Low-voltage power grid residual current monitoring and early warning system and method based on intelligent network

Abstract

The invention discloses a low-voltage power grid residual current monitoring and early warning system and method based on an intelligent network, and belongs to the field of safety monitoring. The method comprises the steps that a residual current sensor is installed to monitor current change, and current fluctuation is intelligently analyzed; establishing a machine learning model based on historical data, and training normal fluctuation characteristics of the residual current; the sudden change of the current is judged through a time sequence anomaly detection algorithm, and a new data optimization model is regularly utilized; the residual current fluctuation is graded, when the current deviation reaches a threshold value, the system automatically sends a notification of the corresponding grade according to the early warning grade, the position of a fault node is automatically judged through the geographic position of a sensor and a network topology structure, and a fault section is judged in combination with power grid topology data; after early warning is triggered, the system performs statistical analysis on abnormal current data, and the system automatically optimizes the acquisition frequency and the early warning threshold according to the real-time current data and the fault probability. According to the invention, through monitoring and intelligent early warning, the safety and response speed of the power grid are significantly improved.

Description

Technical Field

[0001] The present invention belongs to the field of safety monitoring, and more specifically, particularly relates to a low-voltage power grid residual current monitoring and early warning system and method based on an intelligent network. Background Art

[0002] In low-voltage power grids, with the increase in the number of monitoring devices and sensors, the amount of residual current data collected in real time is very large. Especially in high-frequency acquisition mode, the pressure of data storage, transmission and calculation increases significantly, which may cause system delays and affect the real-time nature of early warning. The centralized data processing method is difficult to efficiently process all data during high-incidence periods of faults, which can easily cause network congestion and affect system operation efficiency. There are many reasons for abnormal current fluctuations in the power grid, such as electrical aging, load fluctuations, environmental factors, etc., which can easily lead to false alarms and missed alarms. The current intelligent early warning system may not be able to effectively distinguish between slight fluctuations and real faults during abnormal detection, resulting in frequent false alarms or ignoring potential risks. The residual current characteristics of different fault types (such as short circuit, leakage, etc.) are similar, and conventional algorithms are difficult to accurately distinguish, affecting the accuracy of fault judgment.

[0003] The operating environment of low-voltage power grids is complex and changeable. Traditional fixed threshold monitoring and standardized algorithm models are often difficult to adapt to changes in different regions, loads and environmental conditions. For example, current fluctuations in severe weather may be similar to equipment failures, and existing algorithms have poor adaptability and stability in extreme cases. Due to the large differences in operating characteristics of different power grid areas, standardized algorithms are difficult to take into account all situations, and lack targeted fault analysis, resulting in uneven early warning effects. Low-voltage power grids are widely distributed, especially in remote or complex terrain areas. Real-time data transmission may encounter problems such as network delays and signal attenuation, affecting the immediacy of early warnings. Delays in data transmission may prevent fault information from being transmitted to the operation and maintenance center in the first place, increasing the difficulty of identifying hidden dangers and affecting the rapid response to emergency failures. Summary of the invention

[0004] In view of the above or existing problems of the low-voltage power grid residual current monitoring and early warning system and method based on intelligent network, the present invention is proposed.

[0005] In order to solve the above technical problems, the present invention provides the following technical solutions: The embodiment of the present invention provides a low-voltage power grid residual current monitoring and early warning method based on an intelligent network, including: installing a residual current sensor to monitor current changes and perform intelligent analysis on current fluctuations; establishing a machine learning model based on historical data to train normal fluctuation characteristics of residual current; judging sudden changes in current through a time series anomaly detection algorithm, and immediately generating an early warning signal if an abnormal fluctuation exceeding a set threshold is detected, and regularly optimizing the model using new data to improve the model's recognition rate of abnormal currents in different scenarios; The residual current fluctuation is divided into levels. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the sensor's geographical location and network topology, and the fault section is determined in combination with the power grid topology data. After the warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates a report. According to the real-time current data and fault probability, the system automatically optimizes the collection frequency and warning threshold.

[0006] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, wherein: the installation of residual current sensor to monitor current changes and perform intelligent analysis on current fluctuations includes: Preprocess the current data, remove noise and standardize it, input the preprocessed historical data into the model, build a normal current fluctuation pattern, and during online monitoring, the system compares the difference between the current data and the normal pattern; Extract features from the collected current data and input the features into the XGBoost and SVM models for classification and judgment; The classification process is as follows: Initialize the model F0(x), and at each iteration, build a new weak learner such as a decision tree to fit the residual , using the trained model F(x), calculate the final score and output the classification result. The formula is as follows: Among them, w t is the weight of each tree, h t (x) is the output of the tth tree.

[0007] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, wherein: the machine learning model is established based on historical data to train the normal fluctuation characteristics of the residual current, including: An LSTM model is established based on historical data. The normal fluctuation characteristics of residual current are trained by extracting multi-dimensional features of current time series data and performing feature engineering such as time domain, frequency domain and environmental factors. The normal current mode is identified by combining the LSTM model deep learning algorithm, and the model is continuously updated during system operation. Initialize the LSTM network, set the number of hidden layers L and the number of hidden units H, and adjust the network weights through the back propagation algorithm: Where η is the learning rate, is the gradient of the loss function with respect to the weight; Continue iterating until the validation set error converges.

[0008] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, wherein: the sudden change of current is judged by the time series abnormality detection algorithm, if abnormal fluctuation exceeding the set threshold is detected, an early warning signal is immediately generated, and the model is optimized regularly using new data to improve the model's recognition rate of abnormal current in different scenarios, including: A dynamic threshold T is set according to the normal fluctuation range in the historical data to determine whether there is abnormal fluctuation. Assuming that under normal circumstances, the mean value of current fluctuation is μ and the standard deviation is σ, the threshold T is set as: Among them, k is an adjustment parameter. When the current fluctuation value exceeds T, it can be judged as abnormal; Assume that the actual value of the current at time t is x t , the predicted value is x t1 ; The residual calculation formula is: e t =x t -x t1 If the residual e t Exceeding the threshold T, that is: |e t ∣>T, it is judged as abnormal fluctuation and a warning signal is generated; Continuously obtain current data from monitoring equipment, including normal and abnormal current fluctuation records. The data covers various equipment operating modes, environmental conditions and load levels. New data are marked as normal or abnormal. For known abnormal situations, manual annotation or external fault detection records are used as reference.

[0009] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, the residual current fluctuation is divided into levels, and when the current deviation reaches a threshold, the system automatically issues a notification of the corresponding level according to the early warning level, including: Based on historical data analysis and scenario settings, current fluctuations are divided into multiple levels such as normal, slightly abnormal, moderately abnormal, and severely abnormal. Each level defines a different threshold range, representing different degrees of risk; If the current fluctuation is within the range of ±1σ of the average value μ, it is considered normal fluctuation; If the current deviation is between ±1σ and ±2σ, it indicates a slight fluctuation and is a slight abnormality; If the current deviation is between ±2σ and ±3σ, it is moderately abnormal; If the current deviation exceeds ±3σ, there is a serious safety hazard and it is a serious abnormality; When the current fluctuation value exceeds the threshold of each level, the system automatically determines the risk level of the current fluctuation.

[0010] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, wherein: the location of the fault node is automatically determined by the geographical location of the sensor and the network topology structure, and the fault section is determined in combination with the power grid topology data, including: After detecting an anomaly, the system uses the sensor's geographic location and network topology to locate the faulty node through the following steps: Combined with the topological model, the system searches for the sensor that triggers the abnormal signal and its neighboring nodes to preliminarily determine the source of the abnormality. Based on the geographic coordinates and node relationships, the system identifies the sensor location where the abnormal signal is most concentrated and determines the faulty node. After determining the fault node, the system uses the topological model to check the adjacent power lines and branch paths to determine whether the fault propagates along the power line. The fault propagation along the power line depends on the current fluctuations on the adjacent lines, the propagation characteristics of the fault waveform along the time axis, and the physical characteristics of the power line. The propagation model is based on the following formula: in, I (u,v) is the real-time current value of the line (u, v), is the normal current value of the line (u, v); if ΔI (u,v) >Threshold, indicating that the fault has propagated to the line.

[0011] As a preferred solution of the low-voltage power grid residual current monitoring and early warning method based on intelligent network described in the present invention, wherein: after the early warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates a report, and according to the real-time current data and the fault probability, the system automatically optimizes the acquisition frequency and the early warning threshold, including: Perform multi-dimensional statistics on abnormal data, including: calculating the magnitude of the current deviation from the normal value to determine the severity of the abnormality; analyzing the frequency of abnormal fluctuations per unit time to evaluate the stability of the area or equipment; recording the duration of a single abnormal fluctuation to identify the difference between short-term fluctuations and continuous abnormalities; Identify potential patterns in abnormal data through machine learning and data mining techniques, such as the relationship between specific fault types and specific current fluctuation characteristics; Based on historical data and real-time abnormal data, the system calculates the probability of failure. When an early warning is triggered, the system analyzes whether it is a high-risk period or a high-failure probability node; According to the time periods when historical failures occurred, evaluate whether the current time is a high-risk period and define the failure frequency function: Among them, N t is the number of failures in time period t, and T is the total length of time; If the current f(t)>threshold, it is determined to be a high-risk period; Analyze the failure risk of current equipment or areas through the failure probability model: If P(fault|Ft)>threshold, the system determines it as a high failure probability node; If the system determines that it is currently in a high failure probability stage, it will automatically increase the collection frequency; if the power grid is in a stable state, the system will reduce the collection frequency.

[0012] A low-voltage power grid residual current monitoring and early warning system based on an intelligent network comprises: a current monitoring module, which is used to install a residual current sensor to monitor current changes and perform intelligent analysis on current fluctuations; a fluctuation feature training module, which is used to establish a machine learning model based on historical data and train the normal fluctuation characteristics of the residual current; a detection and judgment module, which is used to judge the sudden change of the current through a time series anomaly detection algorithm, and immediately generate an early warning signal if an abnormal fluctuation exceeding a set threshold is detected, and regularly use new data to optimize the model to improve the model's recognition rate of abnormal currents in different scenarios; a fault early warning module, which is used to classify the residual current fluctuations. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the early warning level, automatically judges the location of the fault node through the geographical location of the sensor and the network topology, and judges the fault section in combination with the power grid topology data; an optimization analysis module, which is used for the system to perform statistical analysis on the abnormal current data after the early warning is triggered, and automatically generate a report. According to the real-time current data and the fault probability, the system automatically optimizes the acquisition frequency and the early warning threshold.

[0013] A computing device, comprising: at least one processor, memory, and input-output unit; The memory is used to store computer programs, and the processor is used to call the computer programs stored in the memory to execute the steps of the low-voltage power grid residual current monitoring and early warning method based on an intelligent network.

[0014] A computer-readable storage medium includes instructions, which, when executed on a computer, enable the computer to execute the steps of a low-voltage power grid residual current monitoring and early warning method based on an intelligent network.

[0015] The beneficial effects of the present invention are as follows: through real-time monitoring and intelligent early warning, the system can quickly identify abnormal residual current fluctuations, issue early warning signals in a timely manner, and help operation and maintenance personnel discover and locate potential faults in advance, significantly improving the safety and response speed of the power grid and reducing the risk of accidents. The present invention uses advanced algorithms such as machine learning and time series analysis, and through deep learning and anomaly detection of historical data, it can accurately identify the difference between normal fluctuations and faults, reducing the possibility of false alarms and missed alarms. Especially in complex power grid environments, the system can adapt to different loads and environmental conditions to achieve more reliable fault judgment. BRIEF DESCRIPTION OF THE DRAWINGS

[0016] In order to more clearly illustrate the technical solutions of the embodiments of the present invention, the accompanying drawings required for use in the description of the embodiments will be briefly introduced below. Obviously, the accompanying drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other accompanying drawings can be obtained based on these accompanying drawings without paying creative labor.

[0017] Figure 1 A flowchart of a low-voltage power grid residual current monitoring and early warning method based on an intelligent network is provided in an embodiment of the present invention.

[0018] Figure 2 A schematic structural diagram of a low-voltage power grid residual current monitoring and early warning system based on an intelligent network is provided in an embodiment of the present invention.

[0019] Figure 3 A schematic structural diagram of a medium according to an embodiment of the present invention is schematically shown.

[0020] Figure 4 A schematic diagram of the structure of a computing device according to an embodiment of the present invention is schematically shown.

[0021] In the drawings, the same or corresponding reference numerals represent the same or corresponding parts. DETAILED DESCRIPTION

[0022] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and easy to understand, the specific implementation methods of the present invention are described in detail below in conjunction with the accompanying drawings.

[0023] In the following description, many specific details are set forth to facilitate a full understanding of the present invention, but the present invention may also be implemented in other ways different from those described herein, and those skilled in the art may make similar generalizations without violating the connotation of the present invention. Therefore, the present invention is not limited to the specific embodiments disclosed below.

[0024] Secondly, the term "one embodiment" or "embodiment" as used herein refers to a specific feature, structure, or characteristic that may be included in at least one implementation of the present invention. The term "in one embodiment" that appears in different places in this specification does not necessarily refer to the same embodiment, nor does it refer to a separate or selective embodiment that is mutually exclusive with other embodiments.

[0025] Example 1 Reference below Figure 1 , Figure 1 This is a flow chart of a low-voltage power grid residual current monitoring and early warning method based on an intelligent network provided by an embodiment of the present invention. It should be noted that the embodiments of the present invention can be applied to any applicable scenario.

[0026] Figure 1 The process of the low-voltage power grid residual current monitoring and early warning method based on the intelligent network provided by an embodiment of the present invention shown includes: S1: Install a residual current sensor to monitor current changes and perform intelligent analysis of current fluctuations.

[0027] Preferably, the current data is preprocessed to remove noise and standardize, the preprocessed historical data is input into the model, a normal current fluctuation pattern is constructed, and during online monitoring, the system compares the difference between the current data and the normal pattern; Features are extracted from the collected current data and input into the XGBoost and SVM models for classification. The classification process is as follows: Initialize the model F0(x), and at each iteration, build a new weak learner such as a decision tree to fit the residual , using the trained model F(x), calculate the final score and output the classification result. The formula is as follows: Among them, w t is the weight of each tree, h t (x) is the output of the tth tree.

[0028] Furthermore, the historical data is segmented into windows of fixed length (e.g., 5-minute windows) according to the time intervals of current fluctuations, and the fluctuation pattern is analyzed within each time window; Extract common current features from each window, such as average current, maximum value, minimum value, standard deviation, rate of change, etc.; use the processed data as a training set, and use a machine learning model such as LSTM to train the normal fluctuation feature pattern. The model will identify the characteristic range of the current during normal operation and be used for comparison during subsequent online monitoring; During real-time monitoring, the system collects current data from the sensor at regular intervals (e.g., 1 second) and extracts features of the real-time data in the same window manner; the extracted real-time feature values ​​are input into the trained normal mode model and compared with the normal current features; Calculate the difference between the current real-time data and the normal mode, and determine whether the difference exceeds the preset threshold. If the difference is large, it is considered as a possible abnormal current fluctuation and a warning signal is generated; The extracted current features are input into the XGBoost and SVM classification models to further analyze the specific types of anomalies (such as short circuit, leakage, etc.). Based on the classification results of the model, the system generates detailed fault category diagnosis results.

[0029] S2: Establish a machine learning model based on historical data and train the normal fluctuation characteristics of residual current.

[0030] Preferably, an LSTM model is established based on historical data, and the normal fluctuation characteristics of the residual current are trained by extracting multi-dimensional features of the current time series data and feature engineering such as time domain, frequency domain and environmental factors, and the normal current mode is identified by combining the LSTM model deep learning algorithm, and the model is continuously updated during system operation; Initialize the LSTM network, set the number of hidden layers L and the number of hidden units H, and adjust the network weights through the back propagation algorithm: Where η is the learning rate, is the gradient of the loss function with respect to the weight; Continue iterating until the validation set error converges.

[0031] Furthermore, the average current and standard deviation are calculated within a time window, the concentration trend of its fluctuation is recorded, and the current data is subjected to Fourier transform or wavelet transform to extract its frequency domain features. For example, the frequency distribution is analyzed by discrete Fourier transform (DFT) to observe whether there are specific frequency components in the current fluctuation to determine the fault mode; environmental factors (such as temperature, humidity, etc.) are introduced as additional features of current fluctuation. If the current is significantly affected by environmental factors, adding these characteristic variables to the LSTM model will help the model accurately identify the normal mode; A multi-layer LSTM structure is used, and the input is multi-dimensional time series features, including the time domain features, frequency domain features and environmental factor features of the current. The shape of the model input is a three-dimensional tensor (number of samples, time steps, number of features). For example, if each sample contains 60 time steps and each step contains 10 features, the input is (number of samples, 60, 10). The LSTM model is trained using historical data, and appropriate loss functions and optimization algorithms are set. Through multiple iterations, the normal operating rules of the current are identified. Each batch of training data allows the model to learn the short-term and long-term dependencies in the time series data, thereby better capturing the normal fluctuation characteristics.

[0032] S3: Through the time series anomaly detection algorithm, sudden changes in current are judged. If abnormal fluctuations exceeding the set threshold are detected, a warning signal is immediately generated. New data is regularly used to optimize the model to improve the model's recognition rate of abnormal current in different scenarios.

[0033] Preferably, a dynamic threshold T is set according to the normal fluctuation range in the historical data to determine whether there is an abnormal fluctuation. Assuming that under normal circumstances, the mean value of the current fluctuation is μ and the standard deviation is σ, the threshold T is set to: Among them, k is an adjustment parameter. When the current fluctuation value exceeds T, it can be judged as abnormal; Assume that the actual value of the current at time t is x t , the predicted value is x t1 ; The residual calculation formula is: e t =x t -x t1 If the residual e t Exceeding the threshold T, that is: |e t ∣>T, it is judged as abnormal fluctuation and a warning signal is generated; Continuously obtain current data from monitoring equipment, including normal and abnormal current fluctuation records. The data covers various equipment operating modes, environmental conditions and load levels. New data are marked as normal or abnormal. For known abnormal situations, manual annotation or external fault detection records are used as reference.

[0034] Furthermore, assuming that the mean value of normal current fluctuation is μ=5A and the standard deviation is σ=0.5 A, Using dynamic threshold formula For example, if k=3, then T=5+3·0.5=6.5A. That is to say, when the real-time current fluctuation exceeds 6.5A, the system will be regarded as abnormal; Smaller k values ​​are sensitive to small fluctuations, while larger k values ​​will only trigger warnings when there are larger fluctuations. During the testing phase, the value of k is usually adjusted repeatedly based on actual operating data and expert opinions to find the best value that meets the requirements of grid stability. The system collects the current value x at the current time t from the sensor. t , as input data; Assume that at a certain time t, the actual current value x t = 6.8A, use the prediction model to calculate the predicted value x t1 , Through the LSTM model or another time series prediction model, the system predicts the current value x at time t t1 , which is the expected value of the current under normal conditions. Assume that the model predicts the value x t1 =5.9 A, the system calculates the residual e at the current time t t , the formula is: e t =x t -x t1 , substituting the data into et=6.8-5.9=0.9A; When |et|>T, that is, 0.9>6.5 does not hold, the threshold is not exceeded and the system will not generate an early warning. If, in subsequent monitoring, it is found that |et| is greater than T, the system will determine it as an abnormal fluctuation and immediately issue an early warning signal.

[0035] S4: The residual current fluctuation is divided into levels. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the sensor's geographical location and network topology, and the fault section is determined in combination with the grid topology data.

[0036] Preferably, according to historical data analysis and scenario setting, the current fluctuation is divided into multiple levels such as normal, slightly abnormal, moderately abnormal and severely abnormal, and each level defines a different threshold range, representing different degrees of risk; If the current fluctuation is within the range of ±1σ of the average value μ, it is considered normal fluctuation; If the current deviation is between ±1σ and ±2σ, it indicates a slight fluctuation and is a slight abnormality; If the current deviation is between ±2σ and ±3σ, it is moderately abnormal; If the current deviation exceeds ±3σ, there is a serious safety hazard and it is a serious abnormality; When the current fluctuation value exceeds the threshold of each level, the system automatically determines the risk level of the current fluctuation.

[0037] Preferably, after detecting an anomaly, the system uses the geographical location of the sensor and the network topology to locate the faulty node through the following steps: Combined with the topological model, the system searches for the sensor that triggers the abnormal signal and its neighboring nodes to preliminarily determine the source of the abnormality. Based on the geographic coordinates and node relationships, the system identifies the sensor location where the abnormal signal is most concentrated and determines the faulty node. After determining the fault node, the system uses the topological model to check the adjacent power lines and branch paths to determine whether the fault propagates along the power line. The propagation of the fault along the power line depends on the current fluctuations on the adjacent lines, the propagation characteristics of the fault waveform along the time axis, and the physical characteristics of the power line. The propagation model is based on the following formula: in, I (u,v) is the real-time current value of the line (u, v), is the normal current value of the line (u, v); if ΔI (u,v) >Threshold, indicating that the fault has propagated to the line.

[0038] Furthermore, assuming xt=6.2A, the deviation from the average value μ=5A is |xt-μ|=1.2A; According to the deviation range, the current fluctuation is classified into the corresponding level: If the current value is between [4.5,5.5]A, it is a normal fluctuation and does not trigger an alarm; If the current value is between [4,4.5]A or (5.5,6]A, it is considered as a slight abnormality; If the current value is between [3.5,4)A or (6,6.5]A, it is judged as moderate abnormality; If the current value is less than 3.5A or greater than 6.5A, it is judged as a serious abnormality; In this example, the current value x t =6.2A is between (6,6.5]A, so the system determines that the current fluctuation is "moderately abnormal".

[0039] S5: After the warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates reports. Based on the real-time current data and fault probability, the system automatically optimizes the collection frequency and warning threshold.

[0040] Preferably, the abnormal data is statistically analyzed in multiple dimensions, including: calculating the magnitude of the current deviation from the normal value to determine the severity of the abnormality; analyzing the frequency of abnormal fluctuations per unit time to evaluate the stability of the area or equipment; recording the duration of a single abnormal fluctuation to identify the difference between a short-term fluctuation and a continuous abnormality; Identify potential patterns in abnormal data through machine learning and data mining techniques, such as the relationship between specific fault types and specific current fluctuation characteristics; Based on historical data and real-time abnormal data, the system calculates the probability of failure. When an early warning is triggered, the system analyzes whether it is a high-risk period or a high-failure probability node; According to the time periods when historical failures occurred, evaluate whether the current time is a high-risk period and define the failure frequency function: Among them, N t is the number of failures in time period t, and T is the total length of time; If the current f(t)>threshold, it is determined to be a high-risk period; Analyze the failure risk of current equipment or areas through the failure probability model: If P(fault|Ft)>threshold, the system determines it as a high failure probability node; If the system determines that it is currently in a high failure probability stage, it will automatically increase the collection frequency; if the power grid is in a stable state, the system will reduce the collection frequency.

[0041] Furthermore, the system sets a time window (for example, 1 hour) and counts the number of abnormal fluctuations that occur within the time period. For example, if 10 abnormal current fluctuations occur within 1 hour, the area or equipment may be unstable. By analyzing the abnormal frequency, the system determines whether the area or equipment is in a high-frequency abnormal state. If the abnormal frequency is higher than the set threshold, the system determines that the area or equipment needs further investigation or maintenance. Whenever an abnormal fluctuation is detected, the system starts to record its duration, from the beginning of the abnormality to the moment of recovery to normal. For example, if an abnormality lasts for 5 seconds, the system will record this value to distinguish between short-term abnormalities and persistent abnormalities. If the duration of abnormal fluctuation is lower than the set threshold (such as 3 seconds), it is a short fluctuation; if it exceeds the threshold, it is a persistent abnormality; persistent abnormality indicates that the equipment has a potential fault, and the system can increase the monitoring level of the equipment and generate additional warnings; The system extracts features from abnormal data, including multi-dimensional information such as deviation amplitude, duration, and abnormal frequency, and constructs an abnormal feature vector. For example, an abnormal feature vector may contain: deviation amplitude = 2.5 A, duration = 5 seconds, frequency = 10 times / hour.

[0042] After introducing the method of the exemplary embodiment of the present invention, next, refer to Figure 2 A low-voltage power grid residual current monitoring and early warning system based on an intelligent network according to an exemplary embodiment of the present invention is described. The system includes: Current monitoring module, used to install residual current sensors to monitor current changes and perform intelligent analysis of current fluctuations; Fluctuation characteristic training module, used to establish a machine learning model based on historical data and train the normal fluctuation characteristics of residual current; The detection and judgment module is used to judge the sudden change of current through the time series anomaly detection algorithm. If an abnormal fluctuation exceeding the set threshold is detected, an early warning signal is immediately generated, and the model is optimized regularly with new data to improve the model's recognition rate of abnormal current in different scenarios; The fault warning module is used to classify the residual current fluctuation. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the sensor's geographical location and network topology, and the fault section is determined in combination with the grid topology data. The optimization analysis module is used to perform statistical analysis on abnormal current data and automatically generate reports after the early warning is triggered. According to the real-time current data and fault probability, the system automatically optimizes the collection frequency and early warning threshold.

[0043] After introducing the method and apparatus of the exemplary embodiment of the present invention, next, reference is made to Figure 3 For a description of a computer-readable storage medium according to an exemplary embodiment of the present invention, please refer to Figure 3 , the computer-readable storage medium shown is a CD 30, on which a computer program (i.e., a program product) is stored. When the computer program is executed by the processor, each step recorded in the above method implementation will be implemented, for example, a residual current sensor is installed to monitor current changes and intelligently analyze current fluctuations; a machine learning model is established based on historical data to train the normal fluctuation characteristics of residual current; a sudden change of current is judged by a time series anomaly detection algorithm, and if an abnormal fluctuation exceeding a set threshold is detected, an early warning signal is immediately generated, and the model is optimized regularly using new data to improve the model's recognition rate of abnormal currents in different scenarios; The residual current fluctuation is divided into levels. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the sensor's geographical location and network topology, and the fault section is determined in combination with the power grid topology data. After the warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates a report. According to the real-time current data and fault probability, the system automatically optimizes the collection frequency and warning threshold. The specific implementation method of each step will not be repeated here.

[0044] It should be noted that examples of the computer-readable storage medium may also include, but are not limited to, phase change memory (PRAM), static random access memory (SRAM), dynamic random access memory (DRAM), other types of random access memory (RAM), read-only memory (ROM), electrically erasable programmable read-only memory (EEPROM), flash memory or other optical or magnetic storage media, which are not listed here one by one.

[0045] After introducing the method, apparatus and medium of the exemplary embodiments of the present invention, next, reference is made to Figure 4 A computing device according to an exemplary embodiment of the present invention.

[0046] Figure 4 A block diagram of an exemplary computing device 40 suitable for implementing embodiments of the present invention is shown, and the computing device 40 may be a computer system or a server. Figure 4 The computing device 40 shown is only an example and should not bring any limitation to the functionality and scope of use of the embodiments of the present invention.

[0047] like Figure 4 As shown, the components of the computing device 40 may include, but are not limited to: one or more processors or processing units 401 , a system memory 402 , and a bus 403 connecting different system components (including the system memory 402 and the processing unit 401 ).

[0048] The computing device 40 typically includes a variety of computer system readable media. These media can be any available media that can be accessed by the computing device 40, including volatile and non-volatile media, removable and non-removable media.

[0049] The system memory 402 may include computer system readable media in the form of volatile memory, such as random access memory (RAM) 4021 and / or cache memory 4022. The computing device 40 may further include other removable / non-removable, volatile / non-volatile computer system storage media. By way of example only, the ROM 4023 may be used to read and write non-removable, non-volatile magnetic media ( Figure 4 is not shown in the Figure 4As shown in FIG. 4 , a disk drive for reading and writing a removable non-volatile disk (e.g., a “floppy disk”) and an optical disk drive for reading and writing a removable non-volatile optical disk (e.g., a CD-ROM, a DVD-ROM, or other optical media) may be provided. In these cases, each drive may be connected to bus 403 via one or more data medium interfaces. System memory 402 may include at least one program product having a set (e.g., at least one) of program modules that are configured to perform the functions of various embodiments of the present invention.

[0050] A program / utility 4025 having a set (at least one) of program modules 4024 may be stored, for example, in system memory 402, and such program modules 4024 include, but are not limited to, an operating system, one or more application programs, other program modules, and program data, each of which or some combination may include an implementation of a network environment. Program modules 4024 generally perform the functions and / or methods of the embodiments described herein.

[0051] The computing device 40 may also communicate with one or more external devices 404 (e.g., a keyboard, a pointing device, a display, etc.). Such communication may be performed via an input / output (I / O) interface 405. Furthermore, the computing device 40 may also communicate with one or more networks (e.g., a local area network (LAN), a wide area network (WAN), and / or a public network, such as the Internet) via a network adapter 406. Figure 4 As shown, the network adapter 406 communicates with other modules (such as the processing unit 401, etc.) of the computing device 40 via the bus 403. It should be understood that although Figure 4 Not shown, other hardware and / or software modules may be used in conjunction with computing device 40 .

[0052] The processing unit 401 executes various functional applications and data processing by running the program stored in the system memory 402, for example, installing a residual current sensor to monitor current changes and intelligently analyzing current fluctuations; establishing a machine learning model based on historical data to train the normal fluctuation characteristics of the residual current; judging the sudden change of the current through the time series anomaly detection algorithm, if an abnormal fluctuation exceeding the set threshold is detected, an early warning signal is immediately generated, and the model is regularly optimized with new data to improve the model's recognition rate of abnormal currents in different scenarios; the residual current fluctuation is divided into levels, and when the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the early warning level, automatically determines the location of the fault node through the geographical location of the sensor and the network topology, and determines the fault section in combination with the power grid topology data; after the early warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates a report, and according to the real-time current data and the probability of failure, the system automatically optimizes the acquisition frequency and the early warning threshold. The specific implementation methods of each step will not be repeated here. It should be noted that although several units / modules or sub-units / sub-modules of the synchronous escape wiring device based on multi-commodity flow are mentioned in the above detailed description, this division is only exemplary and not mandatory. In fact, according to the embodiments of the present invention, the features and functions of two or more units / modules described above can be embodied in one unit / module. Conversely, the features and functions of one unit / module described above can be further divided into multiple units / modules to be embodied.

[0053] Those skilled in the art can clearly understand that, for the convenience and brevity of description, the specific working processes of the systems, devices and units described above can refer to the corresponding processes in the aforementioned method embodiments and will not be repeated here.

[0054] In the several embodiments provided by the present invention, it should be understood that the disclosed systems, devices and methods can be implemented in other ways. The device embodiments described above are only schematic. For example, the division of the units is only a logical function division. There may be other division methods in actual implementation. For example, multiple units or components can be combined or integrated into another system, or some features can be ignored or not executed. Another point is that the mutual coupling or direct coupling or communication connection shown or discussed can be through some communication interfaces, and the indirect coupling or communication connection of devices or units can be electrical, mechanical or other forms.

[0055] The units described as separate components may or may not be physically separated, and the components shown as units may or may not be physical units, that is, they may be located in one place or distributed on multiple network units. Some or all of the units may be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0056] In addition, each functional unit in each embodiment of the present invention may be integrated into one processing unit, or each unit may exist physically separately, or two or more units may be integrated into one unit.

[0057] If the functions are implemented in the form of software functional units and sold or used as independent products, they can be stored in a non-volatile computer-readable storage medium that can be executed by a processor. Based on this understanding, the technical solution of the present invention, or the part that contributes to the prior art, or the part of the technical solution, can be embodied in the form of a software product. The computer software product is stored in a storage medium, including several instructions for enabling a computer device (which can be a personal computer, server, or network device, etc.) to perform all or part of the steps of the method described in each embodiment of the present invention. The aforementioned storage medium includes: U disk, mobile hard disk, read-only memory (ROM, Read-Only Memory), random access memory (RAM, Random Access Memory), disk or optical disk, etc., various media that can store program codes.

[0058] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

[0059] In addition, although the operations of the method of the present invention are described in a specific order in the drawings, this does not require or imply that the operations must be performed in this specific order, or that all the operations shown must be performed to achieve the desired results. Additionally or alternatively, some steps may be omitted, multiple steps may be combined into one step, and / or one step may be decomposed into multiple steps.

Claims

1. A low voltage power grid residual current monitoring and early warning method based on intelligent network, characterized in that: include: Install residual current sensors to monitor current changes and perform intelligent analysis of current fluctuations; Build a machine learning model based on historical data to train the normal fluctuation characteristics of residual current; Through the time series anomaly detection algorithm, sudden changes in current are judged. If abnormal fluctuations exceeding the set threshold are detected, a warning signal is immediately generated. New data is regularly used to optimize the model to improve the model's recognition rate of abnormal current in different scenarios. The residual current fluctuation is divided into levels. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the geographical location of the sensor and the network topology structure, and the fault section is determined in combination with the grid topology data. After the warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates reports. Based on the real-time current data and fault probability, the system automatically optimizes the collection frequency and warning threshold.

2. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: The installation of a residual current sensor to monitor current changes and perform intelligent analysis of current fluctuations includes: Preprocess the current data, remove noise and standardize it, input the preprocessed historical data into the model, build a normal current fluctuation pattern, and during online monitoring, the system compares the difference between the current data and the normal pattern; Extract features from the collected current data and input the features into the XGBoost and SVM models for classification and judgment; The classification process is as follows: Initialize the model F0(x), and at each iteration, build a new weak learner such as a decision tree to fit the residual , using the trained model F(x), calculate the final score and output the classification result. The formula is as follows: Among them, w t is the weight of each tree, h t (x) is the output of the tth tree.

3. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: The machine learning model is established based on historical data to train the normal fluctuation characteristics of residual current, including: An LSTM model is established based on historical data. The normal fluctuation characteristics of residual current are trained by extracting multi-dimensional features of current time series data and performing feature engineering such as time domain, frequency domain and environmental factors. The normal current mode is identified by combining the LSTM model deep learning algorithm, and the model is continuously updated during system operation. Initialize the LSTM network, set the number of hidden layers L and the number of hidden units H, and adjust the network weights through the back propagation algorithm: Where η is the learning rate, is the gradient of the loss function with respect to the weight; Continue iterating until the validation set error converges.

4. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: The above-mentioned timing anomaly detection algorithm is used to judge the sudden change of current. If an abnormal fluctuation exceeding the set threshold is detected, an early warning signal is immediately generated. New data is regularly used to optimize the model to improve the model's recognition rate of abnormal current in different scenarios, including: A dynamic threshold T is set according to the normal fluctuation range in the historical data to determine whether there is abnormal fluctuation. Assuming that under normal circumstances, the mean value of current fluctuation is μ and the standard deviation is σ, the threshold T is set as: Among them, k is an adjustment parameter. When the current fluctuation value exceeds T, it can be judged as abnormal; Assume that the actual value of the current at time t is x t , the predicted value is x t1 ; The residual calculation formula is: e t =x t -x t1 If the residual e t Exceeding the threshold T, that is: |e t ∣>T, it is judged as abnormal fluctuation and a warning signal is generated; Continuously obtain current data from monitoring equipment, including normal and abnormal current fluctuation records. The data covers various equipment operating modes, environmental conditions and load levels. New data are marked as normal or abnormal. For known abnormal situations, manual annotation or external fault detection records are used as reference.

5. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: The residual current fluctuation is divided into levels. When the current deviation reaches a threshold, the system automatically issues a notification of the corresponding level according to the warning level, including: Based on historical data analysis and scenario settings, current fluctuations are divided into multiple levels such as normal, slightly abnormal, moderately abnormal, and severely abnormal. Each level defines a different threshold range, representing different degrees of risk; If the current fluctuation is within the range of ±1σ of the average value μ, it is considered normal fluctuation; If the current deviation is between ±1σ and ±2σ, it indicates a slight fluctuation and is a slight abnormality; If the current deviation is between ±2σ and ±3σ, it is moderately abnormal; If the current deviation exceeds ±3σ, there is a serious safety hazard and it is a serious abnormality; When the current fluctuation value exceeds the threshold of each level, the system automatically determines the risk level of the current fluctuation.

6. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: The method of automatically determining the location of the faulty node by using the geographical location of the sensor and the network topology structure, and determining the faulty section in combination with the grid topology data, includes: After detecting an anomaly, the system uses the sensor's geographic location and network topology to locate the faulty node through the following steps: Combined with the topological model, the system searches for the sensor that triggers the abnormal signal and its neighboring nodes to preliminarily determine the source of the abnormality. Based on the geographic coordinates and node relationships, the system identifies the sensor location where the abnormal signal is most concentrated and determines the faulty node. After determining the fault node, the system uses the topological model to check the adjacent power lines and branch paths to determine whether the fault propagates along the power line. The fault propagation along the power line depends on the current fluctuations on the adjacent lines, the propagation characteristics of the fault waveform along the time axis, and the physical characteristics of the power line. The propagation model is based on the following formula: in, I (u,v) is the real-time current value of the line (u, v), is the normal current value of the line (u, v); if ΔI (u,v) >Threshold, indicating that the fault has propagated to the line.

7. The low-voltage power grid residual current monitoring and early warning method based on intelligent network as claimed in claim 1 is characterized in that: After the warning is triggered, the system performs statistical analysis on the abnormal current data and automatically generates reports. According to the real-time current data and the probability of failure, the system automatically optimizes the acquisition frequency and the warning threshold, including: Perform multi-dimensional statistics on abnormal data, including: calculating the magnitude of the current deviation from the normal value to determine the severity of the abnormality; analyzing the frequency of abnormal fluctuations per unit time to evaluate the stability of the area or equipment; recording the duration of a single abnormal fluctuation to identify the difference between short-term fluctuations and continuous abnormalities; Identify potential patterns in abnormal data through machine learning and data mining techniques, such as the relationship between specific fault types and specific current fluctuation characteristics; Based on historical data and real-time abnormal data, the system calculates the probability of failure. When an early warning is triggered, the system analyzes whether it is a high-risk period or a high-failure probability node; According to the time periods when historical failures occurred, evaluate whether the current time is a high-risk period and define the failure frequency function: Among them, N t is the number of failures in time period t, and T is the total length of time; If the current f(t)>threshold, it is determined to be a high-risk period; Analyze the failure risk of current equipment or areas through the failure probability model: If P(fault|Ft)>threshold, the system determines it as a high failure probability node; If the system determines that it is currently in a high failure probability stage, it will automatically increase the collection frequency; if the power grid is in a stable state, the system will reduce the collection frequency.

8. A low voltage power grid residual current monitoring and early warning system based on intelligent network, characterized in that: include: Current monitoring module, used to install residual current sensors to monitor current changes and perform intelligent analysis of current fluctuations; Fluctuation characteristic training module, used to establish a machine learning model based on historical data and train the normal fluctuation characteristics of residual current; The detection and judgment module is used to judge the sudden change of current through the time series anomaly detection algorithm. If an abnormal fluctuation exceeding the set threshold is detected, an early warning signal is immediately generated, and the model is optimized regularly with new data to improve the model's recognition rate of abnormal current in different scenarios; The fault warning module is used to classify the residual current fluctuation. When the current deviation reaches the threshold, the system automatically issues a notification of the corresponding level according to the warning level. The location of the fault node is automatically determined through the sensor's geographical location and network topology, and the fault section is determined in combination with the grid topology data. The optimization analysis module is used to perform statistical analysis on abnormal current data and automatically generate reports after the early warning is triggered. According to the real-time current data and fault probability, the system automatically optimizes the collection frequency and early warning threshold.

9. A computing device, comprising: at least one processor, memory, and input-output unit; The memory is used to store a computer program, and the processor is used to call the computer program stored in the memory to execute the steps of the low-voltage power grid residual current monitoring and early warning method based on an intelligent network as described in any one of claims 1 to 7.

10. A computer-readable storage medium comprising instructions, which, when executed on a computer, enables the computer to execute the steps of the low-voltage power grid residual current monitoring and early warning method based on an intelligent network as described in any one of claims 1 to 7.

Images