Intelligent power distribution network monitoring and fault detection system

Abstract

The invention discloses an intelligent power distribution network monitoring and fault detection system, and relates to the technical field of power distribution network detection. Comprising a power distribution network region division module, a periodic data acquisition and monitoring module, a data preprocessing and feature extraction module, an intelligent load state evaluation module, a sub-region state division module, a conventional monitoring module and a dynamic adjustment and anomaly response module, therefore, finer monitoring and management can be realized. According to the method, the power distribution network is divided into the sub-regions, and periodic state scanning is combined, so that the system can accurately identify local overload and abnormal load and respond to abnormal fluctuation in time. LSTM is combined to evaluate load fluctuation, automatic learning adapts to load change, the intelligence, flexibility and stability of the system are improved, and safety and reliability of power supply are ensured.

Description

Technical Field

[0001] The present invention relates to the technical field of distribution network detection, and particularly to an intelligent distribution network monitoring and fault detection system. Background Art

[0002] Intelligent distribution network monitoring and fault detection refers to monitoring the operating status of a distribution network and automatically detecting and identifying faults or abnormal conditions therein through advanced sensing technologies, data analysis, artificial intelligence (AI), and machine learning. Specifically, an intelligent distribution network monitoring system can collect data of various devices within the distribution network, including parameters such as voltage, current, power, and load, and use this data for real-time analysis to identify potential equipment failures, line short circuits, overloads, undervoltage, and other problems. At the same time, fault detection technologies can issue early warnings through methods such as comparison with historical data, abnormal pattern recognition, and predictive maintenance to avoid large-scale power outages or system collapses. Through this intelligent monitoring and fault detection, not only can the reliability and efficiency of the power system be improved, but also intelligent fault location and rapid recovery can be achieved, enhancing the self-healing ability of the power grid.

[0003] The prior art has the following deficiencies:

[0004] Existing intelligent distribution network monitoring systems usually continuously track the status of the distribution network through periodic operating status scans to promptly detect abnormal fluctuations. However, local overloads, especially those caused by concentrated loads or temporary high loads, may be difficult to identify in a timely manner. Since a distribution network consists of multiple regions, lines, and devices, local overloads usually do not immediately affect the operation of the entire network. Therefore, monitoring systems based on average network-wide data often overlook this problem. Such local overloads can lead to unstable power supply in certain areas, increased voltage fluctuations, and may trigger protective power outages in severe cases to prevent equipment damage. As a result, the unstable power supply may cause frequent voltage fluctuations and power outages, and even lead to the entire power system being disconnected from the grid, seriously affecting residents' lives and enterprise production, and bringing inconvenience and huge economic losses to users.

[0005] The above information disclosed in the background art section is only used to enhance the understanding of the background of the present disclosure. Therefore, it may include information that does not constitute the prior art known to those of ordinary skill in the art. Summary of the Invention

[0006] The object of the present invention is to provide an intelligent distribution network monitoring and fault detection system. By dividing the distribution network into multiple sub-regions and introducing periodic operation state scanning, the monitoring system can track and accurately identify local overload or abnormal load in real time. Especially in the case of sudden load increase or local fault, it can timely detect and respond to abnormal fluctuations. This regionalized management combines key features such as reactive power demand and total harmonic distortion, improving the accuracy and response ability of the system. Using the long short-term memory network (LSTM) to evaluate the load fluctuation index, the system automatically learns and adapts to load changes, marks and focuses on monitoring abnormal regions in real time, significantly enhancing the intelligence, flexibility and stability of the distribution network, and ensuring the safety and reliability of power supply to solve the problems in the above background technology.

[0007] To achieve the above object, the present invention provides the following technical solutions: an intelligent distribution network monitoring and fault detection system, including a distribution network regional division module, a periodic data acquisition and monitoring module, a data preprocessing and feature extraction module, an intelligent load state evaluation module, a sub-region state division module, a conventional monitoring module, and a dynamic adjustment and abnormal response module:

[0008] The distribution network regional division module divides the entire distribution network into multiple sub-regions to achieve more refined monitoring and management;

[0009] The periodic data acquisition and monitoring module obtains the operation data of each sub-region of the distribution network by implementing periodic operation state scanning, and continuously tracks the operation state of the distribution network;

[0010] The data preprocessing and feature extraction module preprocesses the data obtained from each sub-region, extracts the key features reflecting the load state of the sub-region from it, and under the detection window, quantitatively analyzes the extracted key features and converts them into numerical indicators reflecting the current load state;

[0011] The intelligent load state evaluation module inputs the quantified features into a pre-trained long short-term memory network, and intelligently evaluates the load state of each sub-region through the long short-term memory network;

[0012] The sub-region state division module divides all sub-regions into local voltage abnormal loads and regional load normal according to the evaluation results of the long short-term memory network, and marks the sub-regions classified as "local voltage abnormal loads" for subsequent key monitoring and processing;

[0013] The conventional monitoring module continues to perform conventional monitoring on the sub-regions classified as regional load normal with a preset periodic operation state scanning to ensure the stable operation of the overall distribution network;

[0014] The dynamic adjustment and abnormal response module dynamically adjusts the data acquisition frequency for sub-regions marked as locally voltage-abnormal loads to more accurately identify and respond to abnormal states.

[0015] Preferably, key features reflecting the load status of the sub-region are extracted from the pre-processed data. The extracted features include the reactive power demand within the sub-region and the degree of harmonic distortion in the power system. Under the detection window, the reactive power demand within the sub-region and the degree of harmonic distortion in the power system are analyzed to generate a reactive power load reference value and a total harmonic distortion reference value respectively. The reactive power load reference value quantifies the relationship between the reactive power demand within the sub-region and the total load, and the total harmonic distortion reference value quantifies the ratio of harmonic current to fundamental current in the power system, reflecting the severity of harmonics in the power system.

[0016] Preferably, the specific steps for analyzing the reactive power demand within the sub-region to generate a reactive power load reference value under the detection window are as follows:

[0017] First, collect the reactive power demand data within the sub-region through monitoring devices. The reactive power within the sub-region is calculated by the phase difference between voltage and current, and the calculation expression is as follows:

[0018]

[0019] , where Q is the reactive power of the sub-region, I k is the current at the k-th load point in the sub-region, V k is the voltage at the k-th load point in the sub-region, θ k is the phase difference between the current and voltage at the k-th load point in the sub-region, n is the number of load points in the sub-region, and sin(θ k ) is the sine value of the phase difference between the current and voltage;

[0020] After collecting the reactive power demand data Q, analyze the dynamic characteristics of the reactive power demand to reveal the change trend of the load in the sub-region. The calculation expression is as follows:

[0021] Q weighted =λ·Q+(1 - λ)·Q previous

[0022] , where Q weighted is the weighted reactive power demand, λ is the smoothing factor, which controls the contribution degree of the current value and the historical value to the weighted average result, and Q previous is the reactive power demand value at the previous moment;

[0023] After analyzing the load dynamic characteristics, combine the reactive power demand of the sub-region with the load density of the power system in this region to calculate the load deviation. The load deviation calculation expression is as follows:

[0024]

[0025] , where D is the sub - area load deviation, A j is the area of the j - th load point in the sub - area, I j is the current demand of the j - th load point in the sub - area;

[0026] The reactive power demand Q after comprehensive weighting weighted and the sub - area load deviation D are used to generate the reactive power load reference value. The generation formula is as follows:

[0027] Reactive Power = α·q weighted + β·d

[0028] , where Reactive Power is the reactive power load reference value, α is the weight coefficient of the weighted reactive power demand Q weighted and β is the weight coefficient of the sub - area load deviation D.

[0029] Preferably, the specific steps for analyzing the harmonic distortion degree in the power system under the detection window to generate the total harmonic distortion reference value are as follows:

[0030] First, perform spectral analysis on the voltage and current signals of the power system through fast Fourier transform to extract each harmonic component in the power system. The extraction formula is as follows:

[0031]

[0032] , where X(h) is the amplitude of the h - th harmonic in the frequency domain, called the spectral component, x(m) is the value of the signal at the m - th sampling point, e is the natural base, p is the imaginary unit, 2π is a mathematical constant used to calibrate the periodicity of sine waves and cosine waves, h is the frequency index representing the frequency component of the h - th harmonic, m is the sampling point index of the time - domain signal, and M is the total number of signal sampling points;

[0033] After extracting the harmonic component X(h), quantify the distortion degree of each harmonic component. The formula is as follows:

[0034]

[0035] , where HD(h) is the distortion degree of the h - th harmonic, X(1) is the fundamental wave amplitude, and ω is the adjustment factor;

[0036] After obtaining the distortion degree HD(h) of each harmonic, summarize according to the distortion degrees of all harmonic components and calculate the total harmonic distortion reference value. The calculation expression is as follows:

[0037]

[0038] , where Total Harmonic is the total harmonic distortion reference value, H is the maximum value of the harmonic order, τ is the weighting coefficient, and γ is the reference value control parameter.

[0039] Preferably, the reactive power load reference value and the total harmonic distortion reference value after quantization are input into a pre-trained long short-term memory network. The load fluctuation index is generated through the long short-term memory network, and the load status of the sub-region is intelligently evaluated through the load fluctuation index.

[0040] Preferably, the load fluctuation index generated by intelligently evaluating the load status of the sub-region through a pre-trained long short-term memory network is compared and analyzed with a pre-set load fluctuation index reference threshold, and the state of the sub-region of the distribution network is divided. The division steps are as follows:

[0041] If the load fluctuation index is greater than the pre-set load fluctuation index reference threshold, the current sub-region is divided into a local voltage abnormal load;

[0042] If the load fluctuation index is less than or equal to the pre-set load fluctuation index reference threshold, the current sub-region is divided into a region with normal load.

[0043] Preferably, for the sub-region marked as a local voltage abnormal load, the data acquisition frequency is dynamically adjusted to more accurately identify and respond to the abnormal state. The specific steps are as follows:

[0044] When the load fluctuation index Load Fluctuation is greater than the load fluctuation index reference threshold, the data acquisition frequency is dynamically adjusted to ensure that potential abnormalities can be more accurately captured when there is a sharp load fluctuation in the sub-region. The frequency adjustment formula is as follows:

[0045]

[0046] , where f new is the adjusted data acquisition frequency, f base is the original periodic operation state scanning frequency, ρ is the adjustment coefficient, Load Fluctuation is the load fluctuation index, Load ref is the load fluctuation index reference threshold, is the exponential adjustment factor of the load fluctuation intensity, Δt avg is the average duration of the current load fluctuation, Δt ref is the reference load fluctuation duration, and μ is the adjustment coefficient;

[0047] The adjusted data acquisition frequency f newIt will be further optimized based on the feedback mechanism to adapt to real-time load changes, making the frequency change more flexible and efficient. The formula is as follows:

[0048]

[0049] , where f adjusted is the optimized data acquisition frequency, δ is the dynamic adjustment coefficient, Load avg is the average value of the load fluctuation index, ζ is the exponential weight of the load fluctuation index, N missed is the number of undetected anomalies, N max is the maximum allowable number of undetected anomalies, and θ is the anomaly undetected weight index.

[0050] In the above technical solution, the technical effects and advantages provided by the present invention are as follows:

[0051] By dividing the distribution network into multiple sub-regions, the present invention enables the monitoring system to conduct meticulous monitoring and management of each sub-region, ensuring that the load conditions of each region are independently and real-time tracked. Different from the traditional monitoring system based on the average data of the entire network, this sub-region division can effectively identify local overload or abnormal load. Especially when the load suddenly increases or local equipment fails, it can promptly capture abnormal fluctuations. Key features such as reactive power demand and total harmonic distortion can reflect changes in load imbalance, equipment anomalies, or power quality problems. The system obtains data in real-time through periodic operation status scanning and quickly responds when anomalies are detected, avoiding large-scale power system failures caused by local overload. In addition, for the overload situation of certain specific sub-regions, the system will dynamically adjust the data acquisition frequency to achieve more accurate anomaly identification. Through this precise monitoring and response mechanism, the stability of the distribution network can be significantly improved, reducing voltage fluctuations, power outages, and system collapses, ensuring the reliability and security of power supply.

[0052] The present invention enables the distribution network monitoring system to have stronger intelligence and adaptability by introducing a long short-term memory network (LSTM). The LSTM can analyze historical data and capture the temporal characteristics of load fluctuations, thereby achieving accurate prediction and intelligent evaluation of the load status. The system can not only monitor the load conditions of sub-regions in real time but also automatically learn and adapt to load changes, seasonal fluctuations, and emergencies. After the reactive power load reference value and the total harmonic distortion reference value are input into the LSTM network, the system can generate a load fluctuation index, provide a dynamic assessment of the load status, and compare and analyze it with a preset threshold. Once the load fluctuation index exceeds the preset threshold, the system automatically determines the local voltage abnormal load, marks the area in real time, and conducts key monitoring. This adaptive intelligent analysis not only reduces the need for human intervention but also adjusts the monitoring strategy according to different power demand changes. Through this intelligent processing, the distribution network can automatically adjust under changing load conditions, significantly improving the flexibility, stability, and fault response ability of the system, and ultimately ensuring the efficiency and safety of power supply. BRIEF DESCRIPTION OF THE DRAWINGS

[0053] 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 required for use in the embodiments. Obviously, the drawings described below are only some embodiments recorded in the present invention. For those of ordinary skill in the art, other drawings can also be obtained based on these drawings.

[0054] Figure 1 It is a schematic diagram of the modules of the intelligent distribution network monitoring and fault detection system of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0055] Now, the exemplary embodiments will be described more comprehensively with reference to the accompanying drawings. However, the exemplary embodiments can be implemented in various forms and should not be construed as limited to the examples set forth herein; rather, these exemplary embodiments are provided so that the present disclosure will be more comprehensive and complete, and the concept of the exemplary embodiments will be fully conveyed to those skilled in the art.

[0056] The present invention provides an intelligent distribution network monitoring and fault detection system as Figure 1 shown, including a distribution network area division module, a periodic data collection and monitoring module, a data preprocessing and feature extraction module, an intelligent load status evaluation module, a sub-region status division module, a conventional monitoring module, and a dynamic adjustment and anomaly response module:

[0057] The distribution network area division module divides the entire distribution network into multiple sub-regions to achieve more refined monitoring and management;

[0058] A distribution network is usually composed of multiple regions, lines, and devices, and there may be significant differences in the load characteristics and operating conditions of each region. First, the distribution network is divided into several sub-regions based on factors such as geographical location, load distribution, and device characteristics. Such a division enables the monitoring system to pay more attention to the specific operating conditions of each sub-region rather than relying solely on the average data of the entire network. A reasonable regional division can not only improve the accuracy of monitoring but also optimize resource allocation to ensure that key regions are given priority monitoring and management.

[0059] The periodic data acquisition and monitoring module obtains the operation data of each sub-region of the distribution network by implementing periodic operation status scans and continuously tracks the operation status of the distribution network;

[0060] "By implementing periodic operation status scans, obtaining the operation data of each sub-region of the distribution network, and continuously tracking the operation status of the distribution network" means collecting the real-time operation data of each sub-region of the distribution network regularly (or at preset time intervals) for monitoring and analyzing its current status. In the prior art, the distribution network usually obtains real-time data through sensors (such as current, voltage sensors, power sensors, etc.) installed at various key positions. These sensors will periodically transmit the data to the central monitoring system, and the system reflects information such as the load, device operation status, and power quality of each sub-region in real time by continuously updating these data. Periodic scans not only help the monitoring system track the daily operation of the distribution network but also detect potential abnormal fluctuations in each sub-region, such as local overload, load imbalance, etc. In this way, the system can timely discover problems that may affect power stability, provide a basis for subsequent processing and adjustment, and thus prevent power outages or equipment damage caused by overload, faults, etc.

[0061] These data include key parameters such as voltage, current, power, and load, reflecting the real-time operation status of the distribution network. Periodic scans ensure that the system can timely capture the changes and potential abnormalities during operation, providing basic data for subsequent analysis and processing. The setting of the period should be optimized according to the specific requirements and operating characteristics of the distribution network to balance the frequency of data acquisition and the consumption of system resources.

[0062] The data preprocessing and feature extraction module preprocesses the data obtained from each sub-region and then extracts the key features reflecting the load status of the sub-region. Under the detection window, the extracted key features are quantitatively analyzed and converted into numerical indicators reflecting the current load status;

[0063] The acquired operating data usually contains noise, missing values, and outliers, so preprocessing is required. The preprocessing steps include data denoising, data cleaning, missing value filling, and time synchronization, etc. Through these processes, the quality and consistency of the data are ensured, making the subsequent feature extraction and analysis more accurate and reliable. High-quality data is the basis for intelligent evaluation and fault detection. The preprocessing process may also involve data standardization and normalization to make different features have the same scale and improve the performance of the model.

[0064] Key features reflecting the load status of the sub-region are extracted from the preprocessed data. The extracted features include the reactive power demand within the sub-region and the degree of harmonic distortion in the power system. Under the detection window, the reactive power demand within the sub-region and the degree of harmonic distortion in the power system are analyzed to generate a reactive power load reference value and a total harmonic distortion reference value respectively. The reactive power load reference value quantifies the relationship between the reactive power demand within the sub-region and the total load, and the total harmonic distortion reference value quantifies the ratio of harmonic current to fundamental current in the power system, reflecting the severity of harmonics in the power system.

[0065] A rapid increase in the reactive power demand within the sub-region usually indicates that the current sub-region is in an overloaded state, especially when the load exceeds the normal range. Reactive power is the energy used in the power system to maintain voltage stability and generate electromagnetic fields, and it is mainly consumed by inductive loads (such as motors, transformers, etc.). When the load in the sub-region increases, especially in the case of concentrated or sudden load increases, the reactive power demand will also increase accordingly. When the load is too large, in order to support the operation of the load, the system will require more reactive power to maintain voltage stability and prevent voltage dips or being too low, resulting in abnormal operation of equipment. A rapid increase in reactive power demand is often due to the activation or sudden increase of a large number of inductive loads in the system, exceeding the design capacity or the carrying capacity of the system. This situation will put greater pressure on the power system, possibly leading to severe voltage fluctuations and even voltage collapse. If not adjusted and controlled in time, it may trigger a protective power cut to prevent equipment damage. Therefore, a rapid increase in reactive power demand is a key warning signal, indicating that the current sub-region may have entered an overloaded state, and the stability and reliability of the power system are threatened, requiring further analysis and processing.

[0066] The specific steps for analyzing the reactive power demand within the sub-region under the detection window to generate the reactive power load reference value are as follows:

[0067] First, collect the reactive power demand data within the sub-region through monitoring devices, which involves obtaining the reactive power information related to the sub-region from the real-time monitoring system. The reactive power within the sub-region is calculated through the phase difference between voltage and current, and the calculation expression is as follows:

[0068]

[0069] , where Q is the reactive power of the sub-region, I k is the current at the k-th load point (transformer, motor, etc.) in the sub-region, V k is the voltage at the k-th load point in the sub-region, θ k is the phase difference between the current and voltage at the k-th load point in the sub-region, n is the number of load points in the sub-region, sin(θ k ) is the sine value of the phase difference between the current and voltage;

[0070] The collection of this data provides a basis for subsequent analysis and reflects the reactive power distribution of each load point in the sub-region.

[0071] After collecting the reactive power demand data Q, analyze the dynamic characteristics of the reactive power demand to reveal the change trend of the load in this sub-region. At this time, we cannot simply rely on simple statistics (such as mean or standard deviation), but we need to consider the instantaneous and sudden nature of the load change. Use the exponentially weighted moving average model (EWMA) of the load to emphasize the changes within a time period, and the calculation expression is as follows:

[0072] Q weighted = λ·Q+(1 - λ)·Q previous

[0073] , where Q weighted is the weighted reactive power demand, representing the adjusted evaluation value of the reactive power demand in the current sub-region, considering the past and current reactive power demands. λ is the smoothing factor, which controls the contribution degree of the current value and the historical value to the weighted average result, and its value range is between 0 and 1. Q previous is the reactive power demand value at the previous moment, representing the reactive power demand at the previous time point in the historical data;

[0074] The function of this step is to capture the instantaneous fluctuations of the load by smoothing the recent changes in the reactive power demand, and at the same time avoid the influence of historical data in the far past on the current evaluation, so as to more accurately identify the abnormal state of the load.

[0075] After analyzing the dynamic characteristics of the load, combine the reactive power demand of the sub-region with the load density of the power system in this region to calculate the load deviation and further quantify the abnormal level of the reactive power load. The load deviation calculation expression is as follows:

[0076]

[0077] , where D is the load deviation of the sub-region, A j is the area of the j-th load point in the sub-region, I jis the current demand at the j-th load point within the sub-region;

[0078] The load deviation D reflects the relationship between the load density and the reactive power demand within the sub-region, and can reveal the situation of load concentration or deviation from the normal range. When the D value is high, it indicates that the load is too concentrated or uneven, and there is a risk of overload. This step helps to deeply understand the abnormal characteristics of the reactive power demand.

[0079] The comprehensively weighted reactive power demand Q weighted and the sub-region load deviation D are used to generate the reactive power load reference value, which is used to quantify the load status of the sub-region. The generation formula is as follows:

[0080] Reactive Power = α·Q weighted + β·D

[0081] , where Reactive Power is the reactive power load reference value, α is the weight coefficient of the weighted reactive power demand Q weighted and is used to measure the importance of the weighted reactive power demand Q weighted in the final reference value Reactive Power , β is the weight coefficient of the sub-region load deviation D and measures the importance of the load deviation D in the reactive power load reference value Reactive Power .

[0082] Through this step, the generated reactive power load reference value not only reflects the current reactive power demand level of the sub-region, but also considers the possible impact of the load density on the system stability. When Reactive Power is large, it indicates that the sub-region is in a high overload risk state, and vice versa, it means that the load in this region is relatively stable and the overload risk is small.

[0083] The larger the performance value of the reactive power load reference value generated after analyzing the reactive power demand within the sub-region under the detection window, the greater the load demand within the sub-region, especially the demand for inductive loads, which may have approached or exceeded the carrying capacity of the system. Reactive power is mainly used to maintain voltage stability in the power system. When the load demand in the sub-region increases, the reactive power demand will also increase accordingly, thus increasing the risk of system load and voltage fluctuations. If the reactive power load reference value is high, it usually indicates that the region may be close to the overload state, and the overload state will lead to voltage drop, increased current fluctuations, and may trigger the protective power-off mechanism. Conversely, if the reactive power load reference value is small, it indicates that the load demand in this region is relatively stable and the overload risk is small.

[0084] Severe harmonic distortion in the power system can indeed indicate that the current sub-region is overloaded, especially when the load exceeds the designed carrying capacity of the system. Harmonic distortion is caused by non-linear loads (such as power electronic devices, frequency converters, switched-mode power supplies, etc.) and usually occurs when the load is too heavy or the load distribution is uneven. When the load in a certain sub-region is too large, especially when industrial loads are concentrated or the instantaneous load increases sharply, the proportion of non-linear loads will increase, leading to an exacerbation of harmonic generation, which in turn causes severe harmonic distortion in the power system. During overload, due to the change of the load, the current and voltage waveforms in the power system are distorted, forming high-frequency harmonic components. These harmonics not only affect the stability of the power system and the power quality, but also may cause damage to distribution equipment, such as equipment overheating, insulation aging, etc. The increase in the degree of harmonic distortion is usually accompanied by an abnormal increase in the system load. Therefore, the increase in the total harmonic distortion is often an important signal of system load overload.

[0085] The specific steps to analyze the degree of harmonic distortion in the power system under the detection window to generate the total harmonic distortion reference value are as follows:

[0086] First, perform spectral analysis on the voltage and current signals of the power system through the Fast Fourier Transform (FFT) to extract the harmonic components in the power system. The frequency and amplitude of each harmonic can provide a basis for further analysis. The extraction formula is as follows:

[0087]

[0088] , where X(h) is the amplitude of the h-th harmonic in the frequency domain, called the spectral component. Each X(h) obtained through the FFT transform corresponds to a signal component with a specific frequency. X(h) contains the amplitude and phase information of this frequency component, reflecting the weight and phase difference of this frequency in the original signal. x(m) is the value of the signal at the m-th sampling point, representing the discrete sample value of the time-domain signal of the power system (such as current or voltage signal). e is the natural base, p is the imaginary unit, 2π is a mathematical constant used to calibrate the periodicity of sine waves and cosine waves, h is the frequency index representing the frequency component of the h-th harmonic, m is the sampling point index of the time-domain signal. For each sampling point m, the complex exponential function calculates the contribution of this point to the h-th harmonic, and M is the total number of signal sampling points;

[0089] Through this step, the amplitude value of each harmonic can be accurately obtained, providing the key data required for harmonic distortion analysis.

[0090] After extracting the harmonic component X(h), quantify the distortion degree of each harmonic component. The distortion degree is expressed as the ratio between each harmonic and the fundamental wave (the main frequency component). The formula is as follows:

[0091]

[0092] where HD(h) is the distortion degree of the h-th harmonic, X(1) is the fundamental wave amplitude, that is, the amplitude of the first harmonic. The fundamental wave is usually the signal component in the power system whose frequency is closest to the grid frequency (such as 50 Hz or 60 Hz). ω is an adjustment factor used to weight the distortion degree of higher harmonics;

[0093] The function of this step is to quantify the distortion degree of each harmonic in a non-linear manner, thereby highlighting the impact of high-order harmonics on the power quality of the power system.

[0094] The distortion degree HD(h) of each harmonic is obtained, and the total harmonic distortion reference value is calculated by summarizing the distortion degrees of all harmonic components. The total harmonic distortion reference value is used to measure the harmonic pollution degree of the entire system. To eliminate the bias towards larger harmonic frequencies, the calculation expression is as follows:

[0095]

[0096] where Total Harmonic is the total harmonic distortion reference value, H is the maximum value of the harmonic order, which refers to the maximum harmonic order within the considered harmonic calculation range, τ is a weighting coefficient used to non-linearly weight the influence degree of each harmonic, and γ is a reference value control parameter, which is a parameter for non-linearly controlling the result of the entire harmonic distortion reference value.

[0097] This step enhances the influence of high-frequency harmonics through weighting and non-linearity to reflect the serious impact of high-order harmonics on the power system when the load increases, and further quantifies the harmonic pollution situation of the system more precisely.

[0098] The larger the value of the total harmonic distortion reference value generated after analyzing the harmonic distortion degree in the power system under the detection window, the greater the risk that the current sub-region is in an overloaded state. Harmonic distortion is caused by non-linear loads (such as power electronic devices, frequency converters, etc.). When the load exceeds the design capacity of the power system, the harmonic content in the system will increase significantly, resulting in the distortion of voltage and current waveforms, thus causing power quality problems. A high total harmonic distortion reference value means that there are more harmonic components in the system, which is usually a sign of overloading or imbalance of the load, indicating that the power system is in a stressed state, increasing the risk of faults and equipment damage. Therefore, when the value of the total harmonic distortion reference value is high, it can be determined that the risk of load overload in the current sub-region is relatively large. On the contrary, when the total harmonic distortion is low, it indicates that the system load is relatively balanced and the overload risk is small.

[0099] The intelligent load status evaluation module inputs the quantified features into a pre-trained long short-term memory network to intelligently evaluate the load status of each sub-region through the long short-term memory network;

[0100] Input the quantized reactive power load reference value and total harmonic distortion reference value into a pre-trained long short-term memory network, generate a load fluctuation index through the long short-term memory network, and intelligently evaluate the load status of the sub-region through the load fluctuation index.

[0101] A pre-trained long short-term memory network (Pre-trained Long Short-Term Memory Network, LSTM) refers to a deep learning model obtained by training with a large amount of historical data before the system is put into actual operation. LSTM is a special type of recurrent neural network (RNN) that is particularly good at processing time series data. Its feature is that it can capture long-term dependencies and effectively prevent the problem of gradient vanishing. In an intelligent distribution network monitoring system, LSTM is used to process and analyze various dynamically changing data in the distribution network, such as reactive power load reference values and total harmonic distortion reference values. These data usually have time series characteristics, reflecting the evolution and fluctuations of the system state. Therefore, the LSTM model can use past data information to predict future load changes and anomalies, and then achieve intelligent evaluation of the load status of the sub-region of the distribution network. Through training on historical operation data, LSTM learns the potential laws and complex time series characteristics in the operation of the distribution network, enabling the model to intelligently evaluate the current load status based on new input data (such as reactive power load reference values and harmonic distortion reference values).

[0102] During the training process, the LSTM network continuously adjusts its weights and biases to optimize its prediction ability. Through the input of a large amount of historical data, the model can identify potential patterns in the load changes of each sub-region of the distribution network. For example, whether an increase in reactive power under specific conditions indicates local overload, or whether an increase in the harmonic distortion reference value is related to unbalanced device loads. This learning process based on historical data makes the LSTM network highly adaptable, capable of performing real-time intelligent evaluation of the distribution network according to different time periods and load conditions. In this way, LSTM can generate an index reflecting load fluctuations - the load fluctuation index - by inputting real-time data such as reactive power load reference values and total harmonic distortion reference values. This index can quantify the load status of the current sub-region, helping the system make more accurate judgments and timely responses. Compared with traditional monitoring methods, LSTM, through its powerful time series modeling ability, enables the system to make more accurate and efficient predictions when facing complex and dynamically changing electrical loads, thereby improving the stability and reliability of the distribution network.

[0103] The long short-term memory network is not limited here, and it can achieve inputting the reactive Power power load reference value and the total HarmonicAfter analysis, the load fluctuation index Load is generated Fluctuation Both the long - short - term memory networks of Fluctuation are acceptable. To implement the technical solution of the present invention, the present invention provides a specific implementation method;

[0104] Load fluctuation index Load Fluctuation The generation formula is as follows:

[0105]

[0106] , where in the formula, f 1 、f 2 are the preset proportionality coefficients of the reactive power load reference value Reactive Power and the total harmonic distortion reference value Total Harmonic , and f 1 、f 2 are both greater than 0.

[0107] It can be seen from the load fluctuation index that the larger the value of the reactive power load reference value generated by analyzing the reactive power demand in the sub - region under the detection window, and the larger the value of the total harmonic distortion reference value generated by analyzing the harmonic distortion degree in the power system under the detection window, the larger the value of the load fluctuation index generated by the intelligent evaluation of the load state of the sub - region by the pre - learned long - short - term memory network, indicating that the probability of the current sub - region load being overloaded is greater; conversely, it indicates that the probability of the current sub - region load being overloaded is smaller.

[0108] The preset proportionality coefficients (f 1 and f 2 ) here refer to the important parameters used to balance and weight the influence of different reference values (reactive power load reference value Reactive Power and total harmonic distortion reference value Total Harmonic ) on the load fluctuation index Load Fluctuation . Specifically, the magnitudes of these coefficients reflect the relative importance and contribution ratios of the two reference values in the calculation of the load fluctuation index. In practical applications, the influences of reactive power load and harmonic distortion on load fluctuation may be different. For example, changes in reactive power demand may have a more direct impact on system load fluctuation, while the influence of harmonic distortion may be more reflected in load imbalance under specific conditions. Therefore, f 1 and f 2 are preset weights used to adjust the formula so that it can more accurately reflect the actual situation of the system. These coefficients are usually determined through the analysis of historical data, expert experience or model optimization methods to ensure that the result of the formula can highly coincide with the actual load fluctuation state. Simply put, the preset proportionality coefficient is a way to quantify and weight the importance of each reference value in the process of generating the load fluctuation index.

[0109] The sub-region status division module divides all sub-regions into local voltage abnormal loads and normal regional loads according to the evaluation results of the long short-term memory network, and marks the sub-regions classified as "local voltage abnormal loads" for subsequent key monitoring and processing;

[0110] Compare and analyze the load fluctuation index generated by intelligently evaluating the load status of the sub-region through a pre-trained long short-term memory network with the pre-set reference threshold of the load fluctuation index, and divide the status of the distribution network sub-region. The division steps are as follows:

[0111] If the load fluctuation index is greater than the pre-set reference threshold of the load fluctuation index, the current sub-region is divided into local voltage abnormal loads;

[0112] If the load fluctuation index is less than or equal to the pre-set reference threshold of the load fluctuation index, the current sub-region is divided into normal regional loads.

[0113] Local voltage abnormal load refers to the phenomenon that the voltage deviates from the normal range due to reasons such as excessive load or uneven distribution in the distribution network sub-region; normal regional load means that the load operation state in the distribution network sub-region is stable, and various electrical parameters (such as voltage, current, reactive power, and harmonics, etc.) fluctuate within the design range.

[0114] The function of marking the sub-regions classified as "local voltage abnormal loads" is to accurately identify the problem areas from the entire distribution network for subsequent key monitoring and rapid processing. Through marking, the system can clearly identify which sub-regions have voltage abnormalities or overload problems, and preferentially allocate resources to collect dynamic data, conduct high-frequency monitoring, and perform fault diagnosis on these areas, timely discover the root cause of the problem and take targeted measures (such as load transfer, equipment maintenance, or reactive power compensation, etc.). This marking mechanism can not only improve the monitoring efficiency, avoid the waste of network-wide monitoring resources, but also effectively prevent local problems from expanding into network-wide failures, and ensure the overall stability and power supply quality of the distribution network.

[0115] The conventional monitoring module continues to perform conventional monitoring on the sub-regions classified as normal regional loads with a pre-set periodic operation state scan to ensure the stable operation of the overall distribution network;

[0116] For the sub-regions classified as having normal regional loads, continue with routine monitoring through periodic operation status scanning as preset. The main purpose is to maintain the continuous and stable operation of the distribution network and ensure the timely detection of potential load fluctuations or system anomalies. Even if these sub-regions appear normal under the current load, regular monitoring is still necessary because the power load and system status are dynamically changing, and sudden increases in load or unexpected failures may lead to the instability of the power system. Through periodic scanning, the operation status of each sub-region can be continuously tracked, and possible abnormal fluctuations, such as slight voltage fluctuations, load imbalance, or equipment operation deviations, can be captured in a timely manner. This not only helps prevent potential overload risks but also provides real-time operation data for the system, supporting more accurate dispatching and management to ensure that the overall distribution network is always in a safe and stable working state.

[0117] The dynamic adjustment and anomaly response module dynamically adjusts the data acquisition frequency for sub-regions marked as having local voltage abnormal loads to more accurately identify and respond to abnormal states;

[0118] The specific steps for dynamically adjusting the data acquisition frequency for sub-regions marked as having local voltage abnormal loads to more accurately identify and respond to abnormal states are as follows:

[0119] When the load fluctuation index Load Fluctuation is greater than the reference threshold of the load fluctuation index, dynamically adjust the data acquisition frequency to ensure that potential anomalies can be more accurately captured when there are significant load fluctuations in the sub-region. The frequency adjustment formula is as follows:

[0120]

[0121] , where f new is the adjusted data acquisition frequency, f base is the original periodic operation status scanning frequency, ρ is the adjustment coefficient used to control the influence degree of the load fluctuation index on the data acquisition frequency, which determines how the frequency is adjusted when the load fluctuation intensity increases. Load Fluctuation is the load fluctuation index, Load ref is the reference threshold of the load fluctuation index, is the exponential adjustment factor of the load fluctuation intensity used to control the non-linear influence of the load fluctuation index Load Fluctuation on the adjustment of the data acquisition frequency. Δt avg is the average duration of the current load fluctuation, Δt ref is the reference load fluctuation duration, usually set based on historical data and the operating characteristics of the power grid, used to standardize the influence of the load fluctuation duration. μ is the adjustment coefficient that controls the influence of the load fluctuation duration on the data acquisition frequency;

[0122] In the step, indicates how the intensity of load fluctuations is weighted to affect the acquisition frequency; while considers the time characteristics of load fluctuations. When the intensity of load fluctuations is large and the fluctuation duration is long, the data acquisition frequency will increase significantly, thereby improving the system's response ability to local voltage abnormal loads.

[0123] The adjusted data acquisition frequency f new will be further optimized based on the feedback mechanism to adapt to real-time load changes. Especially when frequent load fluctuations or abnormal states cannot be accurately captured, an adaptive gain function can be introduced to adjust the update rate of the scanning frequency, making the frequency change more flexible and efficient. The formula is as follows:

[0124]

[0125] , where f adjusted is the optimized data acquisition frequency, δ is the dynamic adjustment coefficient used to control the sensitivity of the dynamic adjustment frequency, Load avg is the average value of the load fluctuation index, ζ is the

[0126] exponential weight of the load fluctuation index, used to control the exponential coefficient of the influence degree on the frequency adjustment, n missed is the number of undetected anomalies, indicating the number of abnormal events that cannot be detected in time during the current sub-region monitoring process, N max is the maximum allowable number of undetected anomalies, indicating the maximum allowable number of abnormal undetected times set by the system, used to standardize the undetected situation, and θ is the abnormal undetected weight index, used to control the influence degree on the adjusted frequency.

[0127] Through this feedback mechanism, when load fluctuations are frequent and the abnormal state cannot be successfully captured, the system will automatically increase the data acquisition frequency, thereby improving the recognition accuracy of anomalies. If the abnormal state has been effectively monitored, the increase in frequency will be suppressed to avoid excessive resource waste and system burden. This dynamic adjustment mechanism can respond more precisely to load changes and ensure the effective monitoring of the system in various complex situations.

[0128] Steps for dynamically adjusting the data acquisition frequency for sub-regions marked as locally voltage-abnormal loads aim to monitor local voltage abnormalities with higher precision and real-time by encrypting the data acquisition frequency, so as to promptly identify and respond to possible overloads or other abnormal conditions. In the operation of a distribution network, local voltage abnormalities are often precursors to load overloads or voltage fluctuations caused by factors such as sudden load increases and power system imbalances. When voltage abnormalities occur in certain areas of the distribution network, conventional periodic data acquisition may not be able to promptly capture sudden short-term voltage fluctuations and rapidly changing load states, resulting in delayed detection of abnormal conditions and affecting the system's response and handling.

[0129] By dynamically adjusting the data acquisition frequency, the monitoring precision and frequency can be flexibly adjusted according to the current grid state and load changes in the sub-region. For example, when a locally voltage-abnormal load is detected, the system will identify this abnormal signal and automatically increase the data acquisition frequency to more finely track key parameters such as voltage fluctuations, current changes, and power factors. This high-frequency data acquisition can not only capture instantaneous voltage deviations in real time but also provide more detailed load fluctuation data, helping the system quickly determine whether the load has entered the overload state and avoiding missing abnormal moments due to too long a sampling period.

[0130] In addition, dynamically adjusting the data acquisition frequency also helps the system better predict the load change trend and take control measures in advance, such as adjusting the voltage, distributing the load, or enabling standby equipment, to ensure that the power system can quickly respond when facing local overloads or voltage problems. This can effectively reduce the failure rate of the distribution network, lower the power outage risk, improve the stability and security of the system, and ensure that the power supply to users is not severely affected. Through this flexible monitoring strategy, the distribution network can continuously optimize its operation efficiency in a dynamic environment, ensuring power supply quality and high reliability of the system.

[0131] The present invention divides the distribution network into multiple sub-areas, and the monitoring system can monitor and manage each sub-area in detail to ensure that the load situation in each area is tracked independently and in real time. Unlike the traditional monitoring system based on the average data of the whole network, this sub-area division can effectively identify local overload or abnormal load, especially when the load suddenly increases or the local equipment fails, and can capture abnormal fluctuations in time. Key features such as reactive power demand and total harmonic distortion can reflect changes in load imbalance, equipment abnormality or power quality problems. The system obtains data in real time through periodic operation status scanning, and responds quickly when an abnormality is found, avoiding local overload from causing large-scale power system failures. In addition, for the overload situation of certain specific sub-areas, the system will dynamically adjust the data acquisition frequency to achieve more accurate abnormality identification. Through this precise monitoring and response mechanism, the stability of the distribution network can be significantly improved, the occurrence of voltage fluctuations, power outages and system collapse can be reduced, and the reliability and safety of power supply can be ensured.

[0132] The present invention introduces a long short-term memory network (LSTM) to make the distribution network monitoring system more intelligent and adaptive. LSTM can analyze historical data and capture the time series characteristics of load fluctuations, thereby realizing accurate prediction and intelligent evaluation of load status. The system can not only monitor the load conditions of sub-areas in real time, but also automatically learn and adapt to load changes, seasonal fluctuations and emergencies. After the reactive power load reference value and the total harmonic distortion reference value are input into the LSTM network, the system can generate a load fluctuation index, which provides a dynamic evaluation of the load state and can be compared and analyzed with the preset threshold. Once the load fluctuation index exceeds the preset threshold, the system automatically determines the local voltage abnormal load, marks the area in real time and conducts key monitoring. This adaptive intelligent analysis not only reduces the need for human intervention, but also adjusts the monitoring strategy according to different changes in power demand. Through this intelligent processing, the distribution network can automatically adjust under changing load conditions, significantly improving the flexibility, stability and fault response capabilities of the system, and ultimately ensuring the efficiency and safety of power supply.

[0133] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technicians in this field according to actual conditions.

[0134] The above description is only by way of illustration of certain exemplary embodiments of the present invention. It is undoubted that those skilled in the art can modify the described embodiments in various ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the scope of protection of the claims of the present invention.

[0135] It should be noted that in this text, if there are relational terms such as first and second, etc., they are only used to distinguish one entity or operation from another entity or operation, and do not necessarily require or imply any actual relationship or order between these entities or operations. Moreover, the terms "comprising", "including" or any other variants thereof are intended to cover non-exclusive inclusion, so that a process, method, article or device comprising a series of elements not only includes those elements, but also includes other elements not expressly listed, or also includes elements inherent to such process, method, article or device. Without further limitation, an element defined by the statement "comprising one..." does not exclude the existence of additional identical elements in the process, method, article or device comprising the element.

[0136] It should be understood that in various embodiments of the present application, the magnitude of the sequence numbers of the above processes does not mean the order of execution is prior or subsequent, and the execution order of each process should be determined by its function and internal logic, and should not constitute any limitation to the implementation process of the embodiments of the present application.

[0137] Those of ordinary skill in the art can realize that the units and algorithm steps of each example described in combination with the embodiments disclosed herein can be implemented by electronic hardware, or a combination of computer software and electronic hardware. Whether these functions are executed in a hardware or software manner depends on the specific application and design constraints of the technical solution. Professional technicians can use different methods to implement the described functions for each specific application, but such implementation should not be considered to exceed the scope of the present application.

[0138] 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 foregoing method embodiments, and will not be repeated herein.

[0139] The unit described as a separated component may or may not be physically separated, and the component shown as a unit may or may not be a physical unit, that is, it may be located in one place, or may be distributed to multiple network units. Some or all of the units can be selected according to actual needs to achieve the purpose of the solution of this embodiment.

[0140] In addition, in each embodiment of the present application, the functional units can be integrated into one processing unit, or each unit exists physically alone, or two or more units are integrated into one unit.

[0141] As described above, this is only the specific implementation manner of the present application. However, the protection scope of the present application is not limited thereto. Any person skilled in the art within the technical scope disclosed by the present application can easily think of changes or substitutions, which should all be covered within the protection scope of the present application. Therefore, the protection scope of the present application shall be subject to the protection scope of the claims described above.

[0142] Only some exemplary embodiments of the present invention have been described by way of illustration above. Undoubtedly, for those of ordinary skill in the art, the described embodiments can be modified in various different ways without departing from the spirit and scope of the present invention. Therefore, the above drawings and descriptions are illustrative in nature and should not be construed as limiting the protection scope of the claims of the present invention.

Claims

1. Intelligent distribution network monitoring and fault detection system, characterized in that: It includes distribution network area division module, periodic data collection and monitoring module, data preprocessing and feature extraction module, intelligent load state evaluation module, sub-area state division module, conventional monitoring module and dynamic adjustment and abnormal response module: The distribution network area division module divides the entire distribution network into multiple sub-areas to achieve more refined monitoring and management; The periodic data collection and monitoring module acquires the operation data of each sub-area of ​​the distribution network by implementing periodic operation status scanning, and continuously tracks the operation status of the distribution network; The data preprocessing and feature extraction module preprocesses the data obtained from each sub-area, extracts the key features reflecting the load status of the sub-area, and performs quantitative analysis on the extracted key features under the detection window to convert them into numerical indicators reflecting the current load status; The intelligent load state evaluation module inputs the quantified features into the pre-learned long short-term memory network, and uses the long short-term memory network to intelligently evaluate the load state of each sub-area; The sub-region status division module divides all sub-regions into local voltage abnormal load and regional load normal according to the evaluation results of the long short-term memory network, and marks the sub-regions divided into "local voltage abnormal load" for subsequent key monitoring and processing; The conventional monitoring module continues to perform conventional monitoring with preset periodic operation status scanning for sub-areas that are classified as having normal regional loads, to ensure the stable operation of the overall distribution network; The dynamic adjustment and abnormal response module dynamically adjusts the data acquisition frequency for sub-areas marked as local voltage abnormal loads to more accurately identify and respond to abnormal conditions.

2. The intelligent distribution network monitoring and fault detection system according to claim 1, characterized in that: The key features reflecting the load status of the sub-area are extracted from the preprocessed data. The extracted features include the reactive power demand in the sub-area and the degree of harmonic distortion in the power system. Under the detection window, the reactive power demand in the sub-area and the degree of harmonic distortion in the power system are analyzed to generate a reactive power load reference value and a total harmonic distortion reference value respectively. The reactive power load reference value quantifies the relationship between the reactive power demand in the sub-area and the total load. The total harmonic distortion reference value quantifies the ratio of harmonic current to fundamental current in the power system, reflecting the severity of the harmonics in the power system.

3. The intelligent distribution network monitoring and fault detection system according to claim 2, characterized in that: The specific steps for analyzing the reactive power demand in the sub-area under the detection window to generate the reactive power load reference value are as follows: First, the reactive power demand data in the sub-area is collected through monitoring equipment. The reactive power in the sub-area is calculated by the phase difference between voltage and current. The calculation expression is as follows: ， Where Q is the reactive power of the sub-area, I k is the current of the kth load point in the sub-region, V k is the voltage of the kth load point in the sub-region, θ k is the current and voltage phase difference of the kth load point in the sub-region, n is the number of load points in the sub-region, sin(θ k ) is the sine value of the phase difference between current and voltage; After collecting the reactive power demand data Q, the dynamic characteristics of the reactive power demand are analyzed to reveal the changing trend of the load in the sub-area. The calculation expression is as follows: Q weighted =λ·Q+(1-λ)·Q previous , In the formula, Q weighted is the weighted reactive power demand, λ is the smoothing factor, which controls the contribution of current value and historical value to the weighted average result, Q previous is the reactive power demand value at the previous moment; After analyzing the dynamic characteristics of the load, the reactive power demand of the sub-region is combined with the load density of the power system in the region to calculate the load deviation. The load deviation calculation expression is as follows: ， Where D is the sub-region load deviation, A j is the area of ​​the jth load point in the sub-region, I j is the current demand of the jth load point in the sub-area; Comprehensive weighted reactive power demand Q weighted The reactive power load reference value generated by the sub-area load deviation D is generated as follows: Reactive Power =α·Q weighted +β·D, In the formula, Reactive Power is the reactive power load reference value, α is the weighted reactive power demand Q weighted is the weight coefficient of the sub-region load deviation D.

4. The intelligent distribution network monitoring and fault detection system according to claim 2, characterized in that: The specific steps for analyzing the harmonic distortion level in the power system under the detection window to generate the total harmonic distortion reference value are as follows: First, the voltage and current signals of the power system are analyzed by fast Fourier transform to extract the harmonic components in the power system. The extraction formula is as follows: ， Where X(h) is the amplitude of the hth harmonic in the frequency domain, called the spectral component, x(m) is the value of the signal at the mth sampling point, e is the natural base, p is the imaginary unit, 2π is a mathematical constant used to calibrate the periodicity of sine and cosine waves, h is the frequency index, indicating the frequency component of the hth harmonic, m is the sampling point index of the time domain signal, and M is the total number of signal sampling points; After extracting the harmonic component X(h), the distortion of each harmonic component is quantified, and the formula is as follows: ， Where HD(h) is the distortion of the hth harmonic, X(1) is the fundamental amplitude, and ω is the adjustment factor; The distortion of each harmonic HD(h) is obtained, and the distortion of all harmonic components is summarized to calculate the reference value of total harmonic distortion. The calculation expression is as follows: ， In the formula, Total Harmonic is the total harmonic distortion reference value, H is the maximum value of the harmonic order, τ is the weighting coefficient, and γ is the reference value control parameter.

5. The intelligent distribution network monitoring and fault detection system according to claim 2, characterized in that: The quantized reactive power load reference value and total harmonic distortion reference value are input into the pre-learned long short-term memory network, and the load fluctuation index is generated by the long short-term memory network. The load status of the sub-area is intelligently evaluated by the load fluctuation index.

6. The intelligent distribution network monitoring and fault detection system according to claim 5, characterized in that: The load fluctuation index generated by intelligently evaluating the load status of the sub-area through the pre-learned long short-term memory network is compared and analyzed with the pre-set load fluctuation index reference threshold, and the distribution network sub-area status is divided. The division steps are as follows: If the load fluctuation index is greater than a preset load fluctuation index reference threshold, the current sub-area is divided into a local voltage abnormal load; If the load fluctuation index is less than or equal to a preset load fluctuation index reference threshold, the current sub-area is classified as having normal regional load.

7. The intelligent distribution network monitoring and fault detection system according to claim 6, characterized in that: For sub-areas marked as local voltage abnormal loads, the specific steps for dynamically adjusting the data acquisition frequency to more accurately identify and respond to abnormal conditions are as follows: When the load fluctuation index Load FluctuNtion When the load fluctuation index is greater than the reference threshold, the data acquisition frequency is dynamically adjusted to ensure that potential anomalies can be captured more accurately when a drastic load fluctuation occurs in the sub-area. The frequency adjustment formula is as follows: ， In the formula, f new is the adjusted data collection frequency, f [Nse is the original periodic operation status scanning frequency, ρ is the adjustment coefficient, Load FluctuNtion is the load fluctuation index, Load ref is the load fluctuation index reference threshold, is the exponential adjustment factor of the load fluctuation intensity, Δt Nvg is the average duration of the current load fluctuation, Δt ref is the reference load fluctuation duration, μ is the adjustment factor; Adjusted data acquisition frequency f new Further optimization will be performed based on the feedback mechanism to adapt to real-time load changes, making frequency changes more flexible and efficient. The formula is as follows: ， In the formula, f Ndjusted is the optimized data collection frequency, δ is the dynamic adjustment coefficient, Load Nvg is the average value of the load fluctuation index, ζ is the exponential weight of the load fluctuation index, N missed is the number of undetected anomalies, N max is the maximum number of missed anomalies allowed, and θ is the weight index of missed anomalies.

Images