A method and apparatus for deployment of vacuum interrupter for power distribution fault detection

Abstract

The application provides a vacuum quick-break device deployment method and device for power distribution fault detection. It belongs to the technical field of power distribution network applications. The method comprises the following steps: simulating different types of faults at different positions of a power distribution network simulation model, recording current data corresponding to all candidate deployment positions under fault simulation and corresponding fault types, and forming a data sample set; dividing the data sample set into a training set and an evaluation set; constructing a power distribution network fault detection model based on the training set, and obtaining the overall deployment fault identification accuracy; representing whether a vacuum quick-break device is deployed at each candidate deployment position as a deployment vector; based on a swarm intelligence optimization algorithm, taking the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, and the number of vacuum quick-break devices corresponding to each deployment vector as the joint optimization target, iteratively optimizing the deployment vector to obtain an optimal deployment scheme. The method can optimize device deployment and take into account detection performance and economy.

Description

Technical Field

[0001] This invention belongs to the field of power distribution network application technology, and in particular relates to a method and apparatus for deploying a vacuum fast-acting device for power distribution fault detection. Background Technology

[0002] In the field of power distribution network application technology, to improve power supply reliability, secondary detection equipment such as fault indicators and voltage and current sensors are typically deployed at critical nodes, or intelligent switching equipment with communication capabilities is installed at line segmentation points and branch points to monitor line status and locate faults. This type of deployment aims to build a monitoring network capable of sensing the power grid's operating status, providing a data foundation for subsequent fault analysis and handling.

[0003] However, in practical applications, the above deployment methods have obvious limitations. Sensors or monitoring equipment are usually deployed at local points along the line, with limited monitoring range, making it difficult to achieve comprehensive coverage of increasingly complex distribution networks, especially those with multiple branches and power sources. This point-based monitoring mode is prone to creating detection blind spots, meaning the system cannot effectively detect faults occurring in uncovered areas. On the other hand, increasing the number of deployment points to eliminate blind spots not only significantly increases the costs of equipment procurement, installation, and subsequent communication and maintenance, resulting in resource waste, but also leads to a sharp increase in the amount of information that needs to be processed. This places higher demands on data transmission bandwidth and the processing capacity of the master station system. During peak data concurrency, information congestion or processing delays can easily occur, affecting the timeliness of fault diagnosis.

[0004] To improve monitoring coverage, existing technologies consider deploying devices with rapid disconnection and intelligent detection capabilities at more critical nodes, such as branch lines and microgrid connection points. These devices typically integrate current and voltage sensing with fault identification algorithms, enabling rapid response to short-circuit faults. However, in practical implementation, scientifically determining the deployment location and quantity of these devices to achieve a balance between ensuring monitoring effectiveness and control costs remains a challenge. Existing deployment decisions often rely on engineering experience or configuration based on local needs, lacking a systematic and quantitative method to comprehensively evaluate overall fault detection capabilities and deployment economics. Therefore, in engineering practice, a trade-off often needs to be struck between "comprehensive coverage" and "economic feasibility," making it difficult to obtain a quantitatively validated optimal deployment strategy that balances performance and cost in the early stages of a project. Summary of the Invention

[0005] The purpose of this invention is to provide a method and apparatus for deploying vacuum fast-acting devices for power distribution fault detection. This method and apparatus can achieve scientific allocation and precise deployment of detection resources while ensuring the fault detection capability of the power distribution network. It can also achieve synergistic optimization between the number of devices deployed and the fault identification performance, thus providing effective support for the intelligent operation and maintenance and refined management of the power distribution network.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: In a first aspect, the present invention provides a method for deploying a vacuum instantaneous tripping device for power distribution fault detection, comprising: Obtain the line parameters, distributed power sources, and load data of a certain distribution network, determine all candidate deployment locations for deployable vacuum instantaneous tripping devices within the distribution network, and establish a distribution network operation simulation model; Based on the power distribution network operation simulation model, different types of faults are simulated at different locations in the power distribution network. The current data collected by the vacuum fast-acting devices at all candidate deployment locations and the corresponding fault types are recorded under each fault simulation to form a data sample set. Each data sample includes a data matrix composed of the current data at all candidate deployment locations and the corresponding fault type. The data sample set is preprocessed and divided into a training set and an evaluation set; A deep learning model is trained based on the training set, a power distribution network fault detection model is constructed, and the full deployment fault identification accuracy is obtained based on the evaluation set. Whether a vacuum instantaneous tripping device is deployed at each candidate deployment location is represented as a deployment vector, and multiple deployment vectors are initialized. Based on the swarm intelligence optimization algorithm, the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, as well as the number of vacuum fast-acting devices deployed corresponding to each deployment vector, are used as joint optimization objectives. Multiple deployment vectors are iteratively optimized until the iteration termination condition is met to obtain the optimal deployment scheme. The number and location of vacuum fast-acting devices in the distribution network are then determined based on the optimal deployment scheme.

[0007] Furthermore, the step of iteratively optimizing multiple deployment vectors based on the swarm intelligence optimization algorithm includes: S21. Determine the deployment vector as an individual in the swarm intelligence algorithm, and set the maximum number of iterations or the iteration termination condition and the swarm intelligence algorithm parameters. S22. Initialize multiple individuals according to a preset number of individuals, and determine the multiple individuals as the current individual set, and set the current iteration number to the initial value; S23. In the current iteration, for each individual in the current set of individuals, the current data matrix in the evaluation sample is retained or masked to obtain the processed current data matrix corresponding to that individual, and the processed current data matrix is ​​input into the power grid fault detection model to obtain the fault identification accuracy corresponding to that individual. S24. Based on the difference between the fault identification accuracy when all candidate deployment locations are equipped with vacuum instantaneous disconnection devices and the fault identification accuracy corresponding to the individual, and the number of vacuum instantaneous disconnection devices deployed for the individual, determine the objective function value corresponding to each individual in the current set of individuals, compare the objective function values ​​corresponding to each individual, update the current global optimum and the optimal individual corresponding to the current global optimum, wherein the objective function value is used to characterize the joint optimization result of the degree of fault identification accuracy decay and the number of vacuum instantaneous disconnection devices deployed; S25. Determine whether the maximum number of iterations or the iteration termination condition has been reached. If yes, determine the optimal individual as the optimal deployment scheme. If no, generate an equal number of new individuals from the current individual set according to the individual update and iteration method in the swarm intelligence algorithm. If any new individual does not meet the preset constraint conditions, re-randomly initialize the new individuals that do not meet the preset constraint conditions. S26. Determine the equal number of new individuals as the new current individual set, increment the iteration count by one, and return to step S23.

[0008] Furthermore, the step of forming the data sample set includes: S31. Determine the common fault types in the distribution network, assuming the number of fault types is... , of which Class-based faults are marked as fault labels. , ; S32. Randomly select one type of fault from the fault types, and simulate the selected fault at a random location in the power distribution network operation simulation model, and record the current signals collected by all vacuum fast-acting devices after the fault occurs. S33. Assume the number of vacuum instantaneous tripping devices is... The one, in the first In the simulation, the first The current signal collected by the vacuum instantaneous tripping device is represented as follows:

[0009] in, In the In this simulation The current signals collected by the vacuum instantaneous trip device constitute a current data matrix:

[0010]

[0011] in, Indicates the first The simulation in the second year Current signal collected by a vacuum instantaneous trip device Indicates the first The simulation in the second year A vacuum instantaneous tripping device at a certain time The collected three-phase current signals, , and Respectively represent the first A vacuum instantaneous tripping device at a certain time The collected phase A current, phase B current, and phase C current, Indicates the first The data matrix is ​​composed of the current signals collected by all vacuum instantaneous trip devices in this simulation. Indicates the time when the fault occurred. This indicates the duration of time for the vacuum instantaneous tripping device to acquire the current signal. , Indicates the number of simulations. Indicates by lines and A real matrix composed of columns; S34, the first The data matrix obtained from the simulation Fault label corresponding to this simulation When a data sample is added to the data sample set, The preset total number of simulations has not been reached. When that happens, return to step S32 until... This forms a data sample set:

[0012] in, Represents a set of data samples. Indicates the first The fault label corresponding to each simulation. This represents the total number of simulations.

[0013] Furthermore, the step of preserving or masking the current data matrix in the evaluation sample to obtain the processed current data matrix corresponding to that individual includes: Represent each individual in the current set of individuals as a deployment vector. :

[0014] in, Indicates the first Vacuum fast-break devices are deployed at each of the candidate deployment locations. Indicates the first Vacuum instantaneous trip device will not be deployed at any of the candidate deployment locations. , Indicates the number of candidate deployment locations; The deployment vector Expand by a factor of 3 in row format to form an expanded vector. :

[0015] Then extend the vector Expand in column form Multiply by 1 to form an extended matrix :

[0016] in, Indicates the duration of current signal acquisition by the vacuum instantaneous tripping device; The extended matrix A matrix dot product is performed on the current data matrix corresponding to each evaluation sample in the evaluation set to preserve or mask the current data matrix, thereby obtaining the processed current data matrix corresponding to the individual. The step of inputting the processed current data matrix into the distribution network fault detection model to obtain the fault identification accuracy corresponding to that individual includes: The processed current data matrix is ​​input into the power distribution network fault detection model, and the corresponding fault type is output. Based on the fault types output by the power distribution network fault detection model and the fault types corresponding to each evaluation sample, the fault identification accuracy corresponding to the individual is obtained.

[0017] Furthermore, the joint optimization objective is determined in the following manner: The objective function for optimizing the deployment of the vacuum instantaneous tripping device is defined as follows:

[0018] in, This represents the objective function to be optimized. This represents the fault identification accuracy when vacuum instantaneous tripping devices are deployed at all candidate deployment locations. This represents the fault identification accuracy corresponding to the individual. For the preset weighting coefficients, and ; Constraints are set according to power grid requirements or the current number of devices. These constraints include at least the following: , in, This indicates the minimum number of vacuum instantaneous shut-off devices that can be deployed at a time.

[0019] Furthermore, the swarm intelligence algorithm is a genetic algorithm, and the parameters of the genetic algorithm include at least the number of individuals in the population, selection rate, crossover rate, mutation rate, and re-insertion rate. The method of generating an equal number of new individuals based on the update and iteration method of individuals in the swarm intelligence algorithm includes generating an equal number of new individuals using selection, crossover, mutation, and re-insertion operations.

[0020] Furthermore, the power distribution network fault detection model is a long short-term time series network, which includes: a convolutional layer for extracting local time-series features from a data matrix composed of current data corresponding to all candidate deployment locations; a recurrent layer for performing time-series correlation modeling on the local time-series features extracted by the convolutional layer; and a fully connected layer for mapping the features processed by the recurrent layer and outputting the fault type.

[0021] In a second aspect, the present invention provides a vacuum instantaneous tripping device deployment apparatus for power distribution fault detection, comprising: The simulation modeling module is used to acquire line parameters, distributed power sources, and load data of a power distribution network, determine the candidate deployment locations of all deployable vacuum instantaneous tripping devices in the power distribution network, and establish a power distribution network operation simulation model. The sample construction module is used to simulate different types of faults at different locations in the distribution network based on the distribution network operation simulation model. It records the current data collected by the vacuum fast-acting devices corresponding to all candidate deployment locations and the corresponding fault types under each fault simulation, forming a data sample set. Each data sample includes a data matrix composed of the current data corresponding to all candidate deployment locations and the corresponding fault type. The data preprocessing module is used to preprocess the data sample set and divide it into a training set and an evaluation set. The fault detection model construction module is used to train a deep learning model based on the training set, construct a power distribution network fault detection model, and obtain the full deployment fault identification accuracy based on the evaluation set. The deployment vector initialization module is used to characterize whether a vacuum instantaneous tripping device is deployed at each candidate deployment location as a deployment vector, and to initialize multiple deployment vectors; The iterative optimization module is used to perform iterative optimization on multiple deployment vectors based on the swarm intelligence optimization algorithm, taking the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, and the number of vacuum fast-acting devices deployed corresponding to each deployment vector as joint optimization objectives, until the iteration termination condition is met, to obtain the optimal deployment scheme, and to determine the number and location of vacuum fast-acting devices in the distribution network based on the optimal deployment scheme.

[0022] In a third aspect, the present invention provides an electronic device including a processor and a memory, the processor being configured to execute a computer program stored in the memory to implement a method for deploying a vacuum fast-break device for power distribution fault detection.

[0023] In a fourth aspect, the present invention provides a computer-readable storage medium storing at least one instruction that, when executed by a processor, implements a method for deploying a vacuum fast-break device for power distribution fault detection.

[0024] Compared with the prior art, the beneficial effects of the present invention are as follows: 1. The present invention provides a method for deploying vacuum instantaneous tripping devices for power distribution fault detection. By establishing a power distribution network operation simulation model, different types of faults are simulated at different locations, a data sample set is constructed and the power distribution network fault detection model is trained. On this basis, a swarm intelligence optimization algorithm is combined to iteratively optimize the deployment quantity and deployment location of vacuum instantaneous tripping devices. This enables the optimized determination of the deployment scheme of vacuum instantaneous tripping devices, thereby ensuring the fault detection capability of the power distribution network while taking into account the reasonable configuration of the deployment quantity and deployment location of the devices, and improving the pertinence and economy of the deployment scheme design.

[0025] 2. This invention represents the deployment status of candidate deployment locations as deployment vectors and further expands them into an extended matrix corresponding to the dimension of the current data matrix. This allows for the retention or masking of the current data matrix in the evaluation sample, thereby obtaining the fault identification accuracy corresponding to each deployment scheme. Simultaneously, by constructing a joint optimization objective that includes the difference between the overall deployment fault identification accuracy and the corresponding deployment scheme's fault identification accuracy, as well as the number of vacuum fast-acting devices deployed, the invention can quantify the retention status of measurement points under different deployment schemes and their impact on fault identification results. This enables a comprehensive evaluation of the detection performance and deployment cost of deployment schemes, improving the computability, comparability, and accuracy of the deployment optimization process.

[0026] 3. This invention constructs a power distribution network fault detection model by employing a long short-term time series network. It utilizes convolutional layers to extract local temporal features, recurrent layers to extract temporal correlation features, and fully connected layers to output fault types. Furthermore, it combines selection, crossover, mutation, and re-insertion operations from a genetic algorithm to iteratively update the deployment vector. This improves the effectiveness of fault identification under different fault scenarios and enhances the efficiency and rationality of searching for the optimal deployment scheme. Thus, it obtains the optimal deployment scheme that balances fault detection performance and device configuration efficiency while meeting the constraint of the minimum number of deployments. Attached Figure Description

[0027] The accompanying drawings, which form part of this application, are used to provide a further understanding of the invention. The illustrative embodiments of the invention and their descriptions are used to explain the invention and do not constitute an undue limitation of the invention. In the drawings: Figure 1 This is a flowchart of a method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to an embodiment of the present invention; Figure 2 This is a schematic diagram of a power distribution network structure containing multiple primary and secondary integrated vacuum fast-break devices according to an embodiment of the present invention; Figure 3 This is a schematic diagram of a power distribution network fault containing multiple primary and secondary integrated vacuum fast-break devices according to an embodiment of the present invention; Figure 4 This is a schematic diagram of a power distribution network fault detection model based on a primary and secondary fusion vacuum fast-break device using a long-short-terminal time series network, according to an embodiment of the present invention. Figure 5 This is a flowchart illustrating the optimization model for deploying a primary and secondary fusion vacuum fast-break device using a swarm intelligence optimization algorithm, as described in an embodiment of the present invention. Figure 6 This is a schematic diagram of a vacuum instantaneous tripping device deployment device for power distribution fault detection according to an embodiment of the present invention; Figure 7 This is a structural block diagram of an electronic device according to an embodiment of the present invention. Detailed Implementation

[0028] The present invention will now be described in detail with reference to the accompanying drawings and embodiments. It should be noted that, unless otherwise specified, the embodiments and features described in this application can be combined with each other.

[0029] The following detailed description is exemplary and intended to provide further detailed explanation of the invention. Unless otherwise specified, all technical terms used in this invention have the same meaning as commonly understood by one of ordinary skill in the art to which this application pertains. The terminology used in this invention is for the purpose of describing particular embodiments only and is not intended to limit the scope of exemplary embodiments according to the invention.

[0030] Example 1 This embodiment provides a method for deploying a vacuum instantaneous trip device for power distribution fault detection. Taking a power distribution network structure containing multiple transformer substations and multiple vacuum instantaneous trip devices as an example, the implementation process of the method is described. In this embodiment, a primary and secondary integrated vacuum instantaneous trip device is used as an example; the primary and secondary integrated vacuum instantaneous trip devices mentioned in the disclosure and accompanying drawings are all described as specific examples of vacuum instantaneous trip devices or instantaneous trip devices.

[0031] like Figures 1 to 5 As shown, Figure 1 The overall flow of this embodiment is shown, mainly including steps S1 to S6. Figure 2 A schematic diagram of a distribution network structure containing multiple vacuum instantaneous trip devices is shown. In this diagram, Ti represents a transformer, CB represents a circuit breaker, and SDi represents a vacuum instantaneous trip device. The diagram shows the locations of distribution areas A to F, transformers T1 to T7, circuit breakers CB, and lines numbered #1 to #14, along with the corresponding candidate deployment locations of the vacuum instantaneous trip devices. Figure 3 A schematic diagram is shown when a fault Li occurs at a random location in the power distribution network. Figure 4 The structure of a distribution network fault detection model based on a long short-term time series network is shown, which uses a matrix of current signals collected by M vacuum fast-acting devices as input. Figure 5 The iterative process of solving the optimization model for the deployment of a vacuum instantaneous tripping device using a swarm intelligence optimization algorithm is shown.

[0032] like Figure 1 As shown, step S1 involves acquiring the line parameters, distributed generation sources, and load data of a distribution network, determining all candidate deployment locations for deployable vacuum instantaneous tripping devices within the distribution network, and establishing a distribution network operation simulation model. In this embodiment, basic information such as the structure and line parameters of a distribution network, the access locations and installed capacity of distributed generation sources, and user load data are first acquired.

[0033] Then, based on the actual conditions of the distribution network, determine the locations of all suitable vacuum instantaneous trip devices within the distribution network. Figure 2 Taking the distribution network structure shown as an example, the distribution network includes multiple low-voltage substations, and multiple locations where vacuum instantaneous tripping devices can be deployed are set along the feeder. In the figure, SD1, SD2, ..., SD13 correspond to the candidate deployment locations of the vacuum instantaneous tripping devices.

[0034] Finally, based on the obtained line parameters, distributed power sources, and load data, and combined with the determined candidate deployment locations, a power distribution network operation simulation model is constructed using simulation software. This provides a unified simulation environment for subsequent random fault simulation, data sample construction, fault detection model training, and deployment optimization solutions.

[0035] Step S2 involves simulating different types of faults at different locations within the distribution network based on the aforementioned distribution network operation simulation model. The current data collected by the instantaneous tripping devices at all candidate deployment locations and the corresponding fault types are recorded for each fault simulation, forming a data sample set. Each data sample includes a data matrix composed of the current data from all candidate deployment locations and the corresponding fault type. Specifically, step S2 may further include steps S31 to S34.

[0036] Step S31: Determine the fault types in the distribution network. Let the number of fault types be N, where the i-th type of fault is marked as a fault tag. , In this embodiment, the fault types may include single-phase grounding fault, three-phase grounding fault, single-phase open-circuit fault, three-phase open-circuit fault, metallic grounding fault, non-metallic grounding fault, transformer fault, overload fault, etc.

[0037] Step S32: Randomly select one type of fault from the fault types, and simulate the selected fault at a random location in the power distribution network operation simulation model, recording the current signals collected by all vacuum instantaneous trip devices after the fault occurs. For example... Figure 3 As shown, the fault It can randomly appear at different locations on the feeder. By continuously changing the fault type and fault location in the simulation model, current response data under various fault scenarios can be constructed.

[0038] Step S33, assuming the number of vacuum instantaneous trip devices is M, the current signal collected by the j-th vacuum instantaneous trip device in the k-th simulation is expressed as:

[0039] in, In the k-th simulation, the current signals collected by the M vacuum instantaneous trip devices constitute a current data matrix:

[0040]

[0041] in, This represents the current signal collected by the j-th vacuum instantaneous trip device in the k-th simulation. This indicates that in the k-th simulation, the j-th vacuum instantaneous tripping device is at time [time value missing]. The collected three-phase current signals, , and These represent the times of the j-th vacuum instantaneous tripping device. The collected phase A current, phase B current and phase C current, This represents the data matrix composed of the current signals collected by all vacuum instantaneous trip devices in the k-th simulation. The time indicated by T represents the moment the fault occurred, and the time duration for the vacuum instantaneous trip device to collect the current signal is also indicated. k represents the number of simulations. This represents a real matrix consisting of 3M rows and T+1 columns.

[0042] Step S34, the data matrix obtained from the k-th simulation Fault label corresponding to this simulation A data sample is created and added to the data sample set; if k does not reach the preset total number of simulations K, return to step S32, until... This forms a data sample set:

[0043] Where A represents the data sample set, This represents the fault label corresponding to the k-th simulation, where K represents the total number of simulations. Therefore, each data sample consists of a current data matrix and its corresponding fault type.

[0044] Step S3 involves preprocessing the data sample set and dividing it into a training set and an evaluation set. In this embodiment, the data samples in the data sample set are preprocessed, which includes at least standardization, label encoding, and dataset partitioning. Specifically, the current data matrix obtained in step S2 is standardized to eliminate the influence of different dimensions and amplitude ranges on the training of the deep learning model; fault labels are then processed. Labels are encoded to form a fault type representation suitable for model output; and the preprocessed data samples are divided into a training set and an evaluation set according to a preset ratio. Since this embodiment uses the terminology of training set and evaluation set, subsequent model training and accuracy evaluation are based on the training set and evaluation set, respectively.

[0045] Step S4: Train a deep learning model based on the training set to construct a power distribution network fault detection model, and obtain the full deployment fault identification accuracy based on the evaluation set. Specifically, using the preprocessed data from step S3 as input, define the model training elements such as the number of hidden layers and neurons, the number of neurons in the input and output layers, activation functions, loss functions, optimizers, and hyperparameters. The output layer represents the fault labels. The encoding, the input layer is the matrix obtained after normalization in step S3. The new matrix, the matrix dimension is still . The power distribution network fault detection model is a long short-term time series network, which includes: a convolutional layer for extracting local time-series features from a data matrix composed of current data corresponding to all candidate deployment locations; a recurrent layer for performing time-series correlation modeling on the local time-series features extracted by the convolutional layer; and a fully connected layer for mapping the features processed by the recurrent layer and outputting the fault type. Figure 4 The input-output structure of the long-short-term time series network is illustrated schematically, wherein the current signals collected by M vacuum fast-acting devices form a matrix. As a time series input, after processing through convolutional layers, recurrent layers, and fully connected layers, the fault type is output. Based on this, the evaluation set is input into the power distribution network fault detection model to obtain the fault identification accuracy rate when all candidate deployment locations are equipped with instantaneous tripping devices, denoted as Yacc, where Yacc represents the fault identification accuracy rate when all candidate deployment locations are equipped with instantaneous tripping devices.

[0046] Step S5: Represent whether a fast-break device is deployed at each candidate deployment location as a deployment vector, and initialize multiple deployment vectors. In this embodiment, the optimization variable corresponding to each candidate deployment location is defined as a binary variable of 0 or 1, and the number of variables is equal to the number of candidate deployment locations. These optimization variables constitute the deployment vectors.

[0047] in, This indicates that a vacuum instantaneous trip device is deployed at the j-th candidate deployment location. This indicates that the vacuum instantaneous trip device will not be deployed at the j-th candidate deployment location. M represents the number of candidate deployment locations. In the subsequent swarm intelligence optimization process, the deployment vector X is determined as an individual in the swarm intelligence algorithm, and multiple individuals are initialized according to a preset number of individuals. These multiple individuals are determined as the current individual set, and the current iteration number is set to the initial value. Thus, whether a candidate deployment location is deployed or not is uniformly mapped to a deployment vector and individual set form that can be used for optimization iteration.

[0048] Step S6: Based on a swarm intelligence optimization algorithm, using the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, and the number of fast-acting devices deployed corresponding to each deployment vector as joint optimization objectives, iteratively optimizes multiple deployment vectors until the iteration termination condition is met to obtain the optimal deployment scheme. Then, based on the optimal deployment scheme, the number and location of fast-acting devices deployed in the distribution network are determined. Specifically, as shown... Figure 5 As shown, step S6 may further include steps S21 to S26.

[0049] Step S21: The deployment vector is determined as an individual in the swarm intelligence algorithm, and the maximum number of iterations or the iteration termination condition and the swarm intelligence algorithm parameters are set. In this embodiment, the swarm intelligence algorithm is a genetic algorithm, and the parameters of the genetic algorithm include at least the number of individuals in the population, selection rate, crossover rate, mutation rate, and re-insertion rate.

[0050] Step S22: Initialize multiple individuals according to a preset number of individuals, and determine the multiple individuals as the current individual set, and set the current iteration number to the initial value.

[0051] Step S23: In the current iteration, for each individual in the current individual set, the current data matrix in the evaluation sample is retained or masked to obtain the processed current data matrix corresponding to that individual. The processed current data matrix is ​​then input into the distribution network fault detection model to obtain the fault identification accuracy corresponding to that individual. Specifically, the process of retaining or masking the current data matrix in the evaluation sample to obtain the processed current data matrix corresponding to that individual, and inputting the processed current data matrix into the distribution network fault detection model to obtain the fault identification accuracy corresponding to that individual, includes: representing each individual in the current individual set as a deployment vector X; and expanding the deployment vector X by a factor of 3 in row-wise form to form an expanded vector.

[0052] Then extend the vector Expand the columns by a factor of T+1 to form an extended matrix. The format is as follows:

[0053] Where T represents the time length for the vacuum instantaneous tripping device to acquire the current signal. The extended matrix... A matrix dot product is performed on the current data matrix corresponding to each evaluation sample in the evaluation set to preserve or mask the current data matrix, obtaining a processed current data matrix corresponding to the individual. Subsequently, the processed current data matrix is ​​input into the distribution network fault detection model, which outputs the corresponding fault type. Then, based on the fault type output by the distribution network fault detection model and the fault type corresponding to each evaluation sample, the fault identification accuracy corresponding to the individual is obtained, denoted as Zacc, where Zacc represents the fault identification accuracy corresponding to the individual, i.e., within the distribution network. The accuracy of fault detection and identification in the distribution network after deploying vacuum fast-acting devices at the designated locations is assessed. In this embodiment, the aforementioned retention or shielding processing and fault identification accuracy statistics are all based on the evaluation samples in the evaluation set.

[0054] Step S24: Based on the difference between the fault identification accuracy when all candidate deployment locations are equipped with instantaneous tripping devices and the fault identification accuracy corresponding to the individual, and the number of instantaneous tripping devices deployed for the individual, determine the objective function value corresponding to each individual in the current set of individuals, compare the objective function values ​​corresponding to each individual, and update the current global optimum and the optimal individual corresponding to the current global optimum. Further, the joint optimization objective is determined as follows: the optimization objective function for the deployment of vacuum instantaneous tripping devices is defined as...

[0055] Where J represents the objective function, Yacc represents the fault identification accuracy when all candidate deployment locations are equipped with instantaneous tripping devices, and Zacc represents the fault identification accuracy corresponding to the individual. For the preset weighting coefficients, and The above optimization objective function considers two factors simultaneously: one is through... On the one hand, it characterizes the degree of decrease in fault identification accuracy after deployment reduction, and on the other hand, it uses... This characterizes the proportion of deployed instantaneous tripping devices, aiming to minimize the number of devices deployed while ensuring fault detection capabilities. Constraints are set based on grid requirements or the current number of devices, and these constraints include at least:

[0056] Where C represents the preset minimum deployment number of vacuum instantaneous tripping devices. In other words, during the optimization process, the generated individuals must not only satisfy the objective function to be as small as possible, but also meet the constraint that the number of instantaneous tripping devices deployed is not less than the preset minimum deployment number.

[0057] Step S25: Determine whether the maximum number of iterations or the iteration termination condition has been reached; if yes, then the optimal individual is determined as the optimal deployment scheme; if not, then according to the update and iteration method of individuals in the swarm intelligence algorithm, an equal number of new individuals are generated from the current individual set, and if any new individual does not meet the preset constraints, the new individuals that do not meet the preset constraints are re-randomly initialized. Specifically, generating an equal number of new individuals according to the update and iteration method of individuals in the swarm intelligence algorithm includes generating an equal number of new individuals using selection, crossover, mutation, and re-insertion operations. Combined with... Figure 5The process shown is as follows: In the current iteration, the objective function value corresponding to each individual in the current individual set is calculated according to the above-mentioned optimization objective function, with each individual corresponding to one objective function value; the objective function values ​​corresponding to all individuals are compared, and the minimum value is selected as the current global optimum, and the individual Xbest corresponding to the current global optimum is recorded; then, operations such as selection, crossover, mutation, and re-insertion are used to generate an equal number of new individuals based on the current individual set; if the new individuals cannot satisfy the aforementioned constraints, new individuals that do not satisfy the constraints are randomly initialized again; after that, the objective function value corresponding to the generated new individuals is calculated again according to the above-mentioned optimization objective function, and it is determined whether it is less than the current global optimum. If it is less than the current global optimum, the current global optimum and the optimal individual Xbest corresponding to the current global optimum are updated; if both are greater than the current global optimum, the current global optimum and its corresponding optimal individual remain unchanged.

[0058] Step S26: Determine the new set of individuals of equal quantity as the new current set, increment the iteration count by one, and return to step S23. If the maximum number of iterations is reached or the iteration termination condition is met, then the individual Xbest corresponding to the optimal value is the optimal deployment scheme. Thus, the optimal deployment scheme corresponds to the candidate deployment positions in vector Xbest where the element equals 1, i.e., the actual deployment positions where the vacuum instantaneous tripping device needs to be deployed, while the candidate deployment positions in vector Xbest where the element equals 0 will not have the vacuum instantaneous tripping device deployed.

[0059] The vacuum instantaneous tripping device deployment method for power distribution fault detection provided in this embodiment is based on a power distribution network operation simulation model. By simulating different types of faults at different locations, a data sample set containing current data and fault types for all candidate deployment locations is constructed. Based on this, a power distribution network fault detection model is trained, thus establishing a correspondence between deployment schemes and fault detection capabilities. Furthermore, by representing the deployment status of candidate deployment locations as deployment vectors and using an extended matrix to retain or mask the current data matrix in the evaluation samples, the method accurately reflects the retention of measurement point information under different deployment schemes and its impact on fault identification results. On this basis, the difference between the overall deployment fault identification accuracy and the fault identification accuracy under different deployment schemes, along with the number of vacuum instantaneous tripping devices deployed, constitutes a joint optimization objective. This objective is then iteratively solved using a genetic algorithm, enabling the search for the optimal deployment scheme while satisfying the minimum deployment quantity constraint. Therefore, this embodiment not only improves the pertinence, computability, and rationality of the vacuum instantaneous tripping device deployment scheme design but also reduces the number of devices deployed, lowers deployment redundancy, and improves resource allocation efficiency while ensuring power distribution network fault detection capabilities. This achieves a comprehensive balance and optimization between power distribution network fault detection performance and device deployment costs.

[0060] Example 2 like Figure 6 As shown, based on the same inventive concept as the above embodiments, the present invention also provides a vacuum instantaneous tripping device deployment device for power distribution fault detection, comprising: The simulation modeling module is used to acquire line parameters, distributed power sources, and load data of a power distribution network, determine the candidate deployment locations of all deployable vacuum instantaneous tripping devices in the power distribution network, and establish a power distribution network operation simulation model. The sample construction module is used to simulate different types of faults at different locations in the distribution network based on the distribution network operation simulation model. It records the current data collected by the vacuum fast-acting devices corresponding to all candidate deployment locations and the corresponding fault types under each fault simulation, forming a data sample set. Each data sample includes a data matrix composed of the current data corresponding to all candidate deployment locations and the corresponding fault type. The data preprocessing module is used to preprocess the data sample set and divide it into a training set and an evaluation set. The fault detection model construction module is used to train a deep learning model based on the training set, construct a power distribution network fault detection model, and obtain the full deployment fault identification accuracy based on the evaluation set. The deployment vector initialization module is used to characterize whether a vacuum instantaneous tripping device is deployed at each candidate deployment location as a deployment vector, and to initialize multiple deployment vectors; The iterative optimization module is used to perform iterative optimization on multiple deployment vectors based on the swarm intelligence optimization algorithm, taking the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, and the number of vacuum fast-acting devices deployed corresponding to each deployment vector as joint optimization objectives, until the iteration termination condition is met, to obtain the optimal deployment scheme, and to determine the number and location of vacuum fast-acting devices in the distribution network based on the optimal deployment scheme.

[0061] Example 3 like Figure 7 As shown, the present invention also provides an electronic device 100 for implementing the aforementioned method for deploying a vacuum fast-break device for power distribution fault detection; The electronic device 100 includes a memory 101, at least one processor 102, a computer program 103 stored in the memory 101 and executable on at least one processor 102, and at least one communication bus 104.

[0062] The memory 101 can be used to store the computer program 103. The processor 102 implements the vacuum fast-break device deployment method for power distribution fault detection in Embodiment 1 by running or executing the computer program stored in the memory 101 and calling the data stored in the memory 101.

[0063] The memory 101 may primarily include a program storage area and a data storage area. The program storage area may store the operating system, application programs required for at least one function (such as sound playback function, image playback function, etc.), etc.; the data storage area may store data created based on the use of the electronic device 100 (such as audio data), etc. In addition, the memory 101 may include non-volatile memory, such as hard disk, RAM, plug-in hard disk, smart media card (SMC), secure digital (SD) card, flash card, at least one disk storage device, flash memory device, or other non-volatile solid-state storage device.

[0064] At least one processor 102 may be a Central Processing Unit (CPU), or other general-purpose processors, digital signal processors (DSPs), application-specific integrated circuits (ASICs), field-programmable gate arrays (FPGAs), or other programmable logic devices, discrete gate or transistor logic devices, discrete hardware components, etc. Processor 102 may be a microprocessor or any conventional processor. Processor 102 is the control center of electronic device 100, connecting various parts of electronic device 100 via various interfaces and lines.

[0065] The memory 101 in the electronic device 100 stores multiple instructions to implement a method for deploying a vacuum instantaneous tripping device for power distribution fault detection, and the processor 102 can execute multiple instructions to achieve the following: Obtain the line parameters, distributed power sources, and load data of a certain distribution network, determine all candidate deployment locations for deployable vacuum instantaneous tripping devices within the distribution network, and establish a distribution network operation simulation model; Based on the power distribution network operation simulation model, different types of faults are simulated at different locations in the power distribution network. The current data collected by the vacuum fast-acting devices at all candidate deployment locations and the corresponding fault types are recorded under each fault simulation to form a data sample set. Each data sample includes a data matrix composed of the current data at all candidate deployment locations and the corresponding fault type. The data sample set is preprocessed and divided into a training set and an evaluation set; A deep learning model is trained based on the training set, a power distribution network fault detection model is constructed, and the full deployment fault identification accuracy is obtained based on the evaluation set. Whether a vacuum instantaneous tripping device is deployed at each candidate deployment location is represented as a deployment vector, and multiple deployment vectors are initialized. Based on the swarm intelligence optimization algorithm, the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, as well as the number of vacuum fast-acting devices deployed corresponding to each deployment vector, are used as joint optimization objectives. Multiple deployment vectors are iteratively optimized until the iteration termination condition is met to obtain the optimal deployment scheme. The number and location of vacuum fast-acting devices in the distribution network are then determined based on the optimal deployment scheme.

[0066] Example 4 If the modules / units integrated in the electronic device 100 are implemented as software functional units and sold or used as independent products, they can be stored in a computer-readable storage medium. Based on this understanding, all or part of the processes in the methods of the above embodiments of the present invention can also be implemented by a computer program instructing related hardware. The computer program can be stored in a computer-readable storage medium, and when executed by a processor, it can implement the steps of the various method embodiments described above. The computer program includes computer program code, which can be in the form of source code, object code, executable files, or certain intermediate forms. The computer-readable medium can include: any entity or device capable of carrying computer program code, recording media, USB flash drives, portable hard drives, magnetic disks, optical disks, computer memory, and read-only memory (ROM).

[0067] Those skilled in the art will understand that embodiments of the present invention can be provided as methods, systems, or computer program products. Therefore, the present invention can take the form of a completely hardware embodiment, a completely software embodiment, or an embodiment combining software and hardware aspects. Furthermore, the present invention can take the form of a computer program product embodied on one or more computer-usable storage media (including, but not limited to, disk storage, CD-ROM, optical storage, etc.) containing computer-usable program code.

[0068] This invention is described with reference to flowchart illustrations and / or block diagrams of methods, apparatus (systems), and computer program products according to embodiments of the invention. It will be understood that each block of the flowchart illustrations and / or block diagrams, and combinations of blocks in the flowchart illustrations and / or block diagrams, can be implemented by computer program instructions. These computer program instructions can be provided to a processor of a general-purpose computer, special-purpose computer, embedded processor, or other programmable data processing apparatus to produce a machine, such that the instructions, which execute via the processor of the computer or other programmable data processing apparatus, generate instructions for implementing the flowchart illustrations and / or block diagrams. Figure 1 One or more processes and / or boxes Figure 1A device that provides the functions specified in one or more boxes.

[0069] These computer program instructions may also be stored in a computer-readable storage medium that can direct a computer or other programmable data processing device to function in a particular manner, such that the instructions stored in the computer-readable storage medium produce an article of manufacture including instruction means, which are implemented in a process Figure 1 One or more processes and / or boxes Figure 1 The function specified in one or more boxes.

[0070] These computer program instructions may also be loaded onto a computer or other programmable data processing equipment to cause a series of operational steps to be performed on the computer or other programmable equipment to produce a computer-implemented process, thereby providing instructions that execute on the computer or other programmable equipment for implementing the process. Figure 1 One or more processes and / or boxes Figure 1 The steps of the function specified in one or more boxes.

[0071] In the description of this specification, references to terms such as "an embodiment," "example," "specific example," etc., indicate that a specific feature, structure, material, or characteristic described in connection with that embodiment or example is included in at least one embodiment or example of the invention. In this specification, illustrative expressions of the above terms do not necessarily refer to the same embodiment or example. Furthermore, the specific features, structures, materials, or characteristics described may be combined in any suitable manner in one or more embodiments or examples.

[0072] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and not to limit it. Although the present invention has been described in detail with reference to the above embodiments, those skilled in the art should understand that modifications or equivalent substitutions can still be made to the specific implementation of the present invention. Any modifications or equivalent substitutions that do not depart from the spirit and scope of the present invention should be covered within the scope of protection of the claims of the present invention.

Claims

1. A method for deploying a vacuum instantaneous tripping device for power distribution fault detection, characterized in that, include: Obtain the line parameters, distributed power sources, and load data of a certain distribution network, determine all candidate deployment locations for deployable vacuum instantaneous tripping devices within the distribution network, and establish a distribution network operation simulation model; Based on the power distribution network operation simulation model, different types of faults are simulated at different locations in the power distribution network. The current data collected by the vacuum fast-acting devices at all candidate deployment locations and the corresponding fault types are recorded under each fault simulation to form a data sample set. Each data sample includes a data matrix composed of the current data at all candidate deployment locations and the corresponding fault type. The data sample set is preprocessed and divided into a training set and an evaluation set; A deep learning model is trained based on the training set, a power distribution network fault detection model is constructed, and the full deployment fault identification accuracy is obtained based on the evaluation set. Whether a vacuum instantaneous tripping device is deployed at each candidate deployment location is represented as a deployment vector, and multiple deployment vectors are initialized. Based on the swarm intelligence optimization algorithm, the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, as well as the number of vacuum fast-acting devices deployed corresponding to each deployment vector, are used as joint optimization objectives. Multiple deployment vectors are iteratively optimized until the iteration termination condition is met to obtain the optimal deployment scheme. The number and location of vacuum fast-acting devices in the distribution network are then determined based on the optimal deployment scheme.

2. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 1, characterized in that, The steps for iteratively optimizing multiple deployment vectors based on the swarm intelligence optimization algorithm include: S21. Determine the deployment vector as an individual in the swarm intelligence algorithm, and set the maximum number of iterations or the iteration termination condition and the swarm intelligence algorithm parameters. S22. Initialize multiple individuals according to a preset number of individuals, and determine the multiple individuals as the current individual set, and set the current iteration number to the initial value; S23. In the current iteration, for each individual in the current set of individuals, the current data matrix in the evaluation sample is retained or masked to obtain the processed current data matrix corresponding to that individual, and the processed current data matrix is ​​input into the power grid fault detection model to obtain the fault identification accuracy corresponding to that individual. S24. Based on the difference between the fault identification accuracy when all candidate deployment locations are equipped with vacuum instantaneous disconnection devices and the fault identification accuracy corresponding to the individual, and the number of vacuum instantaneous disconnection devices deployed for the individual, determine the objective function value corresponding to each individual in the current set of individuals, compare the objective function values ​​corresponding to each individual, update the current global optimum and the optimal individual corresponding to the current global optimum, wherein the objective function value is used to characterize the joint optimization result of the degree of fault identification accuracy decay and the number of vacuum instantaneous disconnection devices deployed; S25. Determine whether the maximum number of iterations or the iteration termination condition has been reached. If yes, determine the optimal individual as the optimal deployment scheme. If no, generate an equal number of new individuals from the current individual set according to the individual update and iteration method in the swarm intelligence algorithm. If any new individual does not meet the preset constraint conditions, re-randomly initialize the new individuals that do not meet the preset constraint conditions. S26. Determine the equal number of new individuals as the new current individual set, increment the iteration count by one, and return to step S23.

3. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 2, characterized in that, The step of forming a data sample set includes: S31, determining common fault types in the distribution network, assuming the number of fault types is... , of which Class-based faults are marked as fault labels. , ; S32. Randomly select one type of fault from the fault types, and simulate the selected fault at a random location in the power distribution network operation simulation model, and record the current signals collected by all vacuum fast-acting devices after the fault occurs. S33. Assume the number of vacuum instantaneous tripping devices is... The one, in the first In the simulation, the first The current signal collected by the vacuum instantaneous tripping device is represented as follows: in, In the In this simulation The current signals collected by the vacuum instantaneous trip device constitute a current data matrix: in, Indicates the first The simulation in the second year Current signal collected by a vacuum instantaneous trip device Indicates the first The simulation in the second year A vacuum instantaneous tripping device at a certain time The collected three-phase current signals, , and Respectively represent the first A vacuum instantaneous tripping device at a certain time The collected phase A current, phase B current, and phase C current, Indicates the first The data matrix is ​​composed of the current signals collected by all vacuum instantaneous trip devices in this simulation. Indicates the time when the fault occurred. This indicates the duration of time for the vacuum instantaneous tripping device to acquire the current signal. , Indicates the number of simulations. Indicates by lines and A real matrix composed of columns; S34, the first The data matrix obtained from the simulation Fault label corresponding to this simulation When a data sample is added to the data sample set, The preset total number of simulations has not been reached. When that happens, return to step S32 until... This forms a data sample set: in, Represents a set of data samples. Indicates the first The fault label corresponding to each simulation. This represents the total number of simulations.

4. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 3, characterized in that, The step of preserving or masking the current data matrix in the evaluation sample to obtain the processed current data matrix corresponding to that individual includes: Represent each individual in the current set of individuals as a deployment vector. : in, Indicates the first Vacuum fast-break devices are deployed at each of the candidate deployment locations. Indicates the first Vacuum instantaneous trip device will not be deployed at any of the candidate deployment locations. , Indicates the number of candidate deployment locations; The deployment vector Expand by a factor of 3 in row format to form an expanded vector. : Then extend the vector Expand in column form Multiply by 1 to form an extended matrix : in, Indicates the duration of current signal acquisition by the vacuum instantaneous tripping device; The extended matrix A matrix dot product is performed on the current data matrix corresponding to each evaluation sample in the evaluation set to preserve or mask the current data matrix, thereby obtaining the processed current data matrix corresponding to the individual. The step of inputting the processed current data matrix into the distribution network fault detection model to obtain the fault identification accuracy corresponding to that individual includes: The processed current data matrix is ​​input into the power distribution network fault detection model, and the corresponding fault type is output. Based on the fault types output by the power distribution network fault detection model and the fault types corresponding to each evaluation sample, the fault identification accuracy corresponding to the individual is obtained.

5. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 4, characterized in that, The joint optimization objective is determined in the following manner: The objective function for optimizing the deployment of the vacuum instantaneous tripping device is defined as follows: in, This represents the objective function to be optimized. This represents the fault identification accuracy when vacuum instantaneous tripping devices are deployed at all candidate deployment locations. This represents the fault identification accuracy corresponding to the individual. For the preset weighting coefficients, and ; Constraints are set according to power grid requirements or the current number of devices. These constraints include at least the following: ， in, This indicates the minimum number of vacuum instantaneous shut-off devices that can be deployed at a time.

6. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 5, characterized in that, The swarm intelligence algorithm is a genetic algorithm, and the parameters of the genetic algorithm include at least the number of individuals in the population, selection rate, crossover rate, mutation rate, and re-insertion rate. The method of generating an equal number of new individuals based on the update and iteration method of individuals in the swarm intelligence algorithm includes generating an equal number of new individuals using selection, crossover, mutation, and re-insertion operations.

7. The method for deploying a vacuum instantaneous tripping device for power distribution fault detection according to claim 6, characterized in that, The power distribution network fault detection model is a long short-term time series network, which includes: a convolutional layer for extracting local time-series features from a data matrix composed of current data corresponding to all candidate deployment locations; a recurrent layer for performing time-series correlation modeling on the local time-series features extracted by the convolutional layer; and a fully connected layer for mapping the features processed by the recurrent layer and outputting the fault type.

8. A vacuum instantaneous tripping device deployment device for power distribution fault detection, characterized in that, include: The simulation modeling module is used to acquire line parameters, distributed power sources, and load data of a power distribution network, determine the candidate deployment locations of all deployable vacuum instantaneous tripping devices in the power distribution network, and establish a power distribution network operation simulation model. The sample construction module is used to simulate different types of faults at different locations in the distribution network based on the distribution network operation simulation model. It records the current data collected by the vacuum fast-acting devices corresponding to all candidate deployment locations and the corresponding fault types under each fault simulation, forming a data sample set. Each data sample includes a data matrix composed of the current data corresponding to all candidate deployment locations and the corresponding fault type. The data preprocessing module is used to preprocess the data sample set and divide it into a training set and an evaluation set. The fault detection model construction module is used to train a deep learning model based on the training set, construct a power distribution network fault detection model, and obtain the full deployment fault identification accuracy based on the evaluation set. The deployment vector initialization module is used to characterize whether a vacuum instantaneous tripping device is deployed at each candidate deployment location as a deployment vector, and to initialize multiple deployment vectors; The iterative optimization module is used to perform iterative optimization on multiple deployment vectors based on the swarm intelligence optimization algorithm, taking the difference between the overall deployment fault identification accuracy and the fault identification accuracy corresponding to each deployment vector, and the number of vacuum fast-acting devices deployed corresponding to each deployment vector as joint optimization objectives, until the iteration termination condition is met, to obtain the optimal deployment scheme, and to determine the number and location of vacuum fast-acting devices in the distribution network based on the optimal deployment scheme.

9. An electronic device, characterized in that, It includes a processor and a memory, the processor being used to execute a computer program stored in the memory to implement the vacuum fast-break device deployment method for power distribution fault detection as described in any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores at least one instruction, which, when executed by a processor, implements the vacuum fast-break device deployment method for power distribution fault detection as described in any one of claims 1 to 7.

Images