Distribution network relay protection setting method and system based on multi-source data

Abstract

The present invention relates to the field of data processing, and more specifically, to a method and system for setting relay protection for a distribution network based on multi-source data. The method comprises: in a gray wolf optimization algorithm, determining the contribution weight of each relay, obtaining the fitness of the relay according to an objective function, selecting the positions of three optimal relays from a population, adjusting the positions of the remaining relays based on a coordination vector, and correcting the convergence factor to obtain a dynamic step factor. For any target relay, the slope between it and the adjacent relay is calculated to obtain an update coefficient. The dynamic step factor is multiplied by the update coefficient to obtain an adjusted search amount to guide the update of the relay position. In a preset number of consecutive iterations, the number and mean of non-zero update coefficients are recorded, and a convergence value is calculated. Whether to terminate the iteration is determined based on the convergence value, and the position and optimal fitness of the optimal relay are output to improve the search capability and convergence speed of the algorithm.

Description

Technical Field

[0001] The present invention relates to the field of data processing, and more specifically, to a distribution network relay protection setting method and system based on multi-source data. Background Art

[0002] Relay protection is an important part of the distribution network. It mainly takes corresponding protection actions for distribution network faults to ensure the normal operation of the distribution network. During the operation of the distribution network, it is extremely difficult to customize relay protection that can coordinate the relay protection of the entire network. In practical applications, the relay protection action time is long and the error is large, and the optimization effect is not ideal. For example, communication delays and the complexity of the power grid will affect the fault handling speed and the accuracy of relay protection. The existing Grey Wolf Optimizer (GWO) is a swarm intelligence optimization algorithm that simulates the predation behavior of gray wolves. The algorithm randomly initializes a group of wolves (solutions), and then updates the position of the wolves by simulating the predation behavior of the wolf pack, thereby finding the optimal solution and quickly obtaining the relay that needs to be optimized.

[0003] The existing Chinese patent application document with publication number CN117578368A discloses a method for setting relay protection constants of an active distribution network, comprising the following steps: S1: obtaining a recording file of a distribution network feeder; S2: establishing a constant coordination optimization model based on the obtained recording file and drawing a distribution network topology diagram; S3: associating the recording file with the feeder through the constant coordination optimization model and the distribution network topology diagram; S5: determining a first weight value corresponding to each fault recording file; S6: obtaining constant value data according to multiple line fault points, that is, the current level constant value data; S7: finding the current level system to which the fault point position on the distribution network topology diagram belongs; S8: comparing the current level constant value data with the upper level constant value data and the lower level constant value data; S9: determining a constant value adjustment strategy based on the auxiliary analysis results; the present invention balances the failure rate of each line of the distribution network and reduces the failure loss by updating the initial relay protection constants of each of the multiple line switches.

[0004] The above scheme analyzes and determines the recording files of multiple lines in the distribution network, obtains recording data based on the recording file analysis, and then uses the fault recording data to determine the fault frequency of multiple line sections. However, the convergence factor is a key parameter in the current Grey Wolf optimization algorithm. The convergence factor setting decreases linearly. In the early stage of the algorithm, a global search is not fully performed, which may lead to falling into the local optimum, thereby reducing the convergence performance of the algorithm and reducing the computational efficiency of the algorithm. Summary of the invention

[0005] In order to solve the problem that the convergence factor setting is linearly decreasing, and global search is not fully performed in the early stage of the algorithm, which may lead to falling into the local optimum and thus the convergence performance of the algorithm is reduced, the present invention provides solutions in the following aspects.

[0006] In the first aspect, a distribution network relay protection setting method based on multi-source data includes: obtaining the contribution weight of each relay; based on the objective function in the gray wolf optimization algorithm, calculating the fitness of each relay, obtaining three optimal relays in the population, updating the control search direction and step size and the randomness of the search based on the three optimal relays, and adjusting the positions of the remaining relays based on the coordination vector; adjusting the convergence factor based on the positions of the remaining relays in the iteration process based on the coordination vector to obtain a dynamic step size factor; any relay is a target relay, and the slope of the position between the target relay and the two adjacent left and right relays is calculated respectively to obtain the non-zero maximum slope, and the calculation Calculate the weighted sum of all relay starting current setting values ​​and contribution weights, calculate the difference between the iterative position of the target relay on the right adjacent relay and the weighted sum and normalize it, and use the product of the maximum slope and the normalized difference as the iterative update coefficient of the target relay; use the product of the dynamic step factor and the update coefficient as the adjustment search amount to guide the position update of the target relay; based on the adjustment search amount, in the process of continuous preset number of iterations, obtain the number of times the relay update coefficient is non-zero and the mean value of the update coefficient to calculate the convergence value of each relay; based on the convergence value, determine whether to terminate the iteration, and output the position and optimal fitness of the optimal relay.

[0007] The effect is as follows: by calculating the fitness of each relay, the optimal starting current setting value can be found, thereby improving the operating efficiency and reliability of the distribution network, and using the information of the optimal relay to dynamically adjust the search direction and step size to make the search process more flexible and efficient. Through the dynamic step size factor and update coefficient, the algorithm can balance between global search and local search, improve the probability of finding the global optimal solution, and reduce the number of iterations; by considering the position and fitness of adjacent relays, the algorithm can identify and jump out of the local optimal solution and continue to search for the global optimal solution, which helps to improve the stability of the power system.

[0008] Preferably, obtaining the contribution degree of each relay includes:

[0009] Define the target layer: evaluate the contribution of relays in the distribution network; define the criterion layer including but not limited to: fault detection speed, fault isolation capability, false operation rate, maintenance cost, etc.; use the 1-9 scaling method to construct a judgment matrix based on each criterion of the criterion layer, and perform eigenvalue decomposition on the judgment matrix to obtain the weight of each criterion layer;

[0010] The weight of each criterion layer is multiplied and accumulated by the weight of each corresponding relay to obtain the contribution weight of the relay under the overall goal. The starting current setting value of each relay .

[0011] The effect is that by multiplying and accumulating the criterion layer weight and the relay weight, the comprehensive contribution weight of each relay is obtained, which helps to optimize the starting current setting value of the relay. By adjusting the starting current setting value of the relay according to the contribution weight, the fault response speed and fault handling capability of the distribution network can be improved.

[0012] Preferably, the coordination vector includes: and synergy vector , for the synergy vector Make dynamic adjustments;

[0013] Preset initialization convergence factor and create 1 row A random matrix of columns, where The columns represent the dimensions of the relay individual position space, and the value range of each element in the matrix is ; Multiply each element in the random matrix by 2 to obtain an update matrix, the range of which is expanded to ;

[0014] Create a matrix of the same size as the update matrix, with all elements set to 1, labeled as a constant matrix, and take the difference between the update matrix and the constant matrix to get a labeled matrix with element values ​​ranging from , multiply the label matrix by the convergence factor to obtain the synergy vector .

[0015] The effect is that by presetting the initialization convergence factor and creating a random matrix, the algorithm can adapt to different optimization problems and search spaces, enhancing the versatility and adaptability of the algorithm. The introduction of the random matrix adds diversity to the search process, which helps the algorithm explore a wider solution space and avoid premature convergence.

[0016] Preferably, obtaining the dynamic step factor includes:

[0017] Take any of the remaining relays as the target relay and calculate the fitness of the target relay after iteration. , and the average fitness value of the target relay after iteration, calculate the ratio of the fitness value to the average fitness value, and perform normalization processing to analyze the level value of the relay for the average fitness of the population;

[0018] Calculate the square of the absolute value of the spatial distance between the current position of the target relay and the optimal solution, and take the ratio of the square of the spatial size of the upper and lower limits of the current setting value as the total width of the search space; take twice the product of the horizontal value and the total width as the dynamic step factor.

[0019] The effect is: by calculating the ratio of the fitness of the target relay to the average fitness of the population, and combining the spatial distance between the target relay and the optimal solution, the step factor is dynamically adjusted, the total width of the search space is considered, and the step factor is ensured to be proportional to the size of the solution space, thereby achieving a balance between global search and local search, avoiding falling into the local optimum, and ensuring that it can continue to approach the global optimal solution.

[0020] Preferably, the dynamic step factor also satisfies the following relationship:

[0021] ;

[0022] In the formula, Indicates Relay No. The dynamic step size factor after iterations, Indicates the current iteration number, Indicates the maximum number of iterations.

[0023] Preferably, the convergence value satisfies the following relationship:

[0024] ;

[0025] In the formula, Indicates The convergence value of the relay, Indicates the number of times the update coefficient is not zero in the preset number of iterations, Indicates the preset number of iterations, Indicates continuous The proportion of non-zero update coefficients in iterations, Indicates Each relay updates the mean value of the coefficients in a preset number of iterations.

[0026] The effect is that the convergence trend of the relay can be effectively quantified by the proportion of non-zero update coefficients of the relay in the preset number of iterations and the mean value of the update coefficients, thereby dynamically adjusting the search strategy and improving the search efficiency.

[0027] Preferably, judging whether to terminate the iteration based on the convergence value includes:

[0028] In response to the convergence value being greater than a preset threshold or reaching a maximum number of iterations, the iteration is terminated, and the optimal configuration of the distribution network relay protection setting under given constraints and objectives is output.

[0029] In a second aspect, a distribution network relay protection setting system based on multi-source data includes: a processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the above-mentioned distribution network relay protection setting method based on multi-source data is implemented.

[0030] The present invention has the following effects:

[0031] 1. The present invention introduces nonlinear factors of fitness and distance, and according to the dynamic step size factor, increases the proportion of global search in the early stage of search, and gradually turns to local search in the later stage of search, thereby improving the search ability and convergence speed of the algorithm.

[0032] 2. By analyzing the slope and distance between adjacent relay individuals, as well as the distance from the weighted sum of all relay starting current setting values, the present invention can more accurately indicate the movement direction of the relay individuals in the search space, and converge to the optimal solution as quickly as possible with the least number of iterations. BRIEF DESCRIPTION OF THE DRAWINGS

[0033] By reading the following detailed description with reference to the accompanying drawings, the above and other objects, features and advantages of the exemplary embodiments of the present invention will become readily understood. In the accompanying drawings, several embodiments of the present invention are shown in an exemplary and non-limiting manner, and the same or corresponding reference numerals represent the same or corresponding parts, wherein:

[0034] Figure 1 It is a method flow chart of steps S1 to S5 in the distribution network relay protection setting method based on multi-source data in an embodiment of the present invention.

[0035] Figure 2 It is a structural block diagram of a distribution network relay protection setting system based on multi-source data according to an embodiment of the present invention. DETAILED DESCRIPTION

[0036] The following will be combined with the drawings in the embodiments of the present invention to clearly and completely describe the technical solutions in the embodiments of the present invention. Obviously, the described embodiments are part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative work are within the scope of protection of the present invention.

[0037] The specific embodiments of the present invention are described in detail below with reference to the accompanying drawings.

[0038] The distribution network relay protection setting optimization proposed in the present invention targets the parameters to be optimized of the moving current setting value, so as to ensure that the distribution network has sensitivity, speed and selectivity when any line of the distribution network fails.

[0039] Reference Figure 1 The distribution network relay protection setting method based on multi-source data includes steps S1 to S5, which are as follows:

[0040] S1: Get the contribution weight of each relay.

[0041] It is further explained that the relay protection system of the distribution network is a complex network composed of multiple relays and protection devices. They work together to ensure the safe and stable operation of the power grid. Therefore, when optimizing the relay protection setting of the entire distribution network, it is necessary to consider the contribution of each relay and optimize the starting current setting value based on the contribution weight of each relay.

[0042] In this embodiment, the contribution weight of each relay is evaluated using the analytic hierarchy process.

[0043] Specifically, the AHP includes the following steps:

[0044] Define the target layer: evaluate the contribution of relays in the distribution network; define the criterion layer including but not limited to: fault detection speed, fault isolation capability, misoperation rate, maintenance cost, etc., with the bottom layer being each relay; use the 1-9 scaling method to construct a judgment matrix based on each criterion in the criterion layer, and perform eigenvalue decomposition on the judgment matrix to obtain the weight of each criterion layer;

[0045] The weight of each criterion layer is multiplied and accumulated by the weight of each corresponding relay to obtain the contribution weight of the relay under the overall goal. The starting current setting value of each relay .

[0046] Specifically, the contribution weight of the relay satisfies the following relationship:

[0047] ;

[0048] In the formula, Indicates The contribution weight of each relay, Indicates The contribution of each relay is Indicates the total number of relays.

[0049] In addition, the contribution weight of the relay can be obtained through fuzzy analytic hierarchy process (FAHP) and network analytic hierarchy process (ANP).

[0050] S2: Based on the objective function in the gray wolf optimization algorithm, the fitness of each relay is calculated, the three best relays in the population are obtained, the direction and step size of the control search and the randomness of the search are updated based on the three best relays, and the positions of the remaining relays are adjusted based on the coordination vector.

[0051] It is further explained that the gray wolf optimization algorithm uses the minimum total time of protection actions as the objective function. The initial distribution of the population will directly affect the quality of the final optimization result. That is to say, the population initialization should be evenly distributed in the design space. The population initialization based on the optimal point set theory ensures that the gray wolf population initialization is evenly distributed.

[0052] The synergy vectors include: and synergy vector , for the synergy vector Make dynamic adjustments;

[0053] Further explanation, the synergy vector Control direction and compensation, coordinated vector Indicates the randomness of the search, so the coordination vector Make further adjustments.

[0054] Specifically, for the coordination vector The steps for dynamic adjustment are as follows:

[0055] Preset initialization convergence factor and create 1 row A random matrix of columns, where is the dimension of the relay individual position space, and the value range of each element in the matrix is ; Multiply each element in the random matrix by 2 to obtain the update matrix. The range of the update matrix is ​​expanded to ;

[0056] Create a matrix of the same size as the update matrix, with all elements set to 1, labeled as a constant matrix, and take the difference between the update matrix and the constant matrix to get a labeled matrix with element values ​​ranging from , multiply the label matrix by the convergence factor to obtain the synergy vector .

[0057] Specifically, the synergy vector satisfies the following relationship:

[0058] ;

[0059] In the formula, Represents the synergy vector , represents the convergence factor, Indicates that the range of randomly generated elements is of OK The column random matrix, The value range of the generated element is of OK Column matrix, Represents the dimension of the individual position space of the relays.

[0060] That is to say, the above , in order to simulate the approximate prey convergence factor It decreases linearly from 2 to 0. The range of variation is .

[0061] Specifically, the initial gray wolf optimization algorithm obtains the coordination vector Here are the steps:

[0062] During the iteration, a random number is generated based on each dimension of the relay. , the range of random numbers is In the example, the double of the random number generated in each dimension is used as the coordination vector on the corresponding dimension. , when updating the position of individual relays, the coordination vector The position of the current relay in each dimension and the optimal relay The positions in corresponding dimensions are combined to calculate the new position.

[0063] For example, when optimizing a relay, three preset parameters need to be adjusted, such as overcurrent, overvoltage, undervoltage, short circuit, etc. The three parameters are analyzed to obtain the optimal solution. Then this problem has a three-dimensional solution space, and each dimension of the relay corresponds to one of the three parameters.

[0064] It should be noted that since it is unknown whether the current position tends to the global optimum, the proportion of global search should be appropriately increased in the early stage of the algorithm. When the linear decrease occurs, the proportion and degree of change of global search and local search are the same, which may lead to local optimum. Therefore, the convergence factor needs to be adjusted. Improvements are made to achieve the coordination vector improvements.

[0065] S3: Adjustment based on the coordination vector Based on the positions of the remaining relays, the convergence factor is corrected during the iteration process to obtain a dynamic step factor.

[0066] Take any of the remaining relays as the target relay and calculate the fitness of the target relay after iteration. , and the average fitness value of the target relay after iteration, calculate the ratio of fitness to the average fitness value, and perform normalization to analyze the level value of the relay for the average fitness of the population;

[0067] Calculate the square of the absolute value of the spatial distance between the current position of the target relay and the optimal solution, and take the ratio of the square of the spatial size of the upper and lower limits of the current setting value as the total width of the search space; take twice the product of the horizontal value and the total width as the dynamic step factor.

[0068] Specifically, the dynamic step factor satisfies the following relationship:

[0069] ;

[0070] In the formula, Indicates Relay No. The dynamic step size factor after iterations, Indicates The exponential function with base , Indicates Relay No. The fitness corresponding to the position after iterations is, Indicates The mean of the fitness corresponding to the individual positions of all relays after iterations, Indicates Relay No. The position after iterations, Indicates The optimal solution for the iteration is Indicates the upper limit of the starting current setting value. Indicates the lower limit of the starting current setting value.

[0071] That is to say, The value of is small, indicating that Relay No. The position corresponding fitness after the iteration is smaller than that of other relays, and the possibility of a local optimal solution appearing nearby is smaller. Should be larger, the relay performs a global search; otherwise, The value of is large, indicating that Relay No. The position corresponding fitness after the iteration is greater than the individual position corresponding fitness of other relays. The current position may be closer to the optimal solution. Its convergence factor should be smaller, the relay performs a local search; Relay No. The position after iteration With The optimal solution of the iteration The greater the distance, The smaller the possibility of an optimal solution appearing nearby, the larger the search step should be in the next iteration; Relay No. The position after iteration With The optimal solution of the iteration The smaller the distance, The greater the possibility of an optimal solution appearing nearby, the smaller the search step size should be in the next iteration.

[0072] Further explanation: the above correction of the convergence factor solves the convergence factor in the basic GWO algorithm. Problems that arise from linear regression, such as failure to fully conduct global search in the early stages of the algorithm, may lead to falling into local optimality. By introducing nonlinear factors of fitness and distance, according to the dynamic step size factor This allows the proportion of global search to be increased in the early stages of the search, while gradually shifting to local search in the later stages of the search, thereby improving the algorithm's search capability and convergence speed.

[0073] In addition, in another embodiment, the linear law is adjusted so that the improved convergence factor changes with the increase in the number of iterations, so that it decreases slowly in the early stage of the iteration and the global search ratio is increased, and the decrease speed is faster in the later stage and the local search ratio is increased, so as to achieve fast convergence to the optimal solution.

[0074] The dynamic step factor also satisfies the following relationship:

[0075] ;

[0076] In the formula, Indicates Relay No. The dynamic step size factor after iterations, Indicates the current iteration number, Indicates the maximum number of iterations.

[0077] Further explanation: the above steps have an impact on the coordination vector The control step size in is adaptively adjusted, and the search direction cannot be obtained more accurately. Therefore, the dynamic step size factor needs to be further corrected. The specific correction steps are as follows:

[0078] S4: Any relay is the target relay. The slope of the position between the target relay and the two adjacent relays on the left and right sides is calculated respectively to obtain the non-zero maximum slope. The weighted sum of the starting current setting values ​​of all relays and the contribution weights is calculated. The difference between the iterative position of the target relay on the right adjacent relay and the weighted sum is calculated and normalized. The product of the maximum slope and the normalized difference is used as the update coefficient of the target relay iteration; the product of the dynamic step factor and the update coefficient is used as the adjustment search amount to guide the position update of the target relay.

[0079] Specifically, the update coefficient of the relay position after iteration satisfies the following relationship:

[0080] ;

[0081] In the formula, Indicates Relay No. The update coefficient of the position after iterations, represents the maximum value function, Indicates Relay No. After the iteration, the position is The relay is in the The slope between the positions after iterations, Indicates Relay No. After the iteration, the position is The relay is in the The slope between the positions after iterations, Indicates The exponential function with base , Indicates The relay is in the After iterations, the position Indicates the total number of relays, Indicates The contribution weight of each relay, Indicates that the first The setting value of a relay.

[0082] Further explanation: According to the fitness and position relationship of the individual relays in the current iteration, the step size of the individual relays is dynamically adjusted. By analyzing the slope and distance between adjacent individual relays, as well as the distance from the weighted sum of the starting current setting values ​​of all relays, the moving direction of the individual relays in the search space can be indicated more accurately, so as to converge to the optimal solution as quickly as possible with the least number of iterations.

[0083] Specifically, the search volume is adjusted to satisfy the following relationship:

[0084] ;

[0085] In the formula, Indicates Relay No. The adjusted search volume after iterations, Indicates Relay No. The update coefficient of the position after iterations, Indicates Relay No. Dynamic step size factor after iterations.

[0086] The position update of individual relays no longer relies entirely on randomness, but is adjusted according to the actual effect during the search process. When , it means that the current position is relatively good and does not need to be updated again. = or = When , it means that one of the adjacent individuals is closer to the local peak point and needs to be updated in that direction. This dynamic adjustment based on the actual search effect enables the algorithm to approach the global optimal solution more efficiently.

[0087] S5: Based on the adjusted search amount, during the continuous preset number of iterations, the number of times the relay update coefficient is non-zero and the average value of the update coefficient are obtained to calculate the convergence value of each relay; based on the convergence value, it is determined whether to terminate the iteration, and the position of the optimal relay and the optimal fitness are output.

[0088] The convergence value satisfies the following relationship:

[0089] ;

[0090] In the formula, Indicates The convergence value of the relay, Indicates the number of times the update coefficient is not zero in the preset number of iterations, Indicates the preset number of iterations, Indicates continuous The proportion of non-zero update coefficients in iterations, Indicates Each relay updates the mean value of the coefficients in a preset number of iterations.

[0091] For example, the continuous preset number of times is 5 times, when the update coefficient When it is zero, It may fall into the local optimal solution instead of the global optimal solution, which is not the ideal starting current setting value, and the convergence is weaker; when the update coefficient is non-zero, the update coefficient The smaller the value, the The closer the starting current setting value is to the one adjusted by the contribution weight of each relay, the closer it is to the ideal global optimal solution. The mean update coefficient of the iteration The smaller it is, the stronger the convergence.

[0092] In response to the convergence value being greater than a preset threshold or reaching a maximum number of iterations, the iteration is terminated, and the optimal configuration of the distribution network relay protection setting under given constraints and objectives is output.

[0093] Further explanation: the preset threshold is 0.8.

[0094] The present invention also provides a distribution network relay protection setting system based on multi-source data. Figure 2 As shown, the system includes a processor and a memory, the memory stores computer program instructions, and when the computer program instructions are executed by the processor, the distribution network relay protection setting method based on multi-source data according to the first aspect of the present invention is implemented.

[0095] The system also includes other components familiar to those skilled in the art, such as a communication bus and a communication interface. The configuration and functions of these components are known in the art and will not be described in detail here.

[0096] In the present invention, the aforementioned memory may be any tangible medium containing or storing a program that can be used by or in combination with an instruction execution system, apparatus or device. For example, a computer-readable storage medium may be any appropriate magnetic storage medium or magneto-optical storage medium, such as a resistive random access memory RRAM (Resistive Random Access Memory), a dynamic random access memory DRAM (Dynamic Random Access Memory), a static random access memory SRAM (Static Random-Access Memory), an enhanced dynamic random access memory EDRAM (Enhanced Dynamic Random Access Memory), a high-bandwidth memory HBM (High-Bandwidth Memory), a hybrid memory cube HMC (Hybrid Memory Cube), etc., or any other medium that can be used to store the required information and can be accessed by an application, a module, or both. Any such computer storage medium may be part of a device or accessible or connectable to a device. Any application or module described in the present invention may be implemented using computer-readable / executable instructions that may be stored or otherwise maintained by such a computer-readable medium.

[0097] In the description of this specification, "plurality" or "several" means at least two, such as two, three or more, etc., unless otherwise clearly and specifically defined.

[0098] Although this specification has shown and described a number of embodiments of the present invention, it will be apparent to those skilled in the art that such embodiments are provided by way of example only. Those skilled in the art will conceive of many modifications, changes and alternatives without departing from the ideas and spirit of the present invention. It should be understood that in the practice of the present invention, various alternatives to the embodiments of the present invention described herein may be employed.

Claims

1. A distribution network relay protection setting method based on multi-source data, characterized in that: include: Get the contribution weight of each relay, including: Define the target layer: evaluate the contribution of relays in the distribution network; define the criterion layer including: fault detection speed, fault isolation capability, false operation rate, and maintenance cost; use the 1-9 scaling method to construct a judgment matrix based on each criterion of the criterion layer, and perform eigenvalue decomposition on the judgment matrix to obtain the weight of each criterion layer; The weight of each criterion layer is multiplied and accumulated by the weight of each corresponding relay to obtain the contribution weight of the relay under the overall goal. The starting current setting value of each relay ; Based on the objective function in the gray wolf optimization algorithm, the minimum total protection action time is used as the objective function, the fitness of each relay is calculated, and the three optimal relays in the population are obtained. The direction and step size of the control search and the randomness of the search are updated based on the three optimal relays, and the positions of the remaining relays are adjusted based on the coordination vector. The coordination vector includes: and synergy vector , for the synergy vector Make dynamic adjustments; Preset initialization convergence factor and create 1 row A random matrix of columns, where The columns represent the dimensions of the relay individual position space, and the value range of each element in the matrix is ; Multiply each element in the random matrix by 2 to obtain an update matrix, the range of which is expanded to ; Create a matrix of the same size as the update matrix, with all elements set to 1, labeled as a constant matrix, and take the difference between the update matrix and the constant matrix to get a labeled matrix with element values ​​ranging from , multiply the label matrix by the convergence factor to obtain the synergy vector ; Adjustment based on the coordination vector Based on the positions of the remaining relays, the convergence factor is modified during the iteration process to obtain a dynamic step size factor, including: Take any of the remaining relays as the target relay and calculate the fitness of the target relay after iteration. , and the average fitness value of the target relay after iteration, calculate the ratio of the fitness value to the average fitness value, and perform normalization processing to analyze the level value of the relay for the average fitness of the population; Calculate the square of the absolute value of the spatial distance between the current position of the target relay and the optimal solution, and take the ratio of the square of the spatial size of the upper and lower limits of the current setting value as the total width of the search space; take twice the product of the horizontal value and the total width as the dynamic step factor; Any relay is a target relay, and the slope of the position between the target relay and the two adjacent relays on the left and right is calculated respectively to obtain the non-zero maximum slope, and the weighted sum of the starting current setting values ​​of all relays and the contribution weights is calculated, and the difference between the iterative position of the target relay on the right adjacent relay and the weighted sum is calculated and normalized, and the product of the maximum slope and the normalized difference is used as the update coefficient of the target relay iteration; the product of the dynamic step factor and the update coefficient is used as the adjustment search amount to guide the position update of the target relay; Based on the adjustment search amount, in the process of iterating for a preset number of consecutive times, the number of times the relay update coefficient is non-zero and the average value of the update coefficient are obtained to calculate the convergence value of each relay, and the convergence value satisfies the following relationship: ； In the formula, Indicates The convergence value of the relay, Indicates the number of times the update coefficient is not zero in the preset number of iterations, Indicates the preset number of iterations, Indicates continuous The proportion of non-zero update coefficients in iterations, Indicates The relay updates the mean value of the coefficient in the preset number of iterations; based on the convergence value, it is determined whether to terminate the iteration, and the position and optimal fitness of the optimal relay are output.

2. The distribution network relay protection setting method based on multi-source data according to claim 1 is characterized in that: The dynamic step factor also satisfies the following relationship: ； In the formula, Indicates Relay No. The dynamic step size factor after iterations, Indicates the current iteration number, Indicates the maximum number of iterations.

3. The distribution network relay protection setting method based on multi-source data according to claim 1 is characterized in that: The step of judging whether to terminate the iteration based on the convergence value includes: In response to the convergence value being greater than a preset threshold or reaching a maximum number of iterations, the iteration is terminated, and the optimal configuration of the distribution network relay protection setting under given constraints and objectives is output.

4. The distribution network relay protection setting system based on multi-source data is characterized by: include: A processor and a memory, wherein the memory stores computer program instructions, and when the computer program instructions are executed by the processor, a distribution network relay protection setting method based on multi-source data according to any one of claims 1 to 3 is implemented.

Images