FTU-based intelligent power distribution system monitoring method and monitoring device

Abstract

The invention relates to the technical field of intelligent power distribution system monitoring, and discloses an FTU-based intelligent power distribution system monitoring method and monitoring device. The method comprises the following steps: processing current signal data acquired by FTU equipment to obtain amplitude information and phase information, and determining correlation characteristics of amplitude attenuation and propagation distance; extracting spectrum distribution change data based on the association features, performing multi-scale decomposition, and judging a coupling relationship between spectrum distribution change and amplitude attenuation and a dynamic mapping relationship between phase offset data and a propagation distance; generating multi-dimensional feature data based on the association features, the coupling relationship and the dynamic mapping relationship, inputting the multi-dimensional feature data to the joint distribution model, and performing calibration processing on the joint distribution model to obtain an optimized distance relationship model; and the FTU equipment collects dynamic fluctuation data of the current signal and inputs the dynamic fluctuation data to the distance relation model to obtain an accurate positioning result of the fault position. The positioning accuracy of the fault position of the power distribution system is improved.

Description

Technical Field

[0001] This application relates to the field of intelligent monitoring technology for power distribution systems, and in particular to an intelligent power distribution system monitoring method and monitoring device based on FTU. Background Technology

[0002] With rapid socio-economic development and accelerated urbanization, electricity demand is constantly increasing, making the stability and reliability of power distribution systems crucial for power supply. Traditional power distribution system monitoring methods rely on manual inspections and periodic maintenance, which are not only inefficient but also difficult to achieve real-time monitoring and fault early warning of power distribution equipment. Furthermore, with the increasing complexity of power distribution networks, traditional monitoring methods can no longer meet the demands of modern intelligent and automated power distribution systems.

[0003] Intelligent power distribution systems integrate advanced information technology, communication technology, and automation technology to achieve real-time monitoring, remote control, and intelligent analysis of power distribution equipment. Among these systems, the feeder terminal unit (FTU) is a key component, responsible for collecting operational data from power distribution lines, such as voltage, current, and power, and possessing remote communication and control capabilities.

[0004] The application of FTUs provides a foundation for the intelligentization of power distribution systems, but their monitoring methods still face the following problems and challenges: 1. Incomplete data acquisition: Existing FTU monitoring methods often only focus on some key parameters, ignoring other factors that may affect the stability of the power distribution system, such as ambient temperature and humidity; 2. Insufficient data processing capabilities: FTUs collect a huge amount of data, and traditional data processing methods struggle to achieve rapid analysis and effective utilization of this data; 3. Inaccurate fault warnings: Due to a lack of in-depth analysis and learning of historical data, existing FTU monitoring methods suffer from false alarms and missed alarms in fault warnings; 4. Inflexible remote control: Existing FTU monitoring systems often adopt preset control strategies for remote control, lacking the ability to dynamically adjust control parameters based on real-time data; 5. Low system integration: Existing FTU monitoring systems often operate independently, making it difficult to effectively integrate with other intelligent power distribution equipment or systems.

[0005] To address the aforementioned problems, this invention proposes an intelligent power distribution system monitoring method based on FTU, aiming to achieve comprehensive, accurate, and real-time monitoring and management of the power distribution system by improving key technologies such as data acquisition, data processing, fault early warning, and remote control. Summary of the Invention

[0006] To address the aforementioned technical issues, this application provides an intelligent power distribution system monitoring method and device based on FTU, which improves the monitoring accuracy of fault locations in power distribution systems.

[0007] In a first aspect, this application provides a monitoring method for an intelligent power distribution system based on an FTU (Fault-to-Unit) device. Multiple FTU devices are installed at different locations within the power distribution system to acquire current signal data when a fault occurs in the power distribution system. The method includes: Step S1: Perform time-domain and frequency-domain separation processing on the current signal data to obtain amplitude information and phase information. Analyze the amplitude attenuation law based on the amplitude information to determine the correlation characteristics between amplitude attenuation and propagation distance. The propagation distance is the line distance between the current signal data and the FTU device from the fault location. Step S2: Extract spectral distribution change data based on the associated features, perform multi-scale decomposition on the spectral distribution change data, determine the coupling relationship between spectral distribution change and amplitude attenuation, analyze the phase shift influence based on the phase information and the coupling relationship, and obtain the dynamic mapping relationship between phase shift data and propagation distance; Step S3: Generate multi-dimensional feature data based on correlation features, coupling relationships and dynamic mapping relationships, establish a joint distribution model, determine the comprehensive relationship between the multi-dimensional feature data and the propagation distance, obtain the line condition data and environmental adaptation data of the power distribution system, calibrate the joint distribution model, and obtain a distance relationship model adapted to complex scenarios; Step S4: The FTU device collects dynamic fluctuation data of real-time current signal, inputs it into the distance relationship model, and combines it with background data from complex scenario analysis to obtain the accurate location result of the fault.

[0008] Compared with the prior art, the beneficial effects of the present invention are at least as follows: This application separates the time and frequency domains to obtain a raw dataset containing amplitude and phase information, ensuring the accuracy of signal acquisition and providing a high-quality data foundation for subsequent amplitude variation, spectral distribution, and phase shift analysis. By calculating the trend curve of amplitude variation over time and employing empirical mode decomposition (EMD) technology to decompose amplitude variation into multiple intrinsic mode functions, the amplitude attenuation law related to propagation distance is obtained, accurately capturing the characteristics of amplitude attenuation, avoiding the loss of details caused by global averaging, and improving the accuracy of the correlation between amplitude attenuation and propagation distance. Through frequency domain transformation, the spectral distribution is decomposed into multiple scales to analyze the coupling relationship between spectral variation and amplitude attenuation. Multi-scale decomposition ensures that the signal variation characteristics in different frequency ranges can be fully captured, enhancing the robustness of signal characteristics and thus improving the accuracy of fault location.

[0009] This application also utilizes phase difference calculation technology to quantify phase offset, obtaining a dynamic mapping between phase offset and propagation distance. Accurate phase offset calculation, combined with the dynamic mapping of propagation distance, enables precise fault location estimation. Multiple feature data, such as amplitude attenuation, spectral distribution changes, and phase offset, are integrated into a comprehensive analysis framework. A joint distribution model of multi-dimensional features and propagation distance is constructed. By combining multiple features, errors that may arise from single-feature analysis are avoided, enhancing the model's accuracy and robustness. By fusing line conditions and environmental adaptation data and calibrating model parameters, a distance relationship model adapted to complex scenarios is optimized. Real-time calibration enables the model to adapt to various environmental changes and power grid load fluctuations, improving the accuracy and flexibility of fault location. Based on the distance relationship model obtained from previous analysis, dynamic signal changes are processed in real time. If the signal change exceeds a preset threshold, model parameters are adjusted to estimate the fault location. Through real-time adjustment and iterative optimization, the model can cope with signal fluctuations in complex power grid environments, ultimately accurately determining the fault location. Attached Figure Description

[0010] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the description of the embodiments will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0011] Figure 1 This is a schematic diagram of one embodiment of the intelligent power distribution system monitoring method based on FTU in this application. Figure 2 This is a flowchart illustrating the optimization of dynamic mapping results in the embodiments of this application; Figure 3 This is a comparison diagram of fault location errors in the power distribution system in the embodiments of this application; Figure 4 This is a schematic diagram of one embodiment of the intelligent power distribution system monitoring device based on FTU in this application. Detailed Implementation

[0012] This application provides a monitoring method and device for an intelligent power distribution system based on an FTU. The terms "first," "second," "third," "fourth," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments described herein can be implemented in a sequence other than that illustrated or described herein. Furthermore, the terms "comprising" or "having" and any variations thereof are intended to cover a non-exclusive inclusion; for example, a process, method, system, product, or device that includes a series of steps or units is not necessarily limited to those steps or units explicitly listed, but may include other steps or units not explicitly listed or inherent to such processes, methods, products, or devices.

[0013] For ease of understanding, the specific process of the embodiments of this application is described below. Please refer to [link / reference]. Figure 1 One embodiment of the intelligent power distribution system monitoring method based on FTU in this application includes: Step S1: Perform time-domain and frequency-domain separation processing on the current signal data to obtain amplitude and phase information. Analyze the amplitude attenuation law based on the amplitude information to determine the correlation characteristics between amplitude attenuation and propagation distance. The propagation distance is the line distance between the current signal data transmitted from the fault location to the FTU device.

[0014] In modern power distribution systems, FTU (Feeder Terminal Unit) devices are widely used to monitor the operating status of the power system in real time. In complex power systems, signal propagation is affected by various factors, such as line length, changes in electrical load, and weather conditions. Furthermore, in the distribution network, the propagation rate and delay of the signal may vary due to different FTU device locations, factors such as line length and load fluctuations. When a fault occurs, the phase of the current signal changes, and this phase shift reflects the delay during propagation. Existing fault location methods mainly rely on setting up multiple sensors in the distribution network and calculating the fault location through models. However, these methods often ignore the spatiotemporal differences in signal propagation and the influence of different electrical environments on the signal. Therefore, accurately calculating the distance of the signal propagation path and the signal attenuation characteristics during propagation becomes crucial for improving fault detection accuracy.

[0015] Specifically, multiple FTU devices are installed at key nodes in the intelligent power distribution system, such as substation outlets, relay points of distribution lines, and feeder terminals of important loads. These FTU devices at key nodes monitor current signal data in real time, performing time-domain and frequency-domain separation processing on the collected current signal data. Time-domain analysis obtains the amplitude information of the current signal as it changes over time through sampling. The amplitude information represents the strength of the current signal and is the basic data for judging signal attenuation in fault detection. Frequency-domain analysis converts the time-domain signal to the frequency domain using Fourier transform or other frequency-domain transformation methods (such as wavelet transform) to obtain phase information, which represents the phase changes of the current signal. By analyzing these phase changes, the propagation delay and propagation path of the signal can be determined, thus providing a basis for subsequent fault location.

[0016] In power systems, signals undergo attenuation as they travel from the fault location to the FTU (Fault Transfer Unit). This amplitude attenuation is typically caused by multiple factors, including line resistance, load fluctuations, line length, and signal frequency components. Therefore, by analyzing the correlation between amplitude attenuation and propagation distance, the propagation distance can be inferred from the amplitude changes of the current signal. This application analyzes amplitude attenuation based on amplitude information. The amplitude attenuation pattern is quantified by comparing the ratio of the initial amplitude to subsequent amplitudes, thereby identifying a gradual change in amplitude from high to low. To further understand the relationship between amplitude attenuation and propagation distance (the distance along the line from the fault source to the FTU), this application uses methods such as least squares to fit the relationship between the attenuation slope and the propagation distance, obtaining correlation characteristics. Detailed analysis methods will be explained later. Through separate processing in the time and frequency domains, the FTU can extract the amplitude and phase information of the signal separately, laying a solid data foundation for subsequent signal analysis and fault location.

[0017] Step S2: Extract spectral distribution change data based on correlation features, perform multi-scale decomposition on the spectral distribution change data, determine the coupling relationship between spectral distribution change and amplitude attenuation, analyze the phase shift influence based on phase information and coupling relationship, and obtain the dynamic mapping relationship between phase shift data and propagation distance.

[0018] Specifically, spectral distribution variation data is identified from the correlation characteristics of amplitude attenuation. This variation data reflects the differences in signal energy distribution at different frequencies. For the spectral distribution feature data, a frequency domain transformation method is used for multi-scale decomposition to obtain spectral distribution variation data at different scales. By comparing and analyzing the spectral distribution variation data with the amplitude attenuation characteristics, the coupling relationship between the two is determined. If the spectral distribution variation data and the amplitude attenuation characteristics show a consistent trend, the coupling relationship is considered valid. Since changes in spectral distribution affect phase, phase information needs to be analyzed in the frequency domain. By comparing the phase differences of different frequency components, the phase shift of the signal is analyzed. Based on the coupling relationship and phase information, a dynamic mapping relationship between phase shift data and propagation distance is established through regression analysis or other data fitting methods, which will be explained in detail later. The dynamic mapping relationship can be used to estimate the signal propagation distance, especially in complex power grid environments where the signal propagation distance may change due to various factors. Therefore, the dynamic mapping relationship system can calculate the propagation distance in real time based on phase information, thereby improving the accuracy of fault location.

[0019] Step S3: Generate multi-dimensional feature data based on correlation features, coupling relationships and dynamic mapping relationships, establish a joint distribution model, determine the comprehensive relationship between multi-dimensional feature data and propagation distance, obtain line condition data and environmental adaptation data of the power distribution system, calibrate the joint distribution model, and obtain a distance relationship model adapted to complex scenarios.

[0020] Specifically, in long-distance and complex environments, spectral variations can provide additional clues to help confirm whether the signal propagation process is as expected; phase shift not only provides information on the distance of signal propagation but also reveals the specific impact of the electrical environment on the signal during propagation. By analyzing the dynamic mapping relationship between phase shift and propagation distance, the location of the fault source can be accurately estimated. Therefore, by combining the correlation characteristics between amplitude attenuation and propagation distance, the coupling relationship between spectral distribution changes and amplitude attenuation, and the dynamic mapping relationship between phase shift data and propagation distance, multi-dimensional feature data is generated. The joint distribution model describes the joint probability distribution between multi-dimensional features (amplitude attenuation, spectral distribution, and phase shift) and propagation distance using statistical methods. Common joint distribution models include Gaussian mixture models (GMMs), and the model outputs the probability distribution of propagation distance.

[0021] However, the line conditions of the power distribution system (such as line length, conductor type, and electrical characteristics) and environmental adaptation data (such as temperature, humidity, and load changes) can affect signal propagation. Therefore, these external factors need to be incorporated into the model calibration process. How to calibrate will be explained in detail later. The calibrated model can better adapt to changes in different power grid environments and provide more accurate propagation distance estimates. This model can dynamically adjust under different environmental conditions (such as load fluctuations and environmental changes), improving positioning accuracy. Through a joint distributed model, the system can more accurately predict the relationship between phase offset and propagation distance, reducing errors caused by changes in the external environment.

[0022] Step S4: The FTU device collects dynamic fluctuation data of real-time current signal, inputs it into the distance relationship model, and combines it with background data from complex scenario analysis to obtain the accurate location of the fault.

[0023] Specifically, the FTU device monitors current fluctuations in real time, capturing changes in the current signal and reflecting short-term fluctuations, interference, or fault events occurring in the system. For example, a short-circuit fault or equipment failure may cause a sharp fluctuation in the current signal. The collected dynamic fluctuation data is used as input to a previously established distance relationship model. This model, trained based on historical data and correlation features, can map the current current signal fluctuation to the propagation distance from the fault source to the FTU device, thereby determining the location of the fault.

[0024] Background data includes line topography and weather records. Line topography refers to the geographical and physical features along the power distribution line, including terrain undulations, slope, obstacles, urban density, and ground materials. These features affect the propagation speed, attenuation, and signal quality of the current signal. Weather records include meteorological data such as temperature, humidity, wind speed, and precipitation. These factors affect the electrical characteristics of the power distribution line and the signal propagation characteristics. By fusing the line topography and weather record data with real-time current signal data collected by the FTU equipment, and using this data as input to the distance relationship model, the model adjusts its parameters based on the line topography and weather data, thereby adjusting the model's prediction results in real time and reducing errors caused by environmental changes. By incorporating background data such as line topography and weather records into the fault location model, the signal propagation estimation process is further optimized, enhancing the model's accuracy and robustness, especially in complex and variable environments, enabling more precise fault location.

[0025] In one specific embodiment, determining the correlation characteristics between amplitude attenuation and propagation distance in step S1 specifically includes the following steps: Step S11: Set a fixed duration to divide the amplitude information within the target duration into multiple target windows, calculate the average amplitude within each target window and compare them to obtain the first change curve representing amplitude decay; Step S12: Identify the inflection point of the first change curve, and use empirical mode decomposition to perform time-domain signal decomposition on the first change curve based on the inflection point to obtain multiple intrinsic mode functions; Step S13: Calculate the energy spectrum of each intrinsic mode function, obtain the attenuation slope of each curve segment, fit the attenuation slope with the preset propagation distance parameter, and determine the correlation characteristics.

[0026] Specifically, the signal amplitude information within the target duration is divided into multiple target time windows, each with a pre-set duration (e.g., 10 milliseconds). The average amplitude is calculated within each window to obtain the representative amplitude value for that time window. The average amplitude values ​​of each window are connected to form a change curve. The first change curve represents the attenuation process of the signal amplitude over time.

[0027] Inflection points are identified by calculating the first or second derivative of the first variation curve. Inflection points are crucial markers in the signal attenuation process; when the rate of amplitude attenuation changes significantly, the slope of the curve changes, forming an inflection point. Identifying these inflection points allows for further analysis of the signal's attenuation characteristics at different stages, such as rapid and slow attenuation phases. Based on these inflection points, Empirical Mode Decomposition (EMD) is used to decompose the first variation curve. EMD, a time-domain signal decomposition technique, decomposes the first variation curve into multiple Intrinsic Mode Functions (IMFs). Each IMF represents an attenuation stage; for example, in long-distance transmission lines, the first function corresponds to rapid attenuation near the end, and the second to slow attenuation far from the end. This decomposition helps reveal the intrinsic structure of amplitude changes and improves the reliability of the correlation with propagation distance. For each IMF obtained from the decomposition, the energy spectrum of each function is calculated to quantify the attenuation intensity of each segment, i.e., the attenuation slope. A linear regression model is then used to fit the attenuation slope with preset propagation distance parameters to determine correlation characteristics. For example, the correlation between the attenuation slope k and the transmission distance d is determined through an exponential function. Let be the constants that are fitted based on historical data, so as to obtain the quantitative relationship between amplitude attenuation and propagation distance.

[0028] In one specific embodiment, determining the coupling relationship between the spectral distribution change and the amplitude attenuation in step S2 specifically includes the following steps: Step S211: Convert the spectral distribution variation data into a frequency domain representation using a frequency domain transformation method; Step S212: Perform multi-scale decomposition on the transformed spectral distribution change data to obtain multi-scale decomposition results including low-frequency and high-frequency components; Step S213: Based on the multi-scale decomposition results, calculate the energy ratio of the spectral distribution change data at each scale, and perform correlation analysis with the associated features to obtain the correlation coefficient; Step S214: If the correlation coefficient is greater than the preset first threshold, it is determined that there is a strong coupling relationship between the spectrum distribution change data and the amplitude attenuation; otherwise, it is a weak coupling relationship.

[0029] Specifically, for spectral distribution characteristic data, frequency domain transformation methods such as Fourier transform are used to convert the spectral distribution variation data into a frequency domain representation. The transformed spectral distribution variation data is then decomposed into multiple scale levels using wavelet transform; for example, lower scales correspond to high-frequency details, and higher scales correspond to low-frequency approximations, thus obtaining the spectral distribution variation data at each scale. This decomposition helps capture subtle changes in the signal during propagation, especially in fault location scenarios, highlighting distance-related attenuation patterns. Statistical indicators of the spectral distribution variation data at each scale, such as mean and variance, are calculated and matched with corresponding indicators of amplitude attenuation characteristics. The Pearson correlation coefficient between the energy ratio and the attenuation pattern is calculated. If the correlation coefficient exceeds a preset value, such as 0.8, a strong coupling relationship is identified; otherwise, a weak coupling relationship is identified. Through correlation analysis, strong coupling relationships can be discovered during signal propagation, providing important basis for fault location and signal transmission model optimization.

[0030] In one specific embodiment, obtaining the dynamic mapping relationship between phase offset data and propagation distance in step S2 specifically includes the following steps: Step S221: Based on the phase information, obtain the phase value of the current signal data at each time point and generate phase value sequence data; Step S222: Compare the phase value sequence data with the coupling relationship to identify the factors that cause phase shift. These factors include signal attenuation intensity and frequency component interference. Step S223: Calculate the phase difference between adjacent time points in the phase value sequence data to obtain phase offset data, obtain the historical propagation distance corresponding to the phase offset data at each time point to generate paired data, and construct a dataset relating phase offset data and propagation distance based on the paired data; Step S224: Establish a regression model by inputting the associated dataset into the regression model to obtain the dynamic mapping relationship between phase shift and propagation distance; Step S225: Analyze the spatiotemporal differences of current signal data collected by multiple FTU devices, and optimize the dynamic mapping relationship based on the spatiotemporal differences.

[0031] Specifically, the current signal is frequency-domain transformed using Fast Fourier Transform (FFT) or other frequency domain analysis methods to obtain the phase value at each moment. A complete phase value sequence is generated using this phase information, recording the phase change of the current signal at different time points. This phase value sequence is then compared with the coupling relationship between spectral distribution changes and amplitude attenuation extracted in previous steps. This comparison helps analyze the factors affecting the signal's phase change. For example, if the signal amplitude attenuation is significant during a certain period, and high-frequency components of the signal experience interference, the phase shift may be more pronounced. By comparing the relationship between amplitude attenuation, spectral changes, and phase shift, it is determined that the signal attenuation intensity and frequency component interference are the main factors causing phase shift. Other influencing factors typically include the characteristics of the propagation medium (such as line length and conductor type) and high-frequency component interference. By identifying these influencing factors, the system can better understand the pattern of phase shift and thus predict changes during signal propagation.

[0032] Phase difference reflects the time delay or change experienced by the signal during transmission. The phase difference at each historical time point corresponds to a historical propagation distance, which refers to the physical line length between the current signal and the FTU device from the fault source location. This distance can be obtained through the topology and line configuration of the distribution network. At different time points, paired data between historical propagation distance and phase offset are formed, forming a key dataset. A linear regression model is applied to fit the associated dataset, and the mapping is verified and constructed. For example, cross-validation is used to check the fitting accuracy. After training the model, the error is evaluated on the test set. If the error is less than a preset value, such as 5%, the mapping result is confirmed to be in the form of a dynamic function, thereby achieving a precise correspondence between phase offset and propagation distance.

[0033] However, since FTU devices in the distribution network are usually located in different electrical locations, and the signal propagation speed and delay are affected by factors such as line length, load fluctuations, and electrical environment, the phase offset analysis in step S224 usually assumes that the signal propagation is synchronous. However, due to spatiotemporal differences, the signal propagation rate may deviate, especially in the case of long distances or large load fluctuations. If step S224 fails to fully consider the signal propagation delay and synchronization issues, it may lead to a lack of accurate modeling of spatiotemporal factors in the framework, resulting in poor adaptability of the overall model in complex environments, thereby affecting the accuracy of fault detection. Therefore, it is necessary to analyze the spatiotemporal differences of current signal data collected by multiple FTU devices and optimize the dynamic mapping relationship based on the spatiotemporal differences. The specific optimization process will be explained later.

[0034] In one specific embodiment, step S225 specifically includes the following steps: The current signal data collected by multiple FTU devices are aligned to a unified reference through a time synchronization mechanism to obtain the corrected signal acquisition delay data. Based on the signal acquisition delay data, a high-precision clock protocol is used to calculate the timing deviation between each FTU device. The clock of each FTU device is adjusted in real time based on the timing deviation to obtain a consistent timing signal. For consistent time-series signals, a multi-feature coupling analysis framework is used to extract key feature values ​​from the phase offset data. The multi-feature coupling analysis framework integrates multiple signal features by linearly combining amplitude attenuation, spectral distribution change data and phase information to comprehensively analyze the signal propagation process. Determine if there is a deviation in the dynamic mapping relationship between key feature values ​​and propagation distance. If so, optimize the dynamic mapping result.

[0035] Specifically, such as Figure 2 The diagram shows a flowchart for optimizing the dynamic mapping results. It collects raw signal data from each FTU device and calculates the initial delay value of each device relative to a reference base, where the reference base is the selected master terminal clock. Based on the initial delay value, a synchronization algorithm is applied to adjust signal alignment, resulting in corrected delay data. A high-precision clock protocol, such as PTP, is used to calculate the timing deviation between terminals. This deviation is obtained by sending synchronization messages and delay request messages, representing the round-trip time difference. The terminal clocks are adjusted in real-time based on the deviation value to generate a consistent timing signal. This adjustment ensures that the signal maintains timing consistency during the propagation distance mapping, avoiding positioning errors caused by accumulated deviations. Fourier transform is applied to the consistent time-series signal to separate the phase offset data. Under the multi-feature coupling analysis framework, the statistical characteristics of the phase offset, such as the mean and variance, are calculated as key feature values. The multi-feature coupling analysis framework is an integration process that is achieved by linearly combining amplitude attenuation and spectral distribution data with phase data. This allows the extracted key feature values ​​to more comprehensively reflect the propagation characteristics of the fault signal. If there is a deviation between the key feature values ​​and the propagation distance, a dynamic adjustment model is constructed through the analysis framework to determine the correction range of the phase offset data. The specific process will be explained later.

[0036] The correction range is input into a mapping function to calculate the optimized distribution value. This mapping function, based on linear interpolation, outputs a dynamic mapping result. This dynamic mapping process, which closely correlates phase shift and propagation distance, ensures the results adapt to complex FTU device scenarios. By analyzing the spatiotemporal differences in current signal data acquired from multiple FTU devices, and combining high-precision clock synchronization and timing deviation correction techniques, the system can eliminate time delay differences between devices, thus providing consistent data for subsequent signal propagation analysis.

[0037] In one specific embodiment, optimizing the dynamic mapping result includes the following steps: Compare the dynamic mapping relationship with the preset correlation curve and calculate the deviation. The correlation curve is the standard relationship curve between phase offset and propagation distance. If the deviation is greater than the preset standard threshold, the least squares method is used to fit the phase offset data; Based on the fitting results, the correction range of the phase shift data is determined. A mapping function is set based on the linear interpolation method. The correction range is input into the mapping function to generate an optimized distribution of the dynamic mapping results.

[0038] Specifically, also refer to Figure 2 First, a standard relationship curve is defined, representing the propagation distance corresponding to a given phase offset value. This curve can be obtained from historical data, physical models, or experimental data. The calculated dynamic mapping relationship is compared with the preset standard relationship curve to obtain the deviation. If the calculated deviation exceeds a preset standard threshold (e.g., 5%), it indicates a significant deviation between the current dynamic mapping result and the preset correlation curve, requiring optimization. The least squares method is used to fit the current phase offset data. The parameters of the fitted model are calculated using the least squares method to find the model best suited to the actual data, resulting in a fitted model whose output more accurately reflects the relationship between phase offset and propagation distance. Based on the correction range, the original mapping function is interpolated and adjusted. Linear interpolation estimates the unknown value between two known data points based on their linear relationship. The entire mapping function is updated using interpolation methods to generate an optimized dynamic mapping result. The new mapping relationship will more accurately reflect the relationship between phase offset and propagation distance. The optimized dynamic mapping result provides a more accurate propagation distance estimate, helping to quickly and accurately locate the fault source. The accuracy of fault location is further improved through the correction range and interpolation methods.

[0039] In one specific embodiment, determining the comprehensive relationship between multi-dimensional feature data and propagation distance in step S3 specifically includes the following steps: The spectral distribution variation data with strong coupling relationship and amplitude attenuation are used as the first input features; The dynamic mapping result of phase offset data and propagation distance is subjected to Fourier transform to convert the time domain signal into a frequency domain signal and obtain the second input feature; The first and second input features are concatenated to form multi-dimensional feature data. Principal component analysis is used to reduce the dimensionality of the multi-dimensional feature data and extract key features. A joint distribution model is established based on the Gaussian mixture model to fit the relationship between key features and propagation distance; The model's fit is verified by calculating the likelihood function value. If the fit is greater than a preset threshold, the joint distribution model outputs the comprehensive relationship between key features and propagation distance.

[0040] Specifically, while considering amplitude attenuation or spectral changes alone can reflect the signal attenuation process to some extent, it cannot capture the more subtle changes during signal propagation. For example, a signal may encounter different environmental factors during propagation, resulting in different rates of amplitude attenuation and frequency component changes. Combining these two factors as the first input feature allows for a comprehensive consideration of multiple dimensions of signal attenuation. Fourier transform is applied to the dynamic mapping result of phase shift. Fourier transform is the process of converting a time-domain signal into a frequency-domain signal. The frequency components of the signal are obtained to represent its spectral characteristics, which are used as the second input feature. The first and second input features are concatenated as vectors to form a coupling matrix. Each row of the coupling matrix represents a vector representation of a feature combination. Principal component analysis is used to reduce the dimensionality of the coupling matrix and input it into the joint distribution model. The joint distribution model is established based on the Gaussian mixture model. The relationship between these input variables and the propagation distance is shown. The Gaussian mixture model is the process of modeling the data distribution by estimating the weighted sum of multiple Gaussian distributions using the expectation-maximization algorithm. The parameters of the Gaussian mixture model are calculated, including the mean, covariance, and weight of each Gaussian component. The mean represents the distribution center, the covariance represents the correlation between variables, and the weight represents the contribution ratio of each component. The model fit is verified by calculating the likelihood function value. If the likelihood value is higher than a preset threshold, the model is confirmed to be effective. The verified Gaussian mixture model is used as the joint distribution model to output the probabilistic relationship between multi-dimensional features and propagation distance.

[0041] In one specific embodiment, the calibration process for the joint distribution model in step S3 specifically includes the following steps: The line condition data of the power distribution system includes line length, conductor type and insulation material parameters, and the environmental adaptation data includes real-time monitoring values ​​of temperature, humidity and wind speed. The line impedance coefficient and environmental interference coefficient are calculated based on the line condition data and environmental adaptation data. The line impedance coefficient and environmental interference coefficient are input into the joint distributed model to ensure that the model receives dynamic features related to amplitude attenuation and phase shift. Based on the dynamic features, the parameters of the joint distributed model are calibrated to adjust the model's adaptability to complex power distribution environments and obtain a distance relationship model that adapts to complex scenarios.

[0042] Specifically, in power distribution systems, signal propagation characteristics are influenced by various factors, including the physical properties of power lines and external environmental conditions. Line condition data primarily includes line length, conductor type, and insulation material parameters, which directly affect signal attenuation characteristics and propagation speed during transmission. Environmental adaptation data includes factors such as temperature, humidity, and wind speed, which also significantly impact the power distribution system. Temperature changes can cause thermal expansion or contraction of conductors, thus affecting current conductivity; humidity and wind speed can affect signal propagation speed in the air, especially under severe weather conditions, where signal propagation can be significantly affected. To accurately reflect the impact of these environmental factors on signal propagation, the model must consider these variables and adjust signal propagation parameters based on real-time weather data. In practical applications, by calculating the line impedance coefficient and environmental interference coefficient, the model can obtain relevant signal propagation characteristics and input them into a joint distribution model for further processing. During this process, the model performs weighted analysis on the input data to ensure that characteristics such as signal amplitude attenuation, spectral changes, and phase shifts reflect propagation characteristics under different power grid environments. For example, signal attenuation may be faster in low-temperature environments, while signal propagation speed may be significantly affected in high-humidity environments. Taking these factors into account, the model dynamically adjusts its parameters, enabling the generated distance relationship model to more accurately predict the propagation distance between the fault source and the FTU device.

[0043] During model calibration, statistical methods such as Gaussian mixture models were used to fit various characteristics of signal propagation, and the model was optimized based on background data. By calculating the relationship between different input features (such as amplitude attenuation, spectral variation, and phase shift) and propagation distance, the calibrated joint distribution model can more accurately estimate the fault location. The calibration method based on multi-dimensional features enables the joint distribution model to handle more dimensional data, enhancing the model's expressive power and prediction accuracy, especially for better prediction and optimization of signal attenuation and path changes in complex scenarios.

[0044] In one specific embodiment, step S4 specifically includes the following steps: If the dynamic fluctuation data of the current signal exceeds the preset threshold range, the least squares method is used to adjust the model parameters of the distance relationship model in real time, and the updated parameters are reintegrated into the model. Based on the adjusted distance relationship model, the preliminary estimate of the fault location is calculated. Background data for complex scenarios is acquired, including route terrain and weather records. This background data is then input into a distance relationship model. The initial estimate is adjusted using gradient descent and iteratively optimized multiple times to obtain a precise fault location that matches the actual route conditions.

[0045] Specifically, the FTU device monitors and collects dynamic fluctuation information of current signals in the power distribution system in real time. When a fault occurs in the system, the amplitude and phase information of the current signal data will change after the fault occurs. The system infers the location of the fault by inputting these signals into a distance relationship model. The preliminary estimate of the fault location is based on the simulation of signal propagation by the distance relationship model, combined with the attenuation characteristics of the signal during transmission to estimate the distance to the fault source.

[0046] Once the dynamic fluctuations in the current signal exceed a preset threshold, indicating a potentially significant change, the system initiates a real-time model adjustment mechanism. This process utilizes the least squares method to adjust the model's parameters. Least squares is a commonly used numerical optimization method that updates model parameters by minimizing the sum of squared errors between data points and predicted values, making the model better reflect actual observations. In this technique, least squares is used to correct various parameters related to current signal propagation (such as propagation speed and signal attenuation) in the distance relationship model, resulting in a more accurate propagation distance estimate. The adjusted model is then reintegrated into the distance relationship model, and the preliminary estimate of the fault location is recalculated based on the updated parameters. At this point, the system uses the adjusted model to extrapolate the distance to the fault source, obtaining a preliminary fault location. To further improve positioning accuracy, the system incorporates background data from complex scenarios to further optimize the fault location results. Background data includes power distribution system line terrain data and weather records, which provide crucial auxiliary information for signal propagation accuracy.

[0047] By acquiring this background data in real time, the system can further revise the preliminary location results based on actual environmental conditions. The system further optimizes the preliminary estimate using the gradient descent method. Gradient descent is an iterative optimization algorithm that finds the minimum error point in a multi-dimensional space. This method calculates the gradient of the model parameters, indicating the direction of minimizing the error, and adjusts the parameters through multiple iterations until the model fits the optimal solution. In fault location applications, gradient descent is used to adjust the preliminary estimate of the fault location, continuously optimizing the parameters to ensure it more closely reflects actual line conditions. By incorporating complex environmental data such as line terrain and weather records, the model can adapt to different power grid environments and signal propagation characteristics under different weather conditions, ensuring more accurate fault location results. By combining real-time acquired current signal data with background data, the system dynamically optimizes the fault location using the least squares method and gradient descent, thereby improving the accuracy and reliability of fault location. This process fully considers the actual environment and signal propagation characteristics of the power system, providing high-precision fault location even under various complex environmental conditions, ensuring the stable operation of the distribution system and efficient fault handling.

[0048] like Figure 3 The image shows a comparison of fault location errors in a power distribution system. The black line represents the fault location accuracy of the traditional method, and the line represents the location error measured at each test point. Because the traditional method may not consider complex electrical parameters or environmental factors, it typically exhibits a larger error value (higher location error). The gray line (joint distribution model) represents the location error of the joint distribution model. Using the joint distribution model, this method considers multiple signal characteristics such as amplitude attenuation and spectral distribution changes, thus reducing the location error compared to the traditional method. Although the error is improved compared to the traditional method, there is still room for improvement. The dark gray line (distance relationship model) represents the optimized joint distribution model. This method, based on the joint distribution model, further incorporates line conditions, environmental factors, and background factors, and optimizes the impact of time delay and complex scene factors during signal propagation. Therefore, under this method, the location error is significantly reduced, showing the lowest error value, indicating that in complex environments, the location accuracy of the optimized joint distribution model is much higher than that of the traditional method and the standard joint distribution model.

[0049] The above describes a method for monitoring an intelligent power distribution system based on an FTU in an embodiment of this application. The following describes an intelligent power distribution system monitoring device based on an FTU in an embodiment of this application. Please refer to [link / reference]. Figure 4 One embodiment of an FTU-based intelligent power distribution system monitoring device in this application includes: The separation module performs time-domain and frequency-domain separation processing on the current signal data to obtain amplitude and phase information. Based on the amplitude information, it analyzes the amplitude attenuation law and determines the correlation characteristics between amplitude attenuation and propagation distance. The propagation distance is the line distance between the current signal data and the FTU device from the fault location. The mapping module extracts spectral distribution change data based on associated features, performs multi-scale decomposition on the spectral distribution change data, determines the coupling relationship between spectral distribution change and amplitude attenuation, analyzes the phase shift effect based on phase information and coupling relationship, and obtains the dynamic mapping relationship between phase shift data and propagation distance. The preliminary judgment module generates multi-dimensional feature data based on correlation features, coupling relationships, and dynamic mapping relationships, establishes a joint distribution model, determines the comprehensive relationship between multi-dimensional feature data and propagation distance, obtains line condition data and environmental adaptation data of the power distribution system, calibrates the joint distribution model, and obtains a distance relationship model adapted to complex scenarios. The optimization module collects dynamic fluctuation data of real-time current signals from the FTU device, inputs it into the distance relationship model, and combines it with background data from complex scenario analysis to obtain accurate fault location results.

[0050] Those skilled in the art will clearly understand that, for the sake of convenience and brevity, the specific working processes of the systems and units described above can be referred to the corresponding processes in the foregoing method embodiments, and will not be repeated here.

[0051] If the integrated unit is implemented as a software functional unit and sold or used as an independent product, it can be stored in a computer-readable storage medium. Based on this understanding, the technical solution of this application, in essence, or the part that contributes to the prior art, or all or part of the technical solution, can be embodied in the form of a software product. This computer software product is stored in a storage medium and includes several instructions to cause a computer device (which may be a personal computer, server, or network device, etc.) to execute all or part of the steps of the methods described in the various embodiments of this application. The aforementioned storage medium includes various media capable of storing program code, such as USB flash drives, portable hard drives, read-only memory (ROM), random access memory (RAM), magnetic disks, or optical disks.

[0052] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them. Although this application has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some of the technical features. Such modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the spirit and scope of the technical solutions of the embodiments of this application.

Claims

1. A monitoring method for an intelligent power distribution system based on an FTU, characterized in that, Multiple FTU devices are installed at different locations in the power distribution system to acquire current signal data when a fault occurs in the power distribution system. The method includes: Step S1: Perform time-domain and frequency-domain separation processing on the current signal data to obtain amplitude information and phase information. Analyze the amplitude attenuation law based on the amplitude information to determine the correlation characteristics between amplitude attenuation and propagation distance. The propagation distance is the line distance between the current signal data and the FTU device from the fault location. Step S2: Extract spectral distribution change data based on the associated features, perform multi-scale decomposition on the spectral distribution change data, determine the coupling relationship between spectral distribution change and amplitude attenuation, analyze the phase shift influence based on the phase information and the coupling relationship, and obtain the dynamic mapping relationship between phase shift data and propagation distance; Step S3: Generate multi-dimensional feature data based on correlation features, coupling relationships and dynamic mapping relationships, establish a joint distribution model, determine the comprehensive relationship between the multi-dimensional feature data and the propagation distance, obtain the line condition data and environmental adaptation data of the power distribution system, calibrate the joint distribution model, and obtain a distance relationship model adapted to complex scenarios; Step S4: The FTU device collects dynamic fluctuation data of real-time current signal, inputs it into the distance relationship model, and combines it with background data from complex scenario analysis to obtain the accurate location result of the fault.

2. The method according to claim 1, characterized in that, The characteristics determining the correlation between amplitude attenuation and propagation distance in step S1 include: Step S11: Set a fixed duration to divide the amplitude information within the target duration into multiple target windows, calculate the average amplitude within each target window and compare them to obtain a first change curve representing amplitude attenuation; Step S12: Identify the inflection point of the first change curve, and perform time-domain signal decomposition on the first change curve based on the inflection point using empirical mode decomposition to obtain multiple intrinsic mode functions; Step S13: Calculate the energy spectrum of each intrinsic mode function, obtain the attenuation slope of each curve segment, fit the attenuation slope with the preset propagation distance parameter, and determine the correlation characteristics.

3. The method according to claim 2, characterized in that, Step S2, determining the coupling relationship between the change in spectral distribution and the amplitude attenuation, includes: Step S211: Convert the spectral distribution change data into a frequency domain representation using a frequency domain transformation method; Step S212: Perform multi-scale decomposition on the converted spectral distribution change data to obtain multi-scale decomposition results including low-frequency and high-frequency components; Step S213: Based on the multi-scale decomposition results, calculate the energy ratio of the spectral distribution change data at each scale, and perform correlation analysis with the associated features to obtain the correlation coefficient; Step S214: If the correlation coefficient is greater than the preset first threshold, it is determined that there is a strong coupling relationship between the spectrum distribution change data and the amplitude attenuation; otherwise, it is a weak coupling relationship.

4. The method according to claim 1, characterized in that, The dynamic mapping relationship between the phase offset data and the propagation distance obtained in step S2 includes: Step S221: Based on the phase information, obtain the phase value of the current signal data at each time point, and generate phase value sequence data; Step S222: Compare the phase value sequence data with the coupling relationship to identify the influencing factors that cause phase shift, including signal attenuation intensity and frequency component interference; Step S223: Calculate the phase difference between adjacent times in the phase value sequence data to obtain phase offset data, obtain the historical propagation distance corresponding to the phase offset data at each time point to generate paired data, and construct a dataset relating the phase offset data and the propagation distance based on the paired data; Step S224: Establish a regression model, input the associated dataset into the regression model, and obtain the dynamic mapping relationship between the phase offset and the propagation distance; Step S225: Analyze the spatiotemporal differences of current signal data collected by multiple FTU devices, and optimize the dynamic mapping relationship based on the spatiotemporal differences.

5. The method according to claim 4, characterized in that, Step S225 includes: The current signal data collected by multiple FTU devices are aligned to a unified reference through a time synchronization mechanism to obtain the corrected signal acquisition delay data. Based on the signal acquisition delay data, a high-precision clock protocol is used to calculate the timing deviation value between each FTU device. The clock of each FTU device is adjusted in real time based on the timing deviation value to obtain a consistent timing signal. For the consistent timing signal, the key feature values ​​in the phase offset data are extracted by combining the multi-feature coupling analysis framework. The multi-feature coupling analysis framework integrates multiple signal features by linearly combining amplitude attenuation, spectral distribution change data and phase information to comprehensively analyze the signal propagation process. Determine whether there is a deviation in the dynamic mapping relationship between the key feature value and the propagation distance. If so, optimize the dynamic mapping result.

6. The method according to claim 5, characterized in that, Optimizing the dynamic mapping result includes: The dynamic mapping relationship is compared with a preset correlation curve and the deviation is calculated. The correlation curve is a standard relationship curve between phase offset and propagation distance. If the deviation is greater than a preset standard threshold, the least squares method is used to fit the phase offset data; Based on the fitting results, the correction range of the phase offset data is determined, a mapping function is set based on the linear interpolation method, the correction range is input into the mapping function, and an optimized distribution of the dynamic mapping results is generated.

7. The method according to claim 3, characterized in that, Determining the comprehensive relationship between the multi-dimensional feature data and the propagation distance in step S3 includes: The spectral distribution variation data with strong coupling relationship and amplitude attenuation are used as the first input features; The dynamic mapping result between the phase offset data and the propagation distance is subjected to Fourier transform to convert the time domain signal into a frequency domain signal, thereby obtaining the second input feature; The first input feature and the second input feature are concatenated to form multi-dimensional feature data. Principal component analysis is used to reduce the dimensionality of the multi-dimensional feature data and extract key features. A joint distribution model is established based on the Gaussian mixture model to fit the relationship between the key features and the propagation distance. The model's fit is verified by calculating the likelihood function value. If the fit is greater than a preset threshold, the joint distribution model outputs the comprehensive relationship between the key features and the propagation distance.

8. The method according to claim 1, characterized in that, The calibration process for the joint distribution model in step S3 includes: The line condition data of the power distribution system includes line length, conductor type and insulation material parameters, and the environmental adaptation data includes real-time monitoring values ​​of temperature, humidity and wind speed. The line impedance coefficient and environmental interference coefficient are calculated based on the line condition data and the environmental adaptation data. The line impedance coefficient and the environmental interference coefficient are input into the joint distribution model to ensure that the model receives dynamic features related to amplitude attenuation and phase shift. Based on the dynamic features, the parameters of the joint distribution model are calibrated to adjust the model's adaptability to complex power distribution environments and obtain a distance relationship model that adapts to complex scenarios.

9. The method according to claim 1, characterized in that, Step S4 includes: If the dynamic fluctuation data of the current signal exceeds the preset threshold range, the least squares method is used to adjust the model parameters of the distance relationship model in real time, and the updated parameters are reintegrated into the model. Based on the adjusted distance relationship model, a preliminary estimate of the fault location is calculated. Background data for complex scenarios is acquired, including route terrain and weather records. This background data is then input into the distance relationship model. The initial estimate is adjusted using gradient descent, and the initial estimate is iteratively optimized through multiple rounds to obtain a precise fault location that matches the actual route conditions.

10. An FTU-based intelligent power distribution system monitoring device, used to implement the FTU-based intelligent power distribution system monitoring method as described in any one of claims 1-9, characterized in that, The monitoring device includes: The separation module performs time-domain and frequency-domain separation processing on the current signal data to obtain amplitude information and phase information. Based on the amplitude information, it analyzes the amplitude attenuation law and determines the correlation characteristics between amplitude attenuation and propagation distance. The propagation distance is the line distance between the current signal data and the FTU device from the fault location. The mapping module extracts spectral distribution change data based on the associated features, performs multi-scale decomposition on the spectral distribution change data, determines the coupling relationship between spectral distribution change and amplitude attenuation, analyzes the phase shift influence based on the phase information and the coupling relationship, and obtains the dynamic mapping relationship between phase shift data and propagation distance. The preliminary judgment module generates multi-dimensional feature data based on correlation features, coupling relationships, and dynamic mapping relationships, establishes a joint distribution model, determines the comprehensive relationship between the multi-dimensional feature data and the propagation distance, obtains line condition data and environmental adaptation data of the power distribution system, calibrates the joint distribution model, and obtains a distance relationship model adapted to complex scenarios. The optimization module collects dynamic fluctuation data of real-time current signals from the FTU device, inputs it into the distance relationship model, and combines it with background data from complex scenario analysis to obtain accurate fault location results.

Images