states, and finally the fault interval corresponding to the optimal individual is output.

[0096] This fault location method for distribution networks with distributed power sources, based on an improved lemming algorithm, can efficiently locate faults and has a better ability to escape local optima. It is more effective for locating multiple faults with distorted information. Experimental simulations have verified that it can achieve fast location speed and high location accuracy in distribution networks with different topologies.

[0097] This method for fault location in distribution networks with distributed power sources, based on an improved lemming algorithm, constructs a switch state function and an objective function for the fault location model of the distribution network with distributed power sources. It initializes the population using a best-point set initialization strategy, improving the initial search quality, on top of the standard lemming optimization algorithm. An elite pool strategy is introduced, using multiple high-performing individuals to construct a weighted guidance mechanism, which collectively guides population updates, enhancing the algorithm's ability to escape local optima and helping it find the optimal solution faster. An adaptive Gaussian perturbation strategy is also introduced, providing a larger perturbation in the early stages of iteration and a smaller perturbation in the later stages, balancing global exploration in the early stages with refined local search in the later stages.

[0098] This method for fault location in distribution networks with distributed power sources, based on an improved lemming algorithm, provides a fault location method that combines a multi-strategy optimization mechanism with distribution network fault segment encoding. Each individual represents a candidate fault interval combination. The improved lemming optimization mechanism searches for the optimal individual, and the fault segment is determined based on the optimal individual. In this embodiment, the best point set strategy is applied during the population initialization stage to improve the uniformity of the initial candidate solutions in the search space; the elite pool strategy is applied during the iterative optimization stage to preserve and guide excellent solutions; and the adaptive Gaussian perturbation strategy is applied during the population update or elite individual perturbation stage. The method in this embodiment uses the optimal individual after best point set initialization, elite pool guidance, and adaptive Gaussian perturbation optimization as the fault location result, and outputs the distribution network segment representing the fault state in the optimal individual. The best point set strategy, elite pool strategy, and adaptive Gaussian perturbation strategy work together to improve the initial population quality, convergence speed, and ability to escape local optima of the algorithm.

[0099] This fault location method for distribution networks with distributed power sources, based on an improved lemming algorithm, transforms the fault section location problem into a binary combinatorial optimization problem. It constructs individual codes representing fault states using each candidate section of the distribution network as optimization variables. A fitness function is established based on information uploaded from the distribution network feeder terminals, switch states, protection action information, and network topology. The method introduces a best-point set strategy, an elite pool strategy, and an adaptive Gaussian perturbation strategy. It iteratively optimizes the combination of candidate fault sections using the improved lemming algorithm and converts continuous positions of the algorithm into fault section states through binary mapping rules. Finally, it outputs the fault section corresponding to the optimal individual, achieving efficient and high-precision fault location.

[0100] The simulation experiment verifies the distribution network fault location method based on the improved lemming optimization algorithm in this embodiment as follows:

[0101] The simulation experiment was conducted on the Matlab platform, using an IEEE 33-node distribution network model containing distributed generation sources, such as... Figure 2 , Figure 2 In the diagram, G represents the system's main power supply, and L1, L2, ..., L... 33Represents feeder segments, s1, s2, ..., s 33 These represent line node switches, with K1, K2, and K3 representing the switching switches for distributed power sources DG1, DG2, and DG3, respectively. The relevant parameters are set as follows: the total population size is equivalent to the total number of branches. Dimension Maximum number of iterations The effectiveness of the proposed method was verified by varying the location and number of distributed power sources and setting up fault scenarios of different complexities. Furthermore, the improved lemming algorithm was compared with the standard lemming algorithm (ALA), beluga algorithm (BWO), and sparrow algorithm (SSA), with average iteration time and average location accuracy used as evaluation criteria for global optimization efficiency and convergence speed.

[0102] (1) Simulation analysis of single point of failure

[0103] Fault location experiments were conducted for single-point faults under different conditions of the number and location of distributed generation (DG) deployments. The simulation results are shown in Table 1. Taking the third case of single-point fault in Table 1 as an example: [K1,K2,K3]=[1,0,1] indicates that the first and third distributed generation sources are connected to the grid. 20 When a fault occurs, specifically when a fault occurs in section 20, the fault status information uploaded by the feeder terminal (FTU) is [1,1,−1,−1,−1,0,0,0,0,0,0,0,0,0,0,0,0,1,1,1,−1,−1,−1,−1,−1,−1,−1,−1,−1,0,0,0,0]. The simulation location output is S=[0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0,0,0,0], and the output fault location result is L. 20 . Figure 5 This is a schematic diagram comparing the simulation results of the distribution network fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a single-point fault scenario. Figure 5 In the diagram, the red curve represents the convergence curve of the improved lemming algorithm IALA from the embodiment. Figure 5 As can be seen, compared with the other three algorithms, the improved lemming algorithm IALA in the embodiment converges significantly at an earlier number of iterations and has a faster convergence speed.

[0104] Table 1 Simulation results of single-point faults

[0105]

[0106] (2) Multi-point fault simulation analysis

[0107] Fault location experiments were conducted for multi-point faults under different conditions of DG deployment quantity and location, and the simulation results are shown in Table 2. Taking Case 4 of multi-point faults as an example: [K1,K2,K3]=[1,1,1] indicates that all three distributed power sources are connected to the grid, when L 12 L 19 L 25 When a fault occurs, specifically in sections 12, 19, and 25, the fault status information uploaded by the feeder terminal (FTU) is [1,1,1,1,1,1,1,1,1,1,1,1,−1,−1,−1,−1,−1,−1,1,1,−1,−1,1,1,1,1,−1,−1,−1,−1,0,0,0,0]. The simulation location output is S=[0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,1,0,0,0,0,0,0,0,0]. The output fault location result is L. 12 L 19 L 25 . Figure 7 This is a comparative diagram of simulation results between the distribution network fault location method based on the improved lemming algorithm in the embodiment and the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a multi-point fault scenario. The red curve is the convergence curve of the improved lemming algorithm (IALA) in the embodiment. Figure 7 As can be seen, the improved lemming algorithm IALA in the embodiment escapes the local optimum on the 16th attempt and then locates the faulty section. The lemming algorithm ALA only escapes the local optimum and locates the faulty section on the 51st attempt, and the sparrow algorithm only escapes the local optimum and locates the faulty section on the 43rd attempt. The white whale algorithm BWO, however, gets stuck in local convergence and never escapes, ultimately failing to locate the preset faulty section. Compared to the other three algorithms, the improved lemming algorithm IALA in the embodiment has a faster convergence speed and more stable performance.

[0108] Table 2 Simulation results of multi-point faults

[0109]

[0110] (3) Fault simulation analysis with distorted information

[0111] Due to the complex working environment of the power distribution network, data distortion may occur in FTU equipment during data transmission. When a line fault occurs, the distorted information increases the difficulty of fault location and affects the accuracy of fault location. Therefore, simulation experiments were conducted to address information distortion during single-point and multi-point faults. Under normal conditions, when a network fault occurs, the fault status information encoding displays 1 or -1 for sections with fault current flowing through it, and distorted information displays 0; for sections without fault current flowing through it, the encoding is 0, and distorted information displays 1. The simulation results are shown in Table 3. Taking Case 4, a fault with distorted information, as an example: [K1,K2,K3]=[1,1,1] indicates that three distributed power sources are connected to the grid, and the location of the distorted information is set to L. 11 L 31 The fault status code for section 11 should have been 1, but it was distorted to 0; the fault status code for section 31 should have been 0, but it was distorted to 1. Therefore, the fault status information uploaded by the feeder terminal (FTU) is [1,1,1,1,1,1,1,1,1,1,0,1,1,1,−1,−1,1,−1,−1,−1,−1,−1,−1,−1,−1,0,1,0,0]. The simulation location output is S=[0,0,0,0,0,0,0,0,0,0,0,0,0,0,1,0,0,0,1,0,0,0,0,1,0,0,0,0,0,0,0,0,0,0]. The output fault location result is... . Figure 6 This is a schematic diagram comparing the simulation results of the power grid fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale Algorithm (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a single-point fault scenario with distortion information. Figure 8 This is a schematic diagram comparing the simulation results of the power grid fault location method based on the improved lemming algorithm in the embodiment with the existing Beluga Whale (BWO), Sparrow Algorithm (SSA), and Lemming Algorithm (ALA) in a multi-point fault scenario with distortion information. Figure 6 and Figure 8 As can be seen, the improved lemming algorithm IALA has a faster convergence speed and more stable localization performance.

[0112] Table 3 Fault simulation results containing distortion information

[0113]

[0114] (4) Analysis of algorithm stability and convergence

[0115] To avoid the randomness of the experiment, each algorithm was run independently 100 times in the simulation, and the results of multiple runs were averaged. The average of the simulation results was taken according to the evaluation index. The results of single-point fault location, multi-point fault location, and fault location with distortion information are shown in Tables 4-6.

[0116] Table 4 Comparison of Single-Point Fault Location Algorithm Results

[0117]

[0118] Table 5 Comparison of results from multi-point fault location algorithms

[0119]

[0120] Table 6 Comparison of Fault Location Algorithm Results Including Distortion Information

[0121]

[0122] As shown in Table 4, for single-point fault scenarios, the fault situation is relatively simple, and all four algorithms have high localization accuracy. However, the average number of iterations for the improved lemming algorithm IALA in the embodiments is 4.25, while that for the lemming algorithm ALA is 8.73, the Sparrow algorithm SSA is 7.42, and the Beluga algorithm BWO is 5.56. The average number of iterations for the improved lemming algorithm IALA in the embodiments is less than that for the three comparative algorithms, indicating that the improved lemming algorithm IALA in the embodiments can reach convergence earlier in each run. As shown in Table 5, in multi-point fault scenarios, the fault situation becomes more complex. After 100 independent runs, the improved lemming algorithm IALA in the embodiments has a localization accuracy of 80%, which is significantly higher than the other three algorithms, and its average number of iterations is 23.68, which is also less than that for the three comparative algorithms. This indicates that the improved lemming algorithm IALA in the embodiments has high localization accuracy and fast convergence speed in handling multi-point fault problems. As shown in Table 6, in fault location scenarios containing distorted information, the improved lemming algorithm IALA in the embodiment achieves a localization accuracy of up to 98%, while the localization accuracy of the lemming algorithms ALA, SSA, and BWO is far lower than that of the improved lemming algorithm IALA in the embodiment. Furthermore, the improved lemming algorithm IALA in the embodiment converges on average in 12-13 runs per cycle, which is significantly faster than the comparative algorithms.

[0123] In summary, the improved lemming algorithm IALA demonstrated in this embodiment exhibits high fault location accuracy and fast convergence speed across three distribution network fault scenarios: single-point faults, multi-point faults, and faults containing distorted information. Although the computational load per generation slightly increases due to the introduction of various strategies, the average convergence time remains comparable to other comparative algorithms due to the significantly reduced number of iterations required to reach the optimal value, maintaining its overall optimization efficiency advantage. Furthermore, in complex scenarios such as multi-point faults and fault location with distorted information, the improved lemming algorithm IALA in this embodiment demonstrates superior anti-interference capability and optimization accuracy compared to other comparative algorithms.

[0124] Although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art can still modify the technical solutions described in the foregoing embodiments or make equivalent substitutions for some of the technical features. Any modifications, equivalent substitutions, improvements, etc., made within the spirit and principles of the present invention should be included within the protection scope of the present invention.

Claims

1. A fault location method for distribution networks with distributed power sources based on an improved lemming algorithm, characterized in that: Includes the following steps, S1. Based on the network topology of the distribution network, a switch state coding method is adopted to define the switch state coding vector R and construct the switch state function; S2. Objective function for constructing a fault location model for a distribution network containing distributed generation sources; S3. Encode the fault status information uploaded by the feeder terminal (FTU) of the distribution network using the switch status encoding method in step S1 to obtain the actual switch status value. The desired switch state information is obtained from the switch state function. ; S4. Set the actual switch status value With switch expected state information Substitute the objective function into the algorithm population and calculate the objective function value for each individual. S5. The initial population is generated using the optimal point set strategy. The elite pool strategy and the adaptive Gaussian perturbation strategy are introduced. The objective function value of the individual is calculated using step S4. The improved lemming algorithm is used to iteratively optimize the combination of candidate fault intervals to obtain the optimal individual. S6. Using binary mapping rules, the continuous positions corresponding to the combination of fault intervals are converted into fault segment states, and finally the fault interval corresponding to the optimal individual is output.

2. The fault location method for distribution networks with distributed power sources based on the improved lemming algorithm as described in claim 1, characterized in that: Step S1, specifically, S11. Suppose a distribution network has N branches, i.e., N feeder segments. Each branch of the distribution network may be in either a normal or fault state. A binary mapping rule is used to represent the fault state of the distribution network, and the fault state vector is defined as S = [s1, s2, ..., s...]. j , ..., s N ], where the operating state s of the j-th branch. j : ； S12. Based on the multidirectional nature of fault currents after distributed generation is connected to the distribution network, a switch state coding method is adopted. The direction from the main power source to the end of the line is defined as the positive direction, and otherwise as the negative direction. The switch state coding vector R = [R1, R2, ..., R...] is defined. j , ..., R N ], where R1, R2, ..., R j , ..., R N These represent the operating states of branches 1, 2, ..., j, ..., N in the distribution network, respectively. A switch detects a fault state R in the j-th branch. j : ； S13. The expression for the switching state function applicable to distribution networks containing distributed generation sources is as follows: ， in, Indicates the desired state information of the switch; and Let J represent the switch state functions upstream and downstream of switch j, respectively, with the following expressions: ， ， in, Representation or logic; and These represent the total number of upstream and downstream feeder sections of switch j, respectively. and These represent the number of power sources upstream and downstream of switch j, respectively; and These indicate whether the upstream and downstream DG of switch j are connected to the grid, respectively. and These represent the state quantities of each segment between switch j and the upstream and downstream power sources, respectively. and These represent the state variables of all feeder sections upstream and downstream of switch j, respectively.

3. The fault location method for distribution networks with distributed power sources based on the improved lemming algorithm as described in claim 2, characterized in that: In step S2, the objective function The expression is as follows: ， in, The actual switch status value is given, where M represents the total number of feeder switches. is the penalty coefficient, used to adjust the impact of the fault quantity constraint on the objective function; N is the total number of feeder sections; Indicates the number of faulty sections.

4. The method for fault location in a distribution network with distributed power sources based on the improved lemming algorithm as described in any one of claims 1-3, characterized in that: Step S5, specifically, S51. Set the population size D and the maximum number of iterations T, and initialize the population using the optimal point set strategy; S52. Introduce an elite pool strategy: Calculate the fitness value of each individual in the t-th iteration, which is the objective function value, and construct an elite pool based on the fitness values. ， in, These are the positions of the top three elite individuals in the population at iteration t, ordered from highest to lowest fitness value, with fitness values ​​of [values ​​to be filled in]. ,satisfy ; To construct elite guiding positions; These are the weight coefficients of elite individuals, and their sum is 1. S53. Calculate the energy factor E. If the energy factor E > 1, proceed to the next step S54; otherwise, proceed to the next step S55. S54. Determine whether the random number Rand in the interval (0,1) is less than the first set value. If so, update the lemming position according to the long-distance migration behavior and proceed to the next step S56; otherwise, update the lemming position according to the burrowing behavior and proceed to the next step S56. S55. Determine whether the random number Rand in the interval (0,1) is less than the second set value. If so, update the lemming position according to the foraging behavior and proceed to the next step S56; otherwise, update the lemming position according to the predator avoidance behavior and proceed to the next step S56. S56. Calculate fitness values ​​and update lemming positions. ; S57. Introduce an adaptive Gaussian perturbation strategy to obtain the lemming position after Gaussian perturbation. ; S58. Compare the lemming positions after Gaussian perturbation. and lemming location The best solution is selected as the optimal solution for the t-th iteration. S59. After updating the elite pool according to fitness, if t < T, let t = t + 1 and return to step S53; otherwise, take the optimal solution of the t-th iteration as the optimal individual.

5. The fault location method for distribution networks with distributed power sources based on the improved lemming algorithm as described in claim 4, characterized in that: In step S51, the population is initialized using the optimal point set strategy, specifically as follows: Let G D For a unit cube in D-dimensional Euclidean space, the set of optimal points for Among them, Jia Dian Ji deviation Where n represents the population size, It is only with and positive numbers Relevant constants, Indicates the first The dimension refers to the optimal point parameter used for the j-th branch, where j = 1, 2, ..., D. Indicates the first The position of the initial individual; take , It is to satisfy The smallest prime number; spread the points of the optimal point set evenly and map them to the upper and lower bounds of the variable.

6. The fault location method for distribution networks with distributed power sources based on the improved lemming algorithm as described in claim 4, characterized in that: In step S57, the lemming position after Gaussian perturbation is obtained. : ， in, For disturbance terms; For a Gaussian perturbation random term that follows a standard normal distribution; This represents the adaptive disturbance factor, used to control the disturbance intensity. , where η min This represents the minimum disturbance value; η max t represents the maximum perturbation value; t represents the current iteration number; and T represents the maximum iteration number.

Images