Method and system for detecting short-circuit fault of alternating-current and direct-current micro-grid based on double-frequency resonant winding
By employing a multi-agent approach and a deep neural network combined with dual-frequency resonant windings in AC/DC microgrids, the problem of insufficient sensitivity in microgrid fault detection was solved, achieving high reliability and high accuracy in short-circuit fault detection and location.
Patent Information
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- STATE GRID FUJIAN ELECTRIC POWER CO LTD
- Filing Date
- 2026-01-28
- Publication Date
- 2026-04-17
AI Technical Summary
Existing microgrid fault detection technologies lack sufficient sensitivity in AC/DC hybrid microgrids, making it difficult to accurately detect and locate short-circuit faults, especially affected by inverter power supply current limiting and interference from complex network topology changes.
The system employs multiple agents to divide the AC/DC microgrid cable area, is equipped with a dual-frequency resonant detection module, and uses a deep neural network combined with physical information constraints to locate faults. It utilizes the dual-frequency resonant winding to capture high-frequency signals and determines the fault area through agent communication, and then combines the deep neural network for precise location.
It improves the reliability of fault criteria, reduces the risk of false alarms, and enables rapid and accurate fault area narrowing and precise location, adapting to high-precision positioning under complex working conditions.
Smart Images
Figure CN121878375A_ABST
Abstract
Description
Technical Field
[0001] This application relates to the field of power monitoring technology, and more specifically, to a method and system for detecting short-circuit faults in AC / DC microgrids based on dual-frequency resonant windings. Background Technology
[0002] With the rapid development of renewable energy, AC / DC hybrid microgrids have become an important development direction for modern distribution networks due to their ability to efficiently integrate various distributed power sources and flexibly respond to different types of load demands. Hybrid microgrids containing DC lines have complex system structures and variable operating modes, resulting in short-circuit fault characteristics that differ significantly from traditional AC power grids. This poses a significant challenge to the rapid and accurate detection and location of faults.
[0003] Existing microgrid fault detection technologies mainly face the following problems: Protection methods based on power frequency electrical quantities lack sensitivity in microgrids because their short-circuit current levels are often low and have indistinct characteristics due to the current limiting effect of inverter power supplies, easily leading to failure to operate or false operation. While methods based on high-frequency transient signals, such as traveling wave ranging, have the potential for high speed and accuracy, their application in practical microgrids is constrained by two main factors: First, the detection of high-frequency fault signals is easily affected by current changes during normal line operation and interference from the surrounding electromagnetic environment. Traditional electromagnetic transformers suffer from magnetic saturation problems, making it difficult to accurately capture weak fault characteristic signals. Second, the propagation characteristics of fault traveling waves are significantly affected by line parameters, distributed power supply switching, and network topology changes. Traditional single physical model algorithms struggle to adapt to this variability, leading to decreased location reliability. Summary of the Invention
[0004] To address the aforementioned technical problems, this invention proposes a method and system for detecting short-circuit faults in AC / DC microgrids based on dual-frequency resonant windings.
[0005] The technical solution of this invention is as follows: This invention proposes a short-circuit fault detection method for AC / DC microgrids based on dual-frequency resonant windings, comprising the following steps: A multi-agent approach is used to divide the AC / DC microgrid cables into multiple regions, with each agent equipped with a dual-frequency resonant detection module. The dual-frequency resonance detection module samples the high-frequency resonance signal of the cable excited by a short-circuit fault in real time through two sets of windings with different resonant frequencies wound on the same frame, and outputs a differential signal. When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
[0006] Preferably, the skeleton of the dual-frequency resonant winding is made of polyamide material; the number of turns of the winding is determined by calculating the transfer function derived from its distributed parameter circuit model based on its target resonant frequency.
[0007] Preferably, the transfer function of the distributed parameter circuit model is: ; In the formula: U is the transfer function; U is the output voltage; I is the input current; Angular frequency; The mutual inductance coefficient of the coil conductors; The wave impedance of the equivalent model; For sampling resistors; This is the electrical length parameter.
[0008] Preferably, the deep neural network model is a dual-channel network structure that integrates a hybrid attention mechanism. Its first channel processes electrical feature vectors, and its second channel processes wave process feature vectors. It is trained and inferred using a two-stage localization algorithm based on physical information.
[0009] Preferably, the dual-channel network structure that integrates the hybrid attention mechanism is constructed as follows: The first and second channels are each composed of a fully connected layer and a multi-head self-attention layer, used to extract the depth dependencies of electrical features and wave process features, respectively. The hybrid attention mechanism includes a channel attention module, which calculates the weights of the output feature vectors of the two channels and performs weighted fusion. The weight calculation formula is as follows: ; In the formula, Let be the output feature of the c-th channel, GAP be global average pooling, and MLP be a multilayer perceptron. The fused features are then passed through a three-layer fully connected network containing residual connections to finally output the fault location.
[0010] Preferably, the two-stage localization algorithm based on physical information includes the following steps: The input variables are fed into a deep neural network model, which outputs a preliminary fault area number, thus obtaining a coarse localization of the fault area. Based on coarse localization, a traveling wave velocity constraint derived from a distributed parameter circuit model is introduced to construct a physical information loss function. The physical information loss function and the mean square error loss function are used together as the objective function to jointly optimize the neural network model. The jointly optimized deep neural network model is used to output the precise fault location.
[0011] Preferably, the deep neural network model also integrates an online learning mechanism, which can fine-tune the model weights using real-time monitoring data when the microgrid topology undergoes long-term changes. The fine-tuning strategy adopts an elastic weight consolidation algorithm.
[0012] On the other hand, the present invention also provides an AC / DC microgrid short-circuit fault detection system based on dual-frequency resonant windings, comprising: The area division module uses multiple agents to divide the AC / DC microgrid cables into multiple areas, and each agent is equipped with a dual-frequency resonant detection module. The signal acquisition module and the dual-frequency resonance detection module sample the high-frequency resonance signal of the cable excited by the short circuit fault in real time through two sets of windings with different resonance frequencies wound on the same frame, and output a differential signal. When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
[0013] In another aspect, the present invention also provides an electronic device having a computer program stored thereon, which, when executed by a processor, implements the AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings as described in any embodiment of the present invention.
[0014] In another aspect, the present invention also provides a computer-readable medium for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings as described in any embodiment of the present invention.
[0015] The present invention has the following beneficial effects: 1. The dual-frequency resonant detection module uses a polyamide material skeleton, which fundamentally avoids magnetic saturation problems and is unaffected by surrounding magnetic fields. Through the differential output of two sets of windings with specific resonant frequencies, it can sensitively capture the high-frequency resonant signal unique to faults, while effectively suppressing common-mode interference during normal operation. The fault threshold is set to several times that of normal operation, which greatly improves the reliability of fault judgment and reduces the risk of false alarms.
[0016] 2. A two-tiered localization system of "coarse regional localization + fine point localization" was constructed. The regional localization method based on multi-agent communication can quickly narrow down the fault area to a specific cable segment. Subsequently, a deep neural network integrating a hybrid attention mechanism and physical information constraints is used for precise localization. The dual-channel structure can deeply mine the electrical energy characteristics and wave propagation characteristics of the fault, and adaptively fuse them through an attention mechanism, making the model more focused on key information. A physical information loss function is introduced, incorporating the physical laws of traveling wave propagation as constraints into the model training, ensuring that the localization results are not only data-driven but also conform to physical principles, thus maintaining extremely high localization accuracy even under complex operating conditions. Attached Figure Description
[0017] To more clearly illustrate the technical solutions of the embodiments of this application, the accompanying drawings used in the embodiments of this application will be briefly introduced below. It should be understood that the following drawings only show some embodiments of this application and should not be regarded as a limitation of the scope. For those skilled in the art, other related drawings can be obtained based on these drawings without creative effort.
[0018] Figure 1 This is a schematic diagram of the method flow in Example 1; Figure 2 This is a structural diagram of a microgrid simulation model; Figure 3 This is a diagram of the dual-frequency resonant winding structure. Figure 4 This is a schematic diagram of a distributed parameter circuit model. Detailed Implementation
[0019] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.
[0020] It should be understood that the step numbers used in the text are for ease of description only and are not intended to limit the order in which the steps are performed.
[0021] It should be understood that the terminology used in this specification is for the purpose of describing particular embodiments only and is not intended to limit the invention. As used in this specification and the appended claims, the singular forms “a,” “an,” and “the” are intended to include the plural forms unless the context clearly indicates otherwise.
[0022] The terms “comprising” and “including” indicate the presence of the described feature, whole, step, operation, element and / or component, but do not exclude the presence or addition of one or more other features, wholes, steps, operations, elements, components and / or collections thereof.
[0023] The term “and / or” refers to any combination of one or more of the associated listed items, as well as all possible combinations, and includes these combinations.
[0024] Example 1: To make the objectives, technical solutions, and advantages of this invention clearer, specific embodiments of this application will be described below, with reference to the accompanying drawings. Figure 1 The technical solution of the present invention will be clearly and completely described.
[0025] To address the problems of existing technologies, this invention provides a short-circuit fault detection method for AC / DC microgrids based on dual-frequency resonant windings, comprising the following steps: A multi-agent approach is used to divide the AC / DC microgrid cables into multiple regions, with each agent equipped with a dual-frequency resonant detection module. In this embodiment, the intelligent agent terminal device consists of a dual-frequency winding resonance detection module, a signal processing module, a fault judgment module, and a region positioning module. It can sample high-frequency resonance signals in real time to determine the occurrence of faults and to identify the fault type and fault area.
[0026] Figure 2 The image shows a microgrid simulation model. Taking two agents, A and B, as an example, agent A is responsible for the location tasks of regions 1, 2, and 3, while agent B is responsible for the location tasks of regions 4, 5, and 6. If a new region is added, the tasks are divided and assigned in this manner.
[0027] The dual-frequency resonance detection module samples the high-frequency resonance signal of the cable excited by a short-circuit fault in real time through two sets of windings with different resonant frequencies wound on the same frame, and outputs a differential signal. During operation, each intelligent agent extracts high-frequency signal characteristics during faults through the multi-frequency resonant winding detection module. Unlike traditional magnetic core materials, the multi-frequency resonant winding frame is made of polyamide, which is unaffected by surrounding magnetic fields and does not experience magnetic saturation due to current changes in the circuit under test. Two sets of windings with different numbers of turns are wound on this frame, such as... Figure 3As shown. The number of turns in each winding is calculated by the transfer function of the detection module. The transfer function is obtained through the high-frequency distributed parameter circuit model of the detection module, as shown in the figure. Figure 4 .
[0028] In this embodiment, the process for determining the number of turns for each winding is as follows: The total length of the coil is L, and the mutual inductance per unit length of the coil conductor is... The internal resistance per unit length of coil conductor considering the skin effect at high frequencies is: The inductance per unit length of coil conductor is Capacitance per unit length of coil conductor to ground Inter-turn capacitance The sampling resistor connected to both ends of the coil Let I(t) be the current in the circuit under test, and U(t) be the voltage across the sampling resistor. The transfer function can be derived using transmission line theory as follows: ; in: ; ; In the formula: U is the transfer function; U is the output voltage; I is the input current; Angular frequency; The mutual inductance coefficient of the coil conductors; The wave impedance of the equivalent model; For sampling resistors; This is the electrical length parameter.
[0029] When determining the sampling resistor (e.g.) After converting the desired resonant frequency to angular frequency (=10 ohms), and substituting it into the above formula, the coil's M can be obtained. Then, using the magnetic flux formula and the induced electromotive force formula, we can obtain: ; ; ; In the formula, r is the radius of the winding wire. Therefore, after determining the bobbin size according to the actual application, the number of winding turns N corresponding to the required resonant frequency can be obtained. For example, if the coil bobbin parameter is selected as h=15mm, =4.5mm With a diameter of 12mm and a radius of 0.05mm, to obtain a resonant frequency of 3MHz, the required number of turns is 105 turns, which can be calculated.
[0030] Using this method, two sets of resonant windings are designed, with resonant frequencies of 3MHz (105 turns) and 5MHz (60 turns), wound separately on both sides of the frame. When a short-circuit fault occurs, a high-frequency signal above 3MHz will be generated on the circuit under test passing through the winding frame, producing 2.5V and 1V outputs on the two resonant windings respectively. Connecting the same-name terminals of the two winding outputs will generate a differential output signal of 1.5V. Therefore, the short-circuit fault threshold can be set to 1V (this threshold is approximately 5 times that of normal operation). If an intelligent agent detects the differential signal value output by the dual-frequency resonant detection module... If the set fault threshold is exceeded, it indicates that a short circuit fault has occurred in the microgrid.
[0031] When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. Each agent monitors its differential signal value Vout in real time. In this embodiment, the fault threshold is set to 1V. When a certain agent (e.g., Figure 2 Agent 2 in the middle detected When the voltage is greater than 1V, a short circuit fault is determined to have occurred within its monitored area. Subsequently, the agent sends fault information to neighboring agents (such as agent 1 and agent 3) via the RS485 communication protocol. If the rear agent (such as agent 1) does not detect the fault, but the front agent (such as agent 3) detects the fault, it can be determined that the fault area is located in the cable segment between agent 2 and agent 3.
[0032] The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
[0033] After determining that the fault area lies between Agent 2 and Agent 3, Agent 2, responsible for that area, activates its built-in deep neural network model for precise localization. The input variables for this model include: distributed power supply capacity. High-frequency signal transmission speed and differential signal value Among them, high-frequency signal transmission speed Dielectric constant of cable insulation material The relevant parameters can be obtained by querying a preset conductor material parameter table or by calculating the time difference of signal transmission between adjacent intelligent agents. d represents the distance between agents, and t represents the time difference.
[0034] The deep neural network model employs a dual-channel network structure that integrates a hybrid attention mechanism, including: First channel: Specially handles [ , The electrical characteristic vector formed by [the fault] focuses on the energy characteristics of the fault.
[0035] Second channel: specifically for handling [ , The wave process characteristic vector formed by ] focuses on the propagation characteristics of fault traveling waves.
[0036] Each channel contains a fully connected layer and a multi-head self-attention layer to deeply explore the dependencies within and between features.
[0037] Channel Attention Module: This module calculates the weights of the output features from the two channels using the following formula: ,in, Let be the output feature of channel c, GAP be global average pooling, and MLP be a multilayer perceptron. Through weighted fusion, the model can adaptively focus on feature channels that are more important to the current fault scenario.
[0038] The fused features are then passed through a three-layer fully connected network with residual connections to output the specific location of the fault point within the region, for example, how many meters away from agent 2.
[0039] The training of the deep neural network model employs a two-stage localization algorithm based on physical information to improve accuracy and reliability, including: Coarse localization stage: using a large number of simulated samples (e.g., 135 sample points, covering different...) , , (Combined) The model is initially trained so that it can output a rough fault area number.
[0040] Fine-tuning phase: Introducing physical constraints into the training objective function. This includes the standard mean squared error loss function. It also added a physical information loss function. Its definition is: ; In the formula, The distance is calculated based on the wave speed and propagation time output by the model. λ represents the true known distance between agents, and λ is a hyperparameter. The total loss function is... = + This joint optimization forces the neural network's output not only to conform to the data but also to basic physical laws (the principle of traveling wave propagation), thereby significantly improving the model's positioning accuracy and physical interpretability.
[0041] As a preferred embodiment of this invention, the deep neural network model also integrates an online learning mechanism. When the microgrid topology undergoes long-term changes due to expansion or renovation, the system can utilize new real-time monitoring data and employ an elastic weight consolidation algorithm to fine-tune the model weights, enabling it to adapt to the new environment while avoiding "catastrophic forgetting" of existing knowledge.
[0042] Example 2: This embodiment provides an AC / DC microgrid short-circuit fault detection system based on dual-frequency resonant windings, including: The area division module uses multiple agents to divide the AC / DC microgrid cables into multiple areas, and each agent is equipped with a dual-frequency resonant detection module. The signal acquisition module and the dual-frequency resonance detection module sample the high-frequency resonance signal of the cable excited by the short circuit fault in real time through two sets of windings with different resonance frequencies wound on the same frame, and output a differential signal. When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
[0043] Example 3: This embodiment provides an electronic device that stores a computer program. When the computer program is executed by a processor, it implements the AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings as described in any embodiment of the present invention.
[0044] Example 4: This embodiment provides a computer-readable medium for storing one or more programs, which, when executed by one or more processors, cause the one or more processors to implement the AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings as described in any embodiment of the present invention.
[0045] In this application embodiment, "at least one" refers to one or more, and "more than one" refers to two or more. "And / or" describes the relationship between related objects, indicating that three relationships can exist. For example, A and / or B can represent the existence of A alone, A and B simultaneously, or B alone. A and B can be singular or plural. The character " / " generally indicates that the preceding and following related objects are in an "or" relationship. "At least one of the following" and similar expressions refer to any combination of these items, including any combination of singular or plural items. For example, at least one of a, b, and c can represent: a, b, c, a and b, a and c, b and c, or a and b and c, where a, b, and c can be single or multiple.
[0046] Those skilled in the art will recognize that the units and algorithm steps described in the embodiments disclosed herein can be implemented using electronic hardware, computer software, or a combination of electronic hardware and software. Whether these functions are implemented in hardware or software depends on the specific application and design constraints of the technical solution. Those skilled in the art can use different methods to implement the described functions for each specific application, but such implementation should not be considered beyond the scope of this application.
[0047] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems, devices, and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.
[0048] In the several embodiments provided in this application, any function, if implemented as a software functional unit and sold or used as an independent product, can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or a part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.
[0049] The above description is merely an embodiment of the present invention and does not limit the patent scope of the present invention. Any equivalent structural or procedural transformations made based on the content of the present invention's specification and drawings, or direct or indirect applications in other related technical fields, are similarly included within the patent protection scope of the present invention.
Claims
1. A method for short-circuit fault detection in AC / DC microgrid based on dual-frequency resonant winding, characterized in that, Includes the following steps: A multi-agent approach is used to divide the AC / DC microgrid cables into multiple regions, with each agent equipped with a dual-frequency resonant detection module. The dual-frequency resonance detection module samples the high-frequency resonance signal of the cable excited by a short-circuit fault in real time through two sets of windings with different resonant frequencies wound on the same frame, and outputs a differential signal. When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
2. The dual-frequency resonant winding based AC / DC microgrid short-circuit fault detection method according to claim 1, characterized in that: The skeleton of the dual-frequency resonant winding is made of polyamide material; the number of turns of the winding is determined by calculating the transfer function derived from its distributed parameter circuit model based on its target resonant frequency.
3. The dual-frequency resonant winding based AC / DC microgrid short-circuit fault detection method according to claim 2, characterized in that: The transfer function of the distributed parameter circuit model is: ； where: is the transfer function; U is the output voltage; I is the input current; is the angular frequency; is the coil conductor mutual inductance; is the wave impedance of the equivalent model; is the sampling resistance; is the electrical length parameter.
4. The AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings according to claim 1, characterized in that: The deep neural network model is a dual-channel network structure that integrates a hybrid attention mechanism. Its first channel processes electrical feature vectors, and its second channel processes wave process feature vectors. It is trained and inferred using a two-stage localization algorithm based on physical information.
5. The AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings according to claim 4, characterized in that: The dual-channel network structure that integrates the hybrid attention mechanism is constructed as follows: The first and second channels are each composed of a fully connected layer and a multi-head self-attention layer, used to extract the depth dependencies of electrical features and wave process features, respectively. The hybrid attention mechanism includes a channel attention module, which calculates the weights of the output feature vectors of the two channels and performs weighted fusion. The weight calculation formula is as follows: ； In the formula, Let be the output feature of the c-th channel, GAP be global average pooling, and MLP be a multilayer perceptron. The fused features are then passed through a three-layer fully connected network containing residual connections to finally output the fault location.
6. The AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings according to claim 4, characterized in that: The two-stage localization algorithm based on physical information includes the following steps: The input variables are fed into a deep neural network model, which outputs a preliminary fault area number, thus obtaining a coarse localization of the fault area. Based on coarse localization, a traveling wave velocity constraint derived from a distributed parameter circuit model is introduced to construct a physical information loss function. The physical information loss function and the mean square error loss function are used together as the objective function to jointly optimize the neural network model. The jointly optimized deep neural network model is used to output the precise fault location.
7. The AC / DC microgrid short-circuit fault detection method based on dual-frequency resonant windings according to claim 1, characterized in that: The deep neural network model also integrates an online learning mechanism. When the microgrid topology undergoes long-term changes, it can use real-time monitoring data to fine-tune the model weights. The fine-tuning strategy adopts an elastic weight consolidation algorithm.
8. A short-circuit fault detection system for AC / DC microgrids based on dual-frequency resonant windings, characterized in that, include: The area division module uses multiple agents to divide the AC / DC microgrid cables into multiple areas, and each agent is equipped with a dual-frequency resonant detection module. The signal acquisition module and the dual-frequency resonance detection module sample the high-frequency resonance signal of the cable excited by the short circuit fault in real time through two sets of windings with different resonance frequencies wound on the same frame, and output a differential signal. When an agent detects that the value of the differential signal exceeds the fault threshold, it determines that a short circuit fault has occurred in the microgrid and identifies the fault area through communication between agents. The agent responsible for the fault area takes the detected distributed power capacity, the transmission speed of the high-frequency resonant signal in the conductor, and the value of the differential signal as input variables and inputs them into a pre-trained deep neural network model. The deep neural network model outputs the specific location of the short-circuit fault within the area.
9. An electronic device comprising a memory, a processor, and a computer program stored in the memory and executable on the processor, characterized in that, When the processor executes the program, it implements the AC / DC microgrid short-circuit fault detection based on dual-frequency resonant windings as described in any one of claims 1 to 7.
10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When executed by the processor, the program implements the AC / DC microgrid short-circuit fault detection based on dual-frequency resonant windings as described in any one of claims 1 to 7.