Intelligent grounding inrush current suppression method and device for power distribution network

Abstract

The embodiment of the invention discloses a power distribution network intelligent grounding inrush current suppression method and device, and the method comprises the steps: collecting the state data of a power distribution network, and generating a line state data set, the state data comprising real-time load characteristics and impedance parameters; analyzing the line state data set based on an edge calculation model, and predicting an inrush current peak value and a waveform characteristic of the inrush current peak value at the grounding moment in combination with a nonlinear load fluctuation characteristic caused by the access of a distributed power supply; a particle swarm optimization model is constructed according to the predicted inrush current peak value, waveform characteristics and load characteristics, a search space is constrained through a local heuristic rule, capacitance and resistance parameters of an RC damping loop are dynamically adjusted, and an optimal damping scheme is generated; on the basis of inter-phase cooperative control logic of a three-phase system, dynamically determining grounding trigger time sequences of all phases in combination with the predicted inrush current waveform characteristics; and a grounding instruction is sent to the three-station integrated mechanism, and the three-station integrated mechanism executes an inrush current suppression operation. According to the invention, efficient inrush current suppression is realized.

Description

TECHNICAL FIELD

[0001] The application relates to the technical field of power distribution networks of power systems, in particular to a power distribution network intelligent grounding inrush current suppression method and device. BACKGROUND

[0002] With the large-scale access of distributed power sources (such as photovoltaic and wind power), the load characteristics and impedance parameters of the power distribution network present nonlinear fluctuations, resulting in complex and variable inrush current peak values and waveform characteristics during grounding faults. The traditional grounding inrush current suppression method mostly uses a fixed-parameter RC damping circuit or simple time sequence control, which is difficult to adapt to dynamic load changes caused by distributed power sources, and has problems such as limited suppression effect and slow response speed. In addition, the prior art lacks precise prediction and dynamic optimization capabilities for real-time load characteristics and inrush current waveforms, resulting in low inrush current suppression efficiency and easy damage to equipment or instability of the system.

[0003] Therefore, an intelligent grounding inrush current suppression method is urgently needed to improve the reliability and safety of the power distribution network by combining real-time data analysis and adaptive optimization technology. SUMMARY

[0004] The application provides a power distribution network intelligent grounding inrush current suppression method and device, which realizes efficient inrush current suppression.

[0005] The application provides the following solutions:

[0006] According to a first aspect, a power distribution network intelligent grounding inrush current suppression method is provided, which comprises: collecting state data of the power distribution network to generate a line state data set, wherein the state data comprises real-time load characteristics and impedance parameters; analyzing the line state data set based on an edge computing model, combining the nonlinear load fluctuation characteristics caused by the access of distributed power sources, and predicting the inrush current peak value and its waveform characteristics at the grounding moment; constructing a particle swarm optimization model according to the predicted inrush current peak value, waveform characteristics and load characteristics, and dynamically adjusting the capacitance and resistance parameters of the RC damping circuit by constraining the search space through local heuristic rules to generate an optimal damping scheme; dynamically determining the grounding trigger timing of each phase based on the inter-phase collaborative control logic of a three-phase system in combination with the predicted inrush current waveform characteristics; and sending a grounding instruction to a three-position integrated mechanism, and executing the inrush current suppression operation by the three-position integrated mechanism.

[0007] According to an implementable manner in the embodiments of the application, the collecting state data of the power distribution network to generate a line state data set comprises: deploying a high-precision synchronous sampling module in a feeder terminal unit to collect three-phase current, voltage and phase difference data at a sampling frequency of at least 10 kHz; performing denoising processing on the collected data through a wavelet transform algorithm to extract the time-frequency characteristics of the load dynamic characteristics and the line impedance, and generating a line state data set containing a time stamp.

[0008] According to an implementable manner in the embodiments of the present application, the edge computing model is used to analyze the line state data set, combined with the non-linear load fluctuation characteristics caused by the distributed power access, to predict the inrush current peak value and its waveform characteristics at the grounding moment, which comprises: deploying a light convolutional neural network model on the edge computing node, inputting the historical line state data and the line state data set; extracting the non-linear load fluctuation characteristics caused by the distributed power access through the light convolutional neural network model, wherein the fluctuation characteristics include harmonic components, power fluctuation amplitudes and phase shift characteristics; combining the non-linear load fluctuation characteristics, using a long short-term memory network to analyze the time series characteristics of the load fluctuation, and predicting the inrush current peak value and its waveform characteristics at the grounding moment.

[0009] According to an implementable manner in the embodiments of the present application, the particle swarm optimization model is constructed according to the predicted inrush current peak value, waveform characteristics and load characteristics, and the search space is constrained by local heuristic rules, which comprises: taking the predicted inrush current peak value, waveform characteristics and load characteristics as input, initializing the particle swarm optimization model, setting the particle swarm parameters to construct the initial search space; introducing an adaptive inertia weight, dynamically adjusting the particle search speed according to the iteration process, to balance the efficiency of global search and local search; constructing a local heuristic rule based on the load imbalance degree and the inrush waveform characteristics, to constrain the search space to exclude invalid parameter combinations.

[0010] According to an implementable manner in the embodiments of the present application, the local heuristic rule based on the load imbalance degree and the inrush waveform characteristics comprises: calculating the three-phase load imbalance degree, determining the dynamic threshold of the load imbalance degree based on the real-time collected three-phase current and voltage data; analyzing the inrush waveform characteristics, extracting the peak value, rise time and decay rate, and constructing a feature weight vector; combining the load imbalance degree threshold and the feature weight vector, designing a dynamic priority rule to preferentially select the parameter region that matches the load imbalance degree and can effectively suppress the inrush peak value; applying the dynamic priority rule to constrain the search space of the particle swarm optimization model, to exclude the parameter combinations that cause high inrush peak value or low damping efficiency.

[0011] According to an implementable manner in the embodiments of the present application, the phase-to-phase cooperative control logic based on the three-phase system dynamically determines the phase grounding trigger timing in combination with the predicted inrush waveform characteristics, which comprises: constructing a timing optimization algorithm based on the phase-to-phase coupling effect, calculating the optimal time interval of the phase grounding trigger in combination with the predicted inrush waveform characteristics; analyzing the phase-to-phase interference degree by real-time monitoring the three-phase load imbalance degree; dynamically adjusting the phase grounding trigger timing according to the phase-to-phase interference degree, to minimize the phase-to-phase interference and optimize the inrush suppression effect.

[0012] According to an implementable manner in the embodiments of the present application, the method further comprises: after the inrush current suppression operation is completed, collecting line state data after grounding in real time by the feeder terminal unit, comparing the line state data with the predicted inrush current waveform characteristics, and generating a deviation analysis result; and adaptively adjusting parameters of the edge computing model based on the deviation analysis result.

[0013] According to a second aspect, a power distribution network intelligent grounding inrush current suppression device is provided, the device comprising: a power grid data acquisition unit configured to acquire state data of a power distribution network, generate a line state data set, and the state data comprising real-time load characteristics and impedance parameters; a state data analysis unit configured to analyze the line state data set based on an edge computing model, predict an inrush current peak value and waveform characteristics at the grounding moment in combination with nonlinear load fluctuation characteristics caused by distributed power source access; a damping scheme generation unit configured to construct a particle swarm optimization model according to the predicted inrush current peak value, waveform characteristics and load characteristics, and dynamically adjust capacitance and resistance parameters of an RC damping circuit by local heuristic rule constraint search space, and generate an optimal damping scheme; a trigger timing determination unit configured to dynamically determine grounding trigger timing of each phase based on inter-phase cooperative control logic of a three-phase system in combination with the predicted inrush current waveform characteristics; and a grounding instruction sending unit configured to send a grounding instruction to a three-position integrated mechanism, and the three-position integrated mechanism executes an inrush current suppression operation.

[0014] According to a third aspect, a computer readable storage medium is provided, and the medium stores a computer program, and the program is executed by a processor to implement steps of the method according to any one of the first aspect.

[0015] According to a fourth aspect, an electronic device is provided, comprising: one or more processors; and a memory associated with the one or more processors, the memory being configured to store program instructions, and the program instructions being configured to perform steps of the method according to any one of the first aspect when being read and executed by the one or more processors.

[0016] According to the specific embodiments provided by the present application, the following technical effects are disclosed by the present application:

[0017] The application collects real-time load characteristics and impedance parameters through a feeder terminal unit to generate a line state data set; predicts a ground inrush peak value and waveform by using an edge computing model combined with the nonlinear load fluctuation characteristics of a distributed power source; dynamically adjusts the parameters of an RC damping circuit based on a particle swarm optimization model to generate an optimal damping scheme; optimizes the ground timing of each phase through inter-phase collaborative control logic; and finally executes an inrush suppression operation through a three-station integrated mechanism. The method significantly improves the accuracy and efficiency of inrush suppression, effectively deals with complex load fluctuations caused by the access of a distributed power source, reduces the impact of ground faults on the stability of power distribution network equipment and systems, and improves the reliability and safety of the power distribution network.

[0018] Of course, implementing any product of the present application does not necessarily require all the advantages described above to be achieved at the same time. BRIEF DESCRIPTION OF DRAWINGS

[0019] In order to more clearly illustrate the technical solutions in the embodiments of the present application or the prior art, the following will briefly introduce the drawings needed in the embodiments. Obviously, the drawings described below are only some embodiments of the present application, and other drawings can also be obtained by those skilled in the art without creative labor.

[0020] Figure 1 System architecture diagram applicable to the embodiments of the present application;

[0021] Figure 2 Flowchart of the intelligent ground inrush suppression method for the power distribution network provided by the embodiments of the present application;

[0022] Figure 3 Structure block diagram of the intelligent ground inrush suppression device for the power distribution network provided by the embodiments of the present application;

[0023] Figure 4 Schematic block diagram of an electronic device provided by the embodiments of the present application. DETAILED DESCRIPTION

[0024] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are only some of the embodiments of the present application, not all. Based on the embodiments in the present application, all other embodiments obtained by those skilled in the art belong to the scope of protection of the present application.

[0025] The terms used in the embodiments of the present application are only for the purpose of describing specific embodiments, and are not intended to limit the present application. The singular forms "a", "an" and "the" used in the embodiments of the present application and the appended claims are also intended to include the plural forms, unless the context clearly indicates otherwise.

[0026] It should be understood that the term "and / or" used herein only describes an association relationship of associated objects, which means that there can be three relationships, for example, A and / or B, which can represent three cases of A existing alone, A and B existing together, and B existing alone. In addition, the character " / " herein generally represents an "or" relationship between the associated objects before and after it.

[0027] Depending on the context, the word "if" as used herein can be interpreted as meaning "when" or "upon" or "in response to determining" or "in response to detecting". Similarly, depending on the context, the phrase "if it is determined" or "if (a stated condition or event) is detected" can be interpreted as meaning "when it is determined" or "in response to determining" or "when (a stated condition or event) is detected" or "in response to detecting (a stated condition or event)".

[0028] At present, there are some grounding surge suppression technologies, which mostly use RC damping circuits with fixed parameters or simple time sequence control, and it is difficult to adapt to dynamic load changes caused by distributed power supply, and there are problems of limited suppression effect and slow response speed.

[0029] Therefore, the present application provides a new idea. In order to facilitate the understanding of the present application, first, the system architecture based on the present application is described. Figure 1 An exemplary system architecture to which embodiments of the present application can be applied is shown in FIG. 1, which can include a user device and a power distribution network intelligent grounding surge suppression device located at the server end. Figure 1

[0030] The user can obtain the state data of the power distribution network through the user device, and the user device sends it to the power distribution network intelligent grounding surge suppression device at the server end. The power distribution network intelligent grounding surge suppression device can generate grounding instructions by using the method provided in the embodiments of the present application. The server end can send the grounding instructions to the user terminal, and the user terminal executes the surge suppression operation.

[0031] The user device can include, but is not limited to, such as: smart mobile terminal, smart home device, wearable device, PC (Personal Computer, personal computer), etc. The smart mobile device can include, such as: mobile phone, tablet computer, notebook computer, PDA (Personal Digital Assistant, personal digital assistant), Internet car, etc. The smart home device can include smart TV, smart refrigerator, etc. The wearable device can include, such as: smart watch, smart glasses, virtual reality device, augmented reality device, mixed reality device, etc.

[0032] ​The power distribution network intelligent grounding inrush suppression device can be set as an independent server, a server group, or a cloud server. The cloud server, also known as a cloud computing server or a cloud host, is a host product in the cloud computing service system, which solves the defects of large management difficulty and weak service scalability in traditional physical hosts and virtual private server (VPS, Virtual Private Server) services. In addition to the architecture shown in Figure 1 In addition to the architecture shown in

[0033] It should be understood that Figure 1 The user equipment and the power distribution network intelligent grounding inrush suppression device in

[0034] Figure 2 A power distribution network intelligent grounding inrush suppression method flowchart is provided for the embodiments of the present application. As shown in Figure 2 The method can include the following steps:

[0035] Step 201: Collecting the state data of the power distribution network to generate a line state data set, wherein the state data includes real-time load characteristics and impedance parameters.

[0036] Step 202: Analyzing the line state data set based on an edge computing model, combining the nonlinear load fluctuation characteristics caused by the access of distributed power sources, and predicting the inrush peak value and waveform characteristics of the grounding moment.

[0037] Step 203: Constructing a particle swarm optimization model according to the predicted inrush peak value, waveform characteristics, and load characteristics, constraining the search space through local heuristic rules, dynamically adjusting the capacitance and resistance parameters of the RC damping circuit, and generating an optimal damping scheme.

[0038] Step 204: Based on the inter-phase cooperative control logic of the three-phase system, combining the predicted inrush waveform characteristics, and dynamically determining the grounding trigger timing of each phase.

[0039] Step 205: Sending a grounding instruction to a three-position integrated mechanism to perform an inrush suppression operation.

[0040] As can be seen from the above process, the application collects real-time load characteristics and impedance parameters through the feeder terminal unit to generate a line state data set; predicts the grounding inrush peak value and waveform by using an edge computing model combined with the nonlinear load fluctuation characteristics of the distributed power source; dynamically adjusts the parameters of the RC damping circuit based on a particle swarm optimization model to generate an optimal damping scheme; optimizes the grounding timing of each phase through inter-phase collaborative control logic; and finally executes the inrush suppression operation through a three-station integrated mechanism. This method significantly improves the accuracy and efficiency of inrush suppression, effectively deals with complex load fluctuations caused by the access of distributed power sources, reduces the impact of grounding faults on the stability of distribution network equipment and systems, and improves the reliability and safety of the distribution network.

[0041] The steps in the above process and the effects that can be further produced will be described in detail below in conjunction with embodiments.

[0042] First, the above step 201, i.e., "collecting state data of the distribution network to generate a line state data set, the state data including real-time load characteristics and impedance parameters", will be described in detail in conjunction with an embodiment.

[0043] The state data of the distribution network mainly includes real-time load characteristics and impedance parameters, which reflect the dynamic changes of the electrical characteristics of the distribution network during operation. Real-time load characteristics usually cover three-phase current, voltage, power factor, and dynamic load fluctuation, such as load size, load type, and their trends over time. Impedance parameters include line resistance, inductance, and capacitance characteristics, which directly affect the inrush behavior during grounding faults. Through high-precision data acquisition equipment, such as a feeder terminal unit, these state data can be obtained in real time, ensuring the timeliness and accuracy of the data.

[0044] Specifically, the collection process usually relies on high-precision synchronous sampling modules deployed in the feeder terminal unit. These modules capture transient data such as three-phase current, voltage, and phase difference at a high sampling frequency, such as 10 kHz or higher, to ensure that rapidly changing electrical signals during grounding faults can be captured. The collected data is preprocessed, such as denoising by wavelet transform algorithm, to remove the influence of environmental noise or equipment interference, so as to extract the time-frequency characteristics of load dynamics and line impedance. These characteristics can reflect the nonlinear load fluctuations caused by the access of distributed power sources, such as harmonic components, power fluctuation amplitude, and phase shift, providing a reliable basis for subsequent inrush prediction.

[0045] The line state dataset generation is the result of structuring the collected state data. This dataset typically contains timestamped records covering three-phase current amplitude, voltage amplitude, phase difference, harmonic components, and impedance parameters. The structured design of the dataset facilitates input and analysis for edge computing models, while supporting time series analysis and feature extraction. By generating such a dataset, the system can comprehensively represent the operation state of the distribution network, providing a data basis for surge prediction and optimization control based on edge computing, thereby achieving precise suppression of grounding surges.

[0046] The following describes the step 202, "analysis of the line state dataset based on the edge computing model, combined with the non-linear load fluctuation characteristics caused by distributed power access, to predict the surge peak value and waveform characteristics of the grounding moment", in detail with examples.

[0047] This step utilizes the efficiency and real-time nature of edge computing to process complex dynamic data of the distribution network, accurately predicting surge characteristics during grounding faults, and providing a scientific basis for subsequent suppression operations. The line state dataset contains real-time collected three-phase current, voltage, phase difference, harmonic components, and impedance parameters, reflecting the operation state and electrical characteristics of the distribution network. Through analysis by the edge computing model, the system can extract key features from these data and accurately predict the behavior of surges by combining the special influence of distributed power.

[0048] The deployment of the edge computing model is usually on edge computing nodes close to the data source, which have strong computing capabilities, such as multi-core processors and sufficient memory, to support real-time data processing and complex algorithm operation. The core of the model is a lightweight convolutional neural network that analyzes the line state dataset and extracts non-linear load fluctuation characteristics caused by distributed power access. These features include harmonic components such as 3rd and 5th harmonic amplitudes and phases, power fluctuation amplitudes, and phase shifts caused by load changes. These non-linear characteristics are typical phenomena caused by the access of distributed power (such as photovoltaic and wind power) to the distribution network, which significantly affect the behavior of surges during grounding faults. Through the feature extraction capability of the convolutional neural network, the system can quickly identify these complex patterns, laying the foundation for surge prediction.

[0049] As an implementable solution, based on the edge computing model, the line state data set is analyzed, combined with the nonlinear load fluctuation characteristics caused by the access of distributed power supply, to predict the inrush peak and its waveform characteristics at the grounding moment, including: deploying a lightweight convolutional neural network model on the edge computing node, inputting historical line state data and line state data set; extracting the nonlinear load fluctuation characteristics caused by the access of distributed power supply through the lightweight convolutional neural network model, the fluctuation characteristics including harmonic components, power fluctuation amplitudes and phase shift characteristics; combining the nonlinear load fluctuation characteristics, using long short-term memory network to analyze the time series characteristics of load fluctuation, and predicting the inrush peak and its waveform characteristics at the grounding moment.

[0050] This process integrates the processing of historical and real-time line state data by deploying lightweight convolutional neural network model and long short-term memory network on edge computing node, extracts complex electrical characteristics caused by access of distributed power supply, and predicts inrush behavior at grounding fault. The purpose is to provide accurate prediction results for subsequent damping parameter optimization and grounding timing control, so as to effectively suppress inrush and ensure stable operation of distribution network.

[0051] The lightweight convolutional neural network model deployed on the edge computing node is the core component of the analysis process. The edge computing node is usually a local computing device close to the distribution network equipment, with high-performance processors and memory, capable of processing large amounts of data under low delay conditions. The lightweight convolutional neural network is an optimized deep learning model, which reduces the computational complexity in design and is suitable for running on resource-constrained edge devices. The model takes historical line state data and line state data set as input, historical data including three-phase current, voltage, phase difference and impedance parameters in the past period, and real-time data reflecting the current operation state of the distribution network. Through the analysis of these data, the model can capture the dynamic electrical characteristics of the distribution network and provide a basis for subsequent feature extraction.

[0052] The lightweight convolutional neural network model is specially used to extract nonlinear load fluctuation characteristics caused by the access of distributed power supply. These characteristics include harmonic components, power fluctuation amplitudes and phase shift characteristics. Harmonic components refer to high-order harmonics introduced after the access of distributed power supply (such as photovoltaic or wind power), such as 3rd and 5th harmonics, which change the waveform of current and voltage and affect the inrush characteristics. Power fluctuation amplitude reflects the instantaneous change of load power, such as the change of distributed power output with weather or time. Phase shift characteristics describe the phase difference changes between three-phase current or voltage, which may be exacerbated by load imbalance or dynamic access of distributed power supply. Convolutional neural network can efficiently identify the patterns and rules of these features through its convolutional and pooling layers, providing key inputs for inrush prediction.

[0053] To further analyze the time dimension characteristics of load fluctuations, the system introduces a long short-term memory network. Long short-term memory network is an artificial neural network specially designed for time series data, which is good at capturing long-term dependencies and short-term trends in data. In this method, the long short-term memory network is based on the nonlinear load fluctuation features extracted by the convolutional neural network, and analyzes the change law of these features over time, such as the periodicity of power fluctuations, the mutation of harmonic components or the trend of phase shift. This time series analysis can reveal the dynamic behavior of the distribution network before the occurrence of ground fault, providing more comprehensive information for predicting the transient characteristics of inrush current.

[0054] Based on the time series characteristics analyzed by the long short-term memory network, the system can predict the inrush current peak and its waveform characteristics at the moment of grounding. The inrush current peak refers to the instantaneous maximum value of current when the grounding fault occurs, which is usually much higher than the normal operating current and may cause damage to equipment. The waveform characteristics include the rise time, decay rate and waveform shape of the inrush current, which determine the destructive and suppression difficulty of the inrush current. By integrating the nonlinear features extracted by the convolutional neural network and the time series analysis of the long short-term memory network, the system can generate high-precision prediction results, such as predicting the inrush current peak to be 1500A, the rise time to be 0.2 milliseconds, and the decay rate to conform to the exponential decay model. These prediction results provide a scientific basis for subsequent RC damping circuit parameter optimization and phase-to-ground timing adjustment.

[0055] Furthermore, the long short-term memory network of the present application can use a multi-layer long short-term memory network model, which contains two hidden layers with 128 neurons each, and uses ReLU activation function to enhance the nonlinear fitting ability. The input layer receives time series characteristic data with dimensions (sample number, time step, feature number), such as (1000, 60, 6), representing 1000 samples, 60 time steps per sample, and 6 features (three-phase current, voltage and phase difference) per time step. The output layer of the model is a fully connected layer, which outputs two target variables: inrush current peak (scalar value, unit: ampere) and waveform feature vector (including rise time, decay rate, etc.). The model is optimized by mean square error loss function, using Adam optimizer with learning rate of 0.001, batch size of 32 and training rounds of 50. The training data comes from historical line state data, and the validation data is 10% of the real-time collected data set.

[0056] After the training is completed, the results are verified. The predicted results are compared with the actual grounding experiment data, and the prediction error is calculated, for example, the peak error rate = |(predicted value - actual value) / actual value| x 100%. If the error exceeds 5%, the weights of the long short-term memory network are fine-tuned by gradient descent method, and the model parameters are updated. During the verification process, the fitting degree of the waveform features is also checked, for example, the matching degree of the predicted waveform and the actual waveform is evaluated by cosine similarity. If the similarity is less than 0.9, the time step or the feature selection of the model input is adjusted, and the model is retrained.

[0057] The above step 203, i.e. "constructing a particle swarm optimization model according to the predicted surge peak, waveform features and load characteristics, constraining the search space by local heuristic rules, dynamically adjusting the capacitance and resistance parameters of the RC damping circuit, and generating the optimal damping scheme" is described in detail below in combination with an embodiment.

[0058] The particle swarm optimization model is an optimization algorithm based on swarm intelligence, simulating the foraging behavior of bird flocks, and finding the optimal solution of the objective function through iterative search. In this application, the model takes the predicted surge peak, waveform features and load characteristics as input, uses the global search capability of the particle swarm optimization algorithm, combines the real-time electrical characteristics of the distribution network, and dynamically determines the optimal parameter configuration of the RC damping circuit to maximize the suppression of the grounding surge and ensure system stability. These input data collectively define the optimization objective, i.e. minimizing the surge peak and ensuring that the damping efficiency reaches a set threshold, for example, suppressing the surge peak to within 120% of the rated current. The model initializes a set of particles, each representing a combination of capacitance and resistance parameters of the RC damping circuit, for example, the capacitance range is 0.1 microfarad to 10 microfarad, and the resistance range is 1 ohm to 100 ohms.

[0059] To improve optimization efficiency, the model introduces local heuristic rules to constrain the search space. The local heuristic rules are constraint conditions designed based on load unbalance degree and surge waveform features, aiming to exclude invalid or inefficient parameter combinations. For example, the rule may stipulate that when the three-phase load unbalance degree exceeds 10%, a larger capacitance value is preferred to enhance the damping effect; or according to the rapid rise characteristic of the surge waveform, a smaller resistance value is preferred to speed up energy dissipation. These rules dynamically adjust the boundaries of the search space by analyzing real-time load characteristics and predicted surge behavior, reducing the invalid iterations of the particle swarm, and thus accelerating the convergence to the optimal solution.

[0060] As an implementable method, the particle swarm optimization model constructed in the application includes: initializing the particle swarm optimization model with the predicted inrush peak value, waveform characteristics and load characteristics as input, setting particle swarm parameters to construct an initial search space; introducing an adaptive inertia weight, dynamically adjusting the particle search speed according to the iteration process, to balance the efficiency of global search and local search; constructing a local heuristic rule based on load unbalance degree and inrush waveform characteristics, to constrain the search space to exclude invalid parameter combinations.

[0061] The inertia weight is a key parameter that controls the moving speed of particles, and determines whether the particles tend to explore new solution space or focus on the neighborhood of the current optimal solution during the search process. The traditional fixed inertia weight may lead to local optimization or slow convergence speed, while the adaptive inertia weight achieves a balance between global search and local search by dynamically adjusting according to the iteration process. Specifically, in the early stage of iteration, the inertia weight is larger, for example, set to 0.9, to promote particles to widely explore the parameter space and cover more possible capacitor and resistor combinations; in the later stage of iteration, the weight gradually decreases, for example, to 0.4, to make the particles focus on the local optimal solution and fine-tune the parameters to improve the optimization accuracy. This dynamic adjustment mechanism significantly improves the search efficiency and convergence speed of the algorithm, ensuring that the optimal RC damping parameters are found within limited computing resources.

[0062] The local heuristic rule is a set of constraint conditions based on the real-time operating state of the distribution network, aiming to reduce the invalid iterations of the particle swarm optimization algorithm and improve the calculation efficiency. The load unbalance degree reflects the deviation degree of three-phase current or voltage, usually calculated by the formula: unbalance degree equals the maximum value of three-phase current deviation divided by the average current. Higher unbalance degree may lead to an increase in inrush peak value, so the rule may prefer larger capacitor values to enhance the damping effect. The inrush waveform characteristics include peak size, rise time and decay rate, which determine the transient behavior of the inrush. For example, a fast-rising inrush waveform may require smaller resistance values to accelerate energy dissipation. Based on these characteristics, the local heuristic rule sets dynamic thresholds for parameter range to exclude parameter combinations that cannot effectively suppress inrush or lead to low damping efficiency, such as combinations of too small capacitor or too large resistor.

[0063] The local heuristic rules can constrain the search space through a feature weight vector. For example, the system first calculates the load unbalance degree, if it exceeds 10%, the capacitance search range is limited to 2 microfarad to 10 microfarad to ensure sufficient energy absorption capacity. At the same time, the rise time of the inrush waveform is analyzed, if it is less than 0.2 milliseconds, the resistance value between 1 ohm to 50 ohm is preferred to speed up the damping response. These rules guide the particle swarm optimization algorithm to focus on the high potential parameter region through a dynamic priority mechanism, avoiding searching for invalid combinations, such as those that cause the inrush peak to exceed 120% of the rated current. By combining the load unbalance degree and the inrush waveform characteristics, the rules not only improve the optimization efficiency, but also enhance the adaptability of parameter selection to adapt to the complex load fluctuations caused by the access of distributed power.

[0064] Preferably, the construction of local heuristic rules based on load unbalance degree and inrush waveform characteristics includes: calculating the three-phase load unbalance degree, determining the dynamic threshold of load unbalance degree based on real-time collected three-phase current and voltage data; analyzing the inrush waveform characteristics, extracting the peak value, rise time and decay rate, and constructing a feature weight vector; combining the load unbalance degree threshold and the feature weight vector, designing a dynamic priority rule to preferentially select parameter regions that match the load unbalance degree and can effectively suppress the inrush peak value; applying the dynamic priority rule to constrain the search space of the particle swarm optimization model, excluding parameter combinations that cause high inrush peak or low damping efficiency.

[0065] The load unbalance degree reflects the deviation degree of three-phase current or voltage of the distribution network, which is usually calculated based on real-time collected three-phase current and voltage data. The specific method is to take the maximum value of three-phase current deviation divided by the average current, that is, the unbalance degree is equal to:

[0066] max(|Ia-Iavg|,|Ib-Iavg|,|Ic-Iavg|) / Iavg

[0067] Where Ia, Ib, Ic are three-phase currents, and Iavg is the average value. For example, if the three-phase currents are 100A, 95A and 105A, and the average current is 100A, the unbalance degree is 5%. The determination of the dynamic threshold is based on real-time data analysis, for example, set the unbalance degree threshold to 10%, when the unbalance degree exceeds this value, the system may prefer to select larger capacitance value to enhance the damping effect. This threshold will be dynamically adjusted according to the running state of the distribution network, for example, in the period when the output of distributed power fluctuates greatly, the threshold may be appropriately reduced to improve the sensitivity.

[0068] Analyzing the waveform characteristics of the inrush current and extracting the peak value, rise time, and decay rate are another core step in constructing the rules. The waveform characteristics of the inrush current describe the transient behavior of the current during a ground fault, directly affecting the selection of damping parameters. The peak value represents the maximum current value of the inrush current, such as 1500 A; the rise time refers to the time required for the current to rise from zero to the peak value, such as 0.2 milliseconds; and the decay rate reflects the speed of the inrush current decay, usually described by an exponential decay model. Through signal processing techniques, such as the Fast Fourier Transform or the Wavelet Transform, these features are extracted from the predicted inrush current waveform. Subsequently, the system constructs a feature weight vector to determine the importance of each feature through Principal Component Analysis or the Entropy Method. For example, if the rise time is short, indicating a strong transient impact of the inrush current, the weight vector may assign a higher weight to the rise time, prioritizing smaller resistance values to accelerate energy dissipation.

[0069] Designing dynamic priority rules in combination with the load unbalance threshold and the feature weight vector is the core of the local heuristic rules. Dynamic priority rules prioritize parameter regions that can effectively suppress inrush current peaks and match load unbalance based on the real-time state of the distribution network. These rules quantify the impact of load unbalance and waveform characteristics, dynamically adjusting parameter priorities to avoid selecting parameter combinations that may lead to high inrush current peaks or low damping efficiency, such as combinations with too small a capacitance or too large a resistance.

[0070] Constraining the search space of the particle swarm optimization model using dynamic priority rules is a key step in achieving efficient optimization. Dynamic priority rules set boundaries for the search space, excluding invalid parameter combinations, such as those that result in inrush current peaks exceeding 120% of the rated current or damping efficiency below 80%. In specific implementations, the rules may check the parameter values of particles at each iteration, and if they do not meet the priority conditions, their positions are reset within the constrained range, such as forcing the capacitance value to be above 2 microfarads.

[0071] Dynamic adjustment of the capacitance and resistance parameters of the RC damping circuit is the core task of optimization. The RC damping circuit is a hardware device that absorbs inrush current energy through a combination of capacitance and resistance, and its parameters directly affect the suppression effect. The particle swarm optimization model updates the position and velocity of particles at each iteration, combining global optimal solutions and individual optimal solutions to gradually approach the best parameter combination. For example, in a certain iteration, the model may determine that a combination of 2 microfarads of capacitance and 50 ohms of resistance can reduce the inrush current peak from 1500 amperes to 200 amperes while maintaining the waveform decay rate within a safe range. To balance global search and local search efficiency, the model uses an adaptive inertia weight strategy, with an initial weight set to 0.9 and gradually decaying to 0.4, to ensure that the algorithm explores a wide solution space in the early stages and focuses on local optimal solutions in the later stages.

[0072] The above step 204, i.e., "dynamically determining the grounding triggering timing of each phase based on the phase-to-phase coordinated control logic combined with the predicted inrush waveform characteristics", will be described in detail below in conjunction with an embodiment.

[0073] In a three-phase power distribution network, a ground fault can cause complex phase-to-phase coupling effects, leading to mutual interference of inrush currents among different phases, amplifying peak values or prolonging decay times. By introducing phase-to-phase coordinated control logic, the system can dynamically adjust the grounding triggering time of each phase according to the predicted inrush waveform characteristics, reduce phase-to-phase interference, and ensure that the inrush suppression effect is optimal. This process combines real-time state and prediction data of the power distribution network, embodying the characteristics of intelligent and adaptive control.

[0074] The core of the phase-to-phase coordinated control logic lies in analyzing the coupling effects in the three-phase system. The currents and voltages in the three-phase power distribution network are not completely independent, and the electrical parameters between phases will be mutually influenced due to load imbalance, impedance differences, or the connection of distributed power sources. For example, when a ground fault occurs in one phase, its inrush current may affect other phases through mutual inductance effects or neutral point currents, leading to additional inrush peak values or waveform distortions. The control logic monitors real-time three-phase currents, voltages, and phase differences, and combines predicted inrush waveform characteristics such as peak value, rise time, and decay rate to construct a phase-to-phase coupling model. This model quantifies the degree of interference of each phase grounding operation on other phases, for example, by calculating the weighted sum of phase-to-phase current deviations to evaluate the interference coefficient. This quantitative analysis provides a scientific basis for determining the grounding timing.

[0075] The process of dynamically determining the grounding triggering timing of each phase relies on a timing optimization algorithm. Based on the predicted inrush waveform characteristics, the system calculates the optimal time interval for triggering each phase, usually within the range of 0 to 5 milliseconds. For example, if the prediction shows that the A-phase inrush peak value is high and the rise time is short, the system may preferentially trigger the A-phase grounding, followed by a delay of 0.5 milliseconds to trigger the B-phase and 1 millisecond to trigger the C-phase, in order to avoid the superposition effect caused by simultaneous grounding of three phases. The optimization algorithm usually uses dynamic programming or genetic algorithms to minimize the objective function, such as the total inrush peak value or the phase-to-phase interference coefficient, through iterative calculations. Real-time monitoring of load imbalance will also affect timing adjustment, for example, when the imbalance exceeds 10%, the system may further extend the phase-to-phase triggering interval to reduce interference. This dynamic adjustment ensures that the grounding operation can adapt to the transient state of the power distribution network.

[0076] As an implementable method, the phase-to-phase coordinated control logic based on three-phase system, combined with the predicted inrush waveform characteristics, dynamically determines the phase-to-ground triggering timing, including: constructing a timing optimization algorithm based on phase-to-phase coupling effect, combining the predicted inrush waveform characteristics, calculating the optimal time interval of phase-to-ground triggering; by real-time monitoring of three-phase load unbalance degree, analyzing the degree of phase-to-phase interference; according to the degree of phase-to-phase interference, dynamically adjusting the phase-to-ground triggering timing to minimize the phase-to-phase interference and optimize the inrush suppression effect.

[0077] The construction of the timing optimization algorithm is based on the predicted inrush waveform characteristics. The inrush waveform characteristics include peak value, rise time and decay rate, which are extracted from real-time and historical line state data by edge computing model. For example, the prediction may show that the A-phase inrush peak is 1500 amperes, the rise time is 0.2 milliseconds, while the peak of B-phase and C-phase is lower. The algorithm uses these characteristics, combined with the phase-to-phase coupling model, to evaluate the impact of different triggering timing on inrush. The coupling model is usually based on circuit theory, considering the interaction of mutual inductance and neutral point current. For example, if A-phase is grounded first, it may induce additional inrush in B-phase and C-phase through mutual inductance effect, so the algorithm will calculate the time interval of delaying B-phase and C-phase triggering to avoid peak superposition. The optimization goal is to minimize the total inrush peak and waveform distortion, usually implemented by dynamic programming or particle swarm optimization algorithm, to search the optimal time interval combination by iteration.

[0078] Real-time monitoring of three-phase load unbalance degree to analyze the degree of phase-to-phase interference is a key link in the implementation of the algorithm. The load unbalance degree is calculated by real-time acquisition of three-phase current and voltage data, and the formula is unbalance degree equals to the maximum value of three-phase current deviation divided by average current. For example, when the three-phase current is 100 amperes, 95 amperes and 105 amperes, the unbalance degree is 5%. The degree of phase-to-phase interference is evaluated by quantifying the combined effect of current deviation and phase difference, for example, define the interference coefficient as a function of the weighted sum of current deviation and phase difference. Real-time monitoring data is provided by feeder terminal unit at high sampling frequency, for example, 10 kHz, to ensure capturing transient changes. When the unbalance degree is high or the phase difference is significant, the interference coefficient increases, indicating that the phase-to-phase coupling effect may exacerbate the inrush, and the triggering timing needs to be further optimized to reduce interference.

[0079] The final goal of the technical features is to dynamically adjust the grounding trigger timing of each phase according to the degree of inter-phase interference. The adjustment process is based on the output of the optimization algorithm, which dynamically updates the trigger time in combination with the real-time interference coefficient. For example, if the interference coefficient indicates that the grounding of phase A has a greater impact on phase B, the algorithm may delay the trigger time of phase B by 0.5 milliseconds to 1 millisecond to reduce the coupling effect. The adjusted timing is transmitted to the three-position integrated mechanism through the edge computing node, which executes the grounding operation with millisecond-level precision. Dynamic adjustment ensures that the timing can adapt to the transient state of the power distribution network, such as distributed power output fluctuations or load mutation scenarios, thereby suppressing the inrush current peak to within 120% of the rated current while reducing waveform distortion. This process significantly improves the efficiency and reliability of inrush current suppression by minimizing inter-phase interference.

[0080] The following describes the step 205, i.e., "sending grounding instructions to the three-position integrated mechanism, and executing inrush current suppression operations by the three-position integrated mechanism" in detail in combination with an embodiment.

[0081] The three-position integrated mechanism is an intelligent hardware device that integrates grounding, isolation, and circuit breaking functions, widely used in power distribution networks to manage faults and protect systems. The mechanism usually includes high-precision electromagnetic drive units and control circuits, capable of executing grounding operations with millisecond-level response speed. The three-position integrated mechanism receives grounding instructions from the edge computing node, which includes the grounding trigger timing of each phase and the parameter configuration of the RC damping circuit. For example, the instructions may specify that phase A triggers grounding at 0 milliseconds, phase B delays 0.5 milliseconds, and phase C delays 1 millisecond, while setting the capacitance to 2 microfarads and the resistance to 50 ohms. The fast response capability of the mechanism ensures that the grounding operation can accurately match the predicted inrush waveform characteristics, effectively absorbing inrush energy.

[0082] The generation of grounding instructions is based on the previous optimization results, integrating the predicted inrush waveform characteristics, load imbalance, and RC damping parameters output by the particle swarm optimization model. The edge computing node serves as the control center, responsible for converting these complex data into executable control signals. The instructions are transmitted to the three-position integrated mechanism through a communication interface, such as fiber-optic communication based on IEC 61850 protocol or wireless network, in a highly reliable and low-latency manner. The instruction content usually includes digital signals of trigger timing and parameter settings, ensuring that the mechanism can accurately interpret and execute. For example, if the predicted inrush peak is 1500 amperes, the optimization algorithm may determine that a 0.5 millisecond inter-phase trigger interval can suppress the peak to 200 amperes, and the instructions will explicitly specify these time points.

[0083] The execution of the surge suppression operation is the core function of the three-position integrated mechanism. The grounding operation connects the fault phase to the ground by closing the grounding switch, and simultaneously activates the RC damping circuit to absorb the surge energy. The RC damping circuit uses the combination of capacitance and resistance to dissipate electrical energy, reducing the surge peak and smoothing the waveform. For example, the capacitance absorbs instantaneous energy, and the resistance controls the energy release speed, thereby avoiding the impact on the equipment. After the operation is completed, the mechanism will feedback the execution status, such as the grounding time and actual current data, which are collected by the feeder terminal unit and uploaded to the edge computing node for subsequent deviation analysis and model optimization. This feedback mechanism ensures the closed-loop control capability of the system.

[0084] Preferably, after the completion of the surge suppression operation, the line state data after grounding is collected in real time by the feeder terminal unit, compared with the predicted surge waveform characteristics, and the deviation analysis result is generated; based on the deviation analysis result, the parameters of the edge computing model are adaptively adjusted.

[0085] Specifically, after the completion of the surge suppression operation, the feeder terminal unit is responsible for collecting the line state data after grounding in real time. These data include three-phase current, voltage, phase difference and impedance parameters, which are usually captured at a high sampling frequency, such as 10 kHz, to capture the transient electrical characteristics at the grounding moment and thereafter. The feeder terminal unit is a high-precision monitoring device deployed at the nodes of the distribution network, which can record the line state at a millisecond level resolution to ensure the capture of the actual effect of surge suppression. For example, the data may show that the actual surge peak after grounding of phase A is 200 amperes, the rise time is 0.25 milliseconds, and the decay rate conforms to the exponential decay model. These real-time data are transmitted to the edge computing node through a communication interface, such as an optical fiber network based on IEC61850 protocol, to provide a basis for subsequent analysis.

[0086] Deviation analysis is the process of comparing the actual line state data collected with the predicted surge waveform characteristics. The predicted surge waveform characteristics are generated by the edge computing model, including peak value, rise time and decay rate, etc. For example, the prediction may indicate that the surge peak is 1500 amperes, while the actual measured value is 200 amperes. Deviation analysis generates quantitative indicators by calculating the difference between the predicted value and the actual value, such as peak error rate equal to |predicted value-actual value| / actual value x 100%, or waveform similarity through cosine similarity calculation. If the peak error rate exceeds 5% or the waveform similarity is less than 0.9, it indicates that the prediction model has deviation. The deviation analysis result is generated in the form of a report, containing error statistics, waveform comparison chart and load characteristic deviation, etc., providing a basis for model optimization.

[0087] Adaptive adjustment of the edge computing model's parameters based on the bias analysis results is the key to achieving system continuous improvement. The edge computing model usually includes a lightweight convolutional neural network and a long short-term memory network for predicting surge characteristics. The adjustment process updates the model's weight parameters through optimization algorithms, such as gradient descent. For example, if the bias analysis shows that the peak prediction is too high, the system may reduce the weights related to harmonic components in the convolutional neural network or adjust the hidden layer parameters of the long short-term memory network to improve the time series prediction accuracy. The adjustment frequency is usually set to once after each surge suppression operation or once every 24 hours to ensure that the model adapts to the dynamic changes of the distribution network, such as distributed power output fluctuations or load mutations.

[0088] The above-mentioned method provided by the embodiments of the present application can be applied to various application scenarios, including but not limited to: first, in a city distribution network containing a large number of distributed power sources (such as photovoltaic and wind power), the method effectively suppresses the ground surge caused by the fluctuation of nonlinear loads by real-time monitoring and optimizing the RC damping parameters, thereby ensuring the stability of the power grid. Second, in the distribution network of highland areas, such as high-altitude substations on the Qinghai-Tibet Plateau, the thin air causes a decrease in insulation performance, and ground faults are prone to cause high surges. The method uses edge computing and interphase collaborative control to dynamically adjust the grounding timing, significantly reducing the risk of equipment damage. In addition, in the distribution network of industrial parks, in the face of frequent load mutations and complex power equipment, the method uses adaptive optimization models to quickly respond to dynamic changes and ensure power supply reliability. These scenarios benefit from the intelligence and real-time nature of the method, especially in special environments such as highland areas, improving the fault management capability of the distribution network.

[0089] The above describes certain embodiments of the present specification. Other embodiments are within the scope of the appended claims. In some cases, the actions or steps recited in the claims can be performed in an order different than the order in the embodiments and still achieve the desired result. In addition, the processes depicted in the figures do not necessarily require the particular order shown or sequential order to achieve the desired results. In certain implementations, multitasking and parallel processing can be utilized or can be advantageous.

[0090] According to another aspect, embodiments provide a distribution network intelligent ground surge suppression device. Figure 3 A schematic block diagram of the distribution network intelligent ground surge suppression device according to one embodiment is shown. As shown, the device 300 includes: Figure 3

[0091] A power grid data acquisition unit 301 configured to acquire state data of the distribution network, generate a line state data set, and the state data includes real-time load characteristics and impedance parameters.

[0092] ​The state data analysis unit 302 is configured to analyze the line state data set based on an edge computing model, in combination with the nonlinear load fluctuation characteristics caused by the access of distributed power sources, to predict the inrush peak value and its waveform characteristics at the grounding moment.

[0093] The damping scheme generation unit 303 is configured to construct a particle swarm optimization model according to the predicted inrush peak value, waveform characteristics, and load characteristics, and dynamically adjust the capacitance and resistance parameters of the RC damping circuit by locally constraining the search space through heuristic rules to generate an optimal damping scheme.

[0094] The trigger timing determination unit 304 is configured to dynamically determine the grounding trigger timing of each phase based on the inter-phase coordinated control logic of the three-phase system in combination with the predicted inrush waveform characteristics.

[0095] The grounding instruction sending unit 305 is configured to send a grounding instruction to the three-position integrated mechanism, and the three-position integrated mechanism performs an inrush suppression operation.

[0096] As an implementable way, the power grid data acquisition unit 301 can be configured to deploy a high-precision synchronous sampling module in the feeder terminal unit to acquire three-phase current, voltage, and phase difference data at a sampling frequency of at least 10 kHz when acquiring the state data of the power distribution network and generating the line state data set.

[0097] As an implementable way, the state data analysis unit 302 can be configured to deploy a lightweight convolutional neural network model on an edge computing node, input historical line state data and the line state data set, extract nonlinear load fluctuation characteristics caused by the access of distributed power sources through the lightweight convolutional neural network model, the fluctuation characteristics include harmonic components, power fluctuation amplitudes, and phase offset characteristics, in combination with the nonlinear load fluctuation characteristics, analyze the time series characteristics of load fluctuations using a long short-term memory network to predict the inrush peak value and its waveform characteristics at the grounding moment when analyzing the line state data set based on an edge computing model in combination with the nonlinear load fluctuation characteristics caused by the access of distributed power sources.

[0098] As an implementable manner, the damping scheme generation unit 303 can be configured to initialize the particle swarm optimization model with the predicted inrush peak value, waveform feature and load characteristic as input, set particle swarm parameters to build an initial search space, introduce an adaptive inertia weight to dynamically adjust the particle search speed according to the iteration process to balance the efficiency of global search and local search, and build a local heuristic rule based on the load imbalance degree and the inrush waveform feature to constrain the search space to exclude invalid parameter combinations when constructing the particle swarm optimization model according to the predicted inrush peak value, waveform feature and load characteristic.

[0099] As an implementable manner, the damping scheme generation unit 303 can be configured to calculate the three-phase load imbalance degree, determine the dynamic threshold of the load imbalance degree based on the real-time collected three-phase current and voltage data, analyze the inrush waveform feature, extract the peak value, rise time and decay rate, and build a feature weight vector, combine the load imbalance threshold and the feature weight vector to design a dynamic priority rule to preferentially select the parameter region that matches the load imbalance degree and can effectively suppress the inrush peak value, and apply the dynamic priority rule to constrain the search space of the particle swarm optimization model to exclude parameter combinations that cause high inrush peak value or low damping efficiency when constructing the local heuristic rule based on the load imbalance degree and the inrush waveform feature.

[0100] As an implementable manner, the trigger timing determination unit 304 can be configured to construct a timing optimization algorithm based on the inter-phase coupling effect, calculate the optimal time interval of each phase grounding trigger in combination with the predicted inrush waveform feature, analyze the inter-phase interference degree by real-time monitoring of the three-phase load imbalance degree, and dynamically adjust the grounding trigger timing of each phase according to the inter-phase interference degree to minimize the inter-phase interference and optimize the inrush suppression effect when dynamically determining the grounding trigger timing of each phase based on the inter-phase cooperative control logic of the three-phase system in combination with the predicted inrush waveform feature.

[0101] As an implementable manner, the device 300 can also be configured to, after the inrush suppression operation is completed, collect line state data after grounding in real time through the feeder terminal unit, compare with the predicted inrush waveform feature, generate a deviation analysis result, and adaptively adjust the parameters of the edge computing model based on the deviation analysis result.

[0102] Each of the embodiments in the specification is described in a progressive manner, and the same or similar parts between the embodiments can be referred to each other. Each of the embodiments focuses on the difference from other embodiments. In particular, the device embodiments are described more simply because they are basically similar to the method embodiments, and the relevant parts can be referred to the part of the method embodiments. The device embodiments described above are merely illustrative, and the units described as separate components can or can not be physically separated, and the components displayed as units can or can not be physical units, i.e. they can be located in one place or distributed on multiple network units. Some or all of the modules can be selected to achieve the purpose of the embodiment according to actual needs. Those skilled in the art can understand and implement it without creative labor.

[0103] It should be noted that the user information (including but not limited to user equipment information, user personal information, etc.) and data (including but not limited to data for analysis, stored data, displayed data, etc.) involved in the present application are all information and data authorized by the user or authorized by all parties, and the collection, use and processing of related data need to comply with relevant laws, regulations and standards of relevant countries and regions, and provide corresponding operation portal for user to choose authorization or refusal.

[0104] In addition, the embodiment of the present application also provides a computer readable storage medium, which stores a computer program, and the program is executed by a processor to realize the steps of the method in any one of the preceding method embodiments.

[0105] An electronic device includes one or more processors and a memory associated with the one or more processors, the memory being configured to store program instructions that, when executed by the one or more processors, perform the steps of the method in any one of the preceding method embodiments.

[0106] The present application also provides a computer program product comprising a computer program which, when executed by a processor, implements the steps of the method in any one of the preceding method embodiments.

[0107] In which, Figure 4 An exemplary architecture of an electronic device is shown, which can specifically include a processor 410, a video display adapter 411, a disk drive 412, an input / output interface 413, a network interface 414, and a memory 420. The processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420 can be connected by a communication bus 430.

[0108] The processor 410 can be implemented in a general-purpose CPU, a microprocessor, an application specific integrated circuit (ASIC), or one or more integrated circuits, etc., for executing programs related to the technical solutions provided in the present application.

[0109] The memory 420 can be implemented in a ROM (Read Only Memory), a RAM (Random Access Memory), a static storage device, a dynamic storage device, etc. The memory 420 can store an operating system 421 for controlling the operation of the electronic device 400, a basic input / output system (BIOS) 422 for controlling the low-level operation of the electronic device 400. In addition, a web browser 423, a data storage management system 424, and a smart grounding inrush current suppression device for power distribution network 425, etc. can also be stored. The smart grounding inrush current suppression device for power distribution network 425 can be an application program for implementing the foregoing steps in the embodiments of the present application. In summary, when the technical solutions provided in the present application are implemented by software or firmware, the related program codes are stored in the memory 420 and executed by the processor 410.

[0110] The input / output interface 413 is configured to connect input / output modules to realize information input and output. The input / output modules can be configured as components in the device (not shown in the figure) or externally connected to the device to provide corresponding functions. The input device can include a keyboard, a mouse, a touch screen, a microphone, various sensors, etc., and the output device can include a display, a speaker, a vibrator, an indicator light, etc.

[0111] The network interface 414 is configured to connect a communication module (not shown in the figure) to realize the communication interaction between the device and other devices. The communication module can realize communication through a wired manner (such as USB, network cable, etc.) or a wireless manner (such as mobile network, WIFI, Bluetooth, etc.).

[0112] The bus 430 includes a path for transmitting information between various components (such as the processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, and the memory 420) of the device.

[0113] It should be noted that although the above device only shows the processor 410, the video display adapter 411, the disk drive 412, the input / output interface 413, the network interface 414, the memory 420, the bus 430 and the like, in the specific implementation process, the device can also include other components necessary for normal operation. In addition, those skilled in the art can understand that the above device can also only contain the components necessary to implement the scheme of the present application, and does not have to contain all the components shown in the figure.

[0114] From the above description of the embodiments, those skilled in the art can clearly understand that the present application can be implemented by means of software and the necessary general hardware platform. Based on such understanding, the technical solutions of the present application can be embodied in the form of a computer program product, which can be stored in a storage medium such as ROM / RAM, magnetic disk, optical disk, etc., and includes a number of instructions for making a computer device (which can be a personal computer, server, or network device, etc.) execute the methods described in various embodiments or some parts of the embodiments of the present application.

[0115] The above describes the technical solutions provided by the present application in detail, and the specific examples are applied to explain the principles and implementation modes of the present application. The above description of the embodiments is only for the purpose of helping to understand the method and its core idea of the present application; at the same time, for those skilled in the art, according to the idea of the present application, the specific implementation mode and application range will be changed. In conclusion, the content of the specification should not be understood as a limitation of the present application.

Claims

1. A method for suppressing intelligent grounding inrush current in a power distribution network, characterized in that, The method includes: Collect status data of the power distribution network and generate a line status dataset, which includes real-time load characteristics and impedance parameters; The line status dataset is analyzed based on the edge computing model, and the nonlinear load fluctuation characteristics caused by distributed power supply access are combined to predict the peak inrush current and its waveform characteristics at the moment of grounding. A particle swarm optimization model is constructed based on the predicted inrush peak, waveform characteristics and load characteristics. The search space is constrained by local heuristic rules, and the capacitance and resistance parameters of the RC damping circuit are dynamically adjusted to generate the optimal damping scheme. Based on the phase-to-phase coordinated control logic of the three-phase system, and combined with the predicted inrush current waveform characteristics, the grounding trigger timing of each phase is dynamically determined. A grounding command is sent to the three-station integrated mechanism, which then performs inrush current suppression operation.

2. The intelligent grounding inrush current suppression method for distribution networks according to claim 1, characterized in that, The collection of power distribution network status data to generate a line status dataset includes: A high-precision synchronous sampling module is deployed in the feeder terminal unit to collect three-phase current, voltage and phase difference data at a sampling frequency of at least 10kHz; The collected data is denoised using wavelet transform algorithm to extract the time-frequency characteristics of load dynamics and line impedance, generating a line status dataset containing timestamps.

3. The intelligent grounding inrush current suppression method for distribution networks according to claim 1, characterized in that, The analysis of the line status dataset based on the edge computing model, combined with the nonlinear load fluctuation characteristics caused by distributed power supply access, predicts the peak inrush current and its waveform characteristics at the moment of grounding, including: A lightweight convolutional neural network model is deployed on an edge computing node, with historical line status data and the line status dataset as input. The nonlinear load fluctuation characteristics caused by distributed power source access are extracted using the lightweight convolutional neural network model. These fluctuation characteristics include harmonic components, power fluctuation amplitude, and phase offset characteristics. Based on the aforementioned nonlinear load fluctuation characteristics, a long short-time memory network is used to analyze the time series characteristics of load fluctuations and predict the peak inrush current and its waveform characteristics at the moment of grounding.

4. The intelligent grounding inrush current suppression method for distribution networks according to claim 1, characterized in that, The process of constructing a particle swarm optimization model based on predicted inrush peak value, waveform characteristics, and load characteristics, and constraining the search space using local heuristic rules, includes: Using the predicted inrush peak, waveform characteristics, and load characteristics as inputs, the particle swarm optimization model is initialized, and particle swarm parameters are set to construct the initial search space; An adaptive inertia weight is introduced to dynamically adjust the particle search speed according to the iteration process, so as to balance the efficiency of global search and local search. We construct local heuristic rules based on load imbalance and inrush waveform characteristics to constrain the search space and eliminate invalid parameter combinations.

5. The intelligent grounding inrush current suppression method for distribution networks according to claim 4, characterized in that, The local heuristic rules for constructing the rules based on load imbalance and inrush waveform characteristics include: The three-phase load imbalance is calculated, and the dynamic threshold of the load imbalance is determined based on the real-time collected three-phase current and voltage data. Analyze the characteristics of the surge waveform, extract the peak value, rise time and decay rate, and construct a feature weight vector; Combining the load imbalance threshold and the feature weight vector, a dynamic priority rule is designed to prioritize the parameter region that matches the load imbalance and can effectively suppress the inrush peak. The dynamic priority rule is applied to constrain the search space of the particle swarm optimization model and eliminate parameter combinations that lead to high inrush peak or low damping efficiency.

6. The intelligent grounding inrush current suppression method for distribution networks according to claim 1, characterized in that, The phase-to-phase coordinated control logic based on the three-phase system, combined with the predicted inrush current waveform characteristics, dynamically determines the grounding trigger timing of each phase, including: A timing optimization algorithm based on interphase coupling effect is constructed, and the optimal time interval for grounding triggering of each phase is calculated by combining the predicted inrush current waveform characteristics. By monitoring the three-phase load imbalance in real time, the degree of inter-phase interference can be analyzed; The grounding triggering sequence of each phase is dynamically adjusted according to the interphase interference level to minimize interphase interference and optimize inrush current suppression.

7. The intelligent grounding inrush current suppression method for distribution networks according to claim 1, characterized in that, The method further includes: After the inrush current suppression operation is completed, the feeder terminal unit collects the line status data after grounding in real time, compares it with the predicted inrush current waveform characteristics, and generates deviation analysis results. Based on the deviation analysis results, the parameters of the edge computing model are adaptively adjusted.

8. A smart grounding inrush current suppression device for power distribution networks, characterized in that, The device includes: The power grid data acquisition unit is configured to acquire status data of the distribution network and generate a line status dataset, wherein the status data includes real-time load characteristics and impedance parameters. The status data analysis unit is configured to analyze the line status dataset based on the edge computing model, and predict the peak inrush current and its waveform characteristics at the moment of grounding by combining the nonlinear load fluctuation characteristics caused by the access of distributed power sources. The damping scheme generation unit is configured to construct a particle swarm optimization model based on the predicted inrush peak, waveform characteristics and load characteristics, and dynamically adjust the capacitance and resistance parameters of the RC damping circuit by constraining the search space through local heuristic rules to generate the optimal damping scheme. The trigger timing determination unit is configured to dynamically determine the grounding trigger timing of each phase based on the phase-to-phase coordinated control logic of the three-phase system and the predicted inrush current waveform characteristics. The grounding command sending unit is configured to send a grounding command to the three-station integrated mechanism, which then performs inrush current suppression operation.

9. An electronic device, characterized in that, include: One or more processors; and a memory associated with the one or more processors, the memory being used to store program instructions that, when read and executed by the one or more processors, perform the steps of the method according to any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the program is executed by the processor, it implements the steps of the method according to any one of claims 1 to 7.

Images