Electric energy quality comprehensive monitoring method and system based on intelligent fusion terminal

Abstract

The invention relates to the field of power quality monitoring of a power system, in particular to a comprehensive power quality monitoring method and system based on an intelligent fusion terminal, and the method comprises the steps: obtaining and preprocessing power quality data of the intelligent fusion terminal, dividing electrical partitions, determining dominant harmonic components, and calculating the electrical distance between terminals in the same partition. And based on density peak clustering, combining feature difference and electrical distance weight fusion to obtain a similarity distance, determining an adaptive cutoff distance by using a low-density region adjustment factor, and drawing a decision diagram to identify a clustering center. Screening clustering results, analyzing electric energy quality characteristic dominance, identifying a mixed cluster and re-clustering, and positioning a disturbance source by using an attenuation law in combination with a significant clustering result and an electrical distance. According to the method, the accuracy of disturbance source positioning is improved through adaptive clustering and mixed cluster analysis, and the operation condition of the power grid and multiple disturbance sources can be accurately identified through the adaptive truncation distance and the adjustment factor of the low-density region.

Description

TECHNICAL FIELD

[0001] The present application relates to the field of power system electric energy quality monitoring. In particular, it relates to an electric energy quality comprehensive monitoring method and system based on an intelligent fusion terminal. BACKGROUND

[0002] The importance of intelligent fusion terminals in electric energy quality monitoring is increasingly evident. Electric energy quality not only relates to the stable operation of the power system, but also directly affects the performance and lifespan of various electrical equipment. Through intelligent fusion terminals, key data such as voltage and current in the power grid can be collected in real time, providing a basis for accurate detection of electric energy quality and helping to promptly identify abnormal disturbances in the power grid, ensuring the reliability and quality of power supply, and being crucial for maintaining the operational efficiency of the power system and the overall user experience.

[0003] There are still many problems in using intelligent fusion terminals for electric energy quality monitoring. There are multiple disturbance sources in the power grid, and the electric energy quality problems caused by these disturbance sources are intertwined, making it difficult to accurately distinguish and locate the disturbance sources relying solely on traditional monitoring methods. To address this challenge, clustering methods are often used to match electric energy quality characteristics and electrical distances in an attempt to effectively identify disturbance sources. However, the physical topology of the power grid has clear regional division and connection characteristics, and the electrical characteristics between different regions differ significantly. This leads to ineffective clustering across regions during the clustering process, making the clustering results unable to accurately reflect the actual operation of the power grid. In addition, mixed clusters often appear in the clustering results, and the data in these mixed clusters combines information from different disturbance sources, further increasing the difficulty of accurately identifying multiple disturbance sources. SUMMARY

[0004] To solve the problem of electric energy quality problems caused by the interweaving of multiple disturbance sources in the power grid being difficult to accurately distinguish and locate, and the limitations of existing clustering methods in handling differences in cross-regional electrical characteristics and mixed clusters, which affect the accurate reflection of the actual operation of the power grid and the effective identification of multiple disturbance sources, the present application provides solutions in the following aspects.

[0005] In a first aspect, the method for power quality comprehensive monitoring based on intelligent fusion terminal comprises: obtaining a power quality data set of the intelligent fusion terminal and preprocessing, dividing an electrical partition, and determining a dominant harmonic component; calculating the electrical distance between the intelligent fusion terminals in the same partition based on the dominant harmonic component in the preprocessed power quality data set and the electrical path in the topology structure of the power grid; clustering the intelligent fusion terminals in the same partition based on the density peak, calculating the feature difference and the electrical distance between the two intelligent fusion terminals and performing weighted fusion, weighting the feature difference and the electrical distance according to the preset weight distribution rule, obtaining the similarity distance, determining the adaptive cut-off distance combined with the adjustment factor of the low-density area, calculating the local density of each intelligent fusion terminal and the relative distance between each intelligent fusion terminal and other intelligent fusion terminals with higher density, drawing a decision graph based on the local density and the relative distance, identifying the density peak point to determine the clustering center, and obtaining the preliminary clustering result in the same partition; screening the clustering result, analyzing the dominance of each power quality feature, and judging whether it is a mixed cluster, re-clustering the mixed cluster based on the corresponding power quality feature to obtain a significant clustering result; combining the significant clustering result and the electrical distance, and obtaining the location of the disturbance source by using the power quality attenuation law; wherein the preset weight distribution rule is that the standard deviation of the feature difference of the power quality of all terminals in the same partition is divided by the sum of the standard deviation of the feature difference of the power quality and the standard deviation of the electrical distance, to obtain the weight coefficient of the similarity distance, and the weight coefficient of the electrical distance is obtained by subtracting the weight coefficient of the similarity distance from 1.

[0006] Preferably, the dominant harmonic component comprises: The square of the amplitude of each harmonic voltage is divided by the cumulative sum of the squares of all harmonic voltage amplitudes to obtain the energy proportion of each harmonic component, the harmonic component with the largest energy proportion is selected as the dominant harmonic component, and the dominant harmonic component is identified according to the dominant harmonic component.

[0007] Preferably, the calculation of the electrical distance comprises: Taking any terminal as a target terminal, a plurality of electrical paths between the target terminal and other terminals are obtained, wherein each electrical path comprises a plurality of conductor lines and a plurality of transformers; The resistance components and the reactance components of all conductor lines in the electrical path are summed respectively, and the sum of the sum results is added to obtain the total line impedance of the electrical path; the short-circuit impedances of all transformers in the electrical path are summed to obtain the total transformer impedance of the electrical path; the total line impedance and the total transformer impedance are summed, and the complex modulus value is taken to obtain the electrical distance between the target terminal and other terminals.

[0008] Preferably, the resistance component and the reactance component are obtained by the steps of: Taking any conductor as a target conductor, a complex modulus value between a unit length resistance of a conductor line of the target conductor and a length of the conductor line of the target conductor is calculated to obtain a resistance component of the target conductor; a complex modulus in the complex number is normalized by 1, A complex modulus value between a unit length inductance of the conductor line and the length of the conductor line is obtained to obtain a reactance component of the target conductor.

[0009] Preferably, the step of obtaining the feature difference comprises: A power quality feature vector is constructed for each terminal in the same subarea, wherein the power quality feature vector comprises: voltage deviation, frequency deviation, total harmonic distortion, voltage amplitude of a dominant harmonic component, voltage sag amplitude and voltage sag duration; Taking any intelligent fusion terminal as a target terminal, a Euclidean distance between power quality feature vectors of two intelligent fusion terminals in the same subarea is calculated to obtain the feature difference.

[0010] Preferably, the calculation method of the adjustment factor of the low-density area comprises: An average distance of electrical distances between all terminal pairs in the subarea is calculated, and a distance of a nearest neighbor point pair for density calculation is determined according to the target cutoff ratio and the number of sample pairs; a difference between the average distance and the distance of the nearest neighbor point pair is normalized and added by 1 to obtain the adjustment factor of the low-density area.

[0011] Preferably, the dominant degree of the power quality feature comprises: The total number of intelligent fusion terminals contained in the cluster is counted, the number of terminals whose preset power quality features exceed standard values of corresponding features is screened out, and a ratio to the total number is calculated to obtain the feature dominant degree.

[0012] Preferably, the position of the disturbance source is obtained by using the power quality attenuation law, comprising: A harmonic disturbance event library and a voltage sag event library are established according to historical data and clustering results, and the position of the disturbance source, the power grid topology and the power feature values of the terminals are recorded; An electrical distance from each terminal to the disturbance source point is determined by using an electrical distance calculation method, and an attenuation curve fitting is performed on harmonic and voltage sag disturbances respectively to obtain respective attenuation coefficients and attenuation models; An intelligent fusion terminal most affected by the disturbance is selected as a backtracking point, and the amplitude of the intelligent fusion terminal is substituted into the fitted attenuation model to calculate the electrical distance from the disturbance source to the backtracking point, and finally the position of the disturbance source is determined in combination with the power grid topology.

[0013] Preferably, the power quality data set comprises: voltage deviation, frequency deviation, total harmonic distortion, voltage amplitude, instantaneous voltage value, voltage sag duration and voltage sag amplitude. According to the power grid topology, the electrical partition is divided, and according to the preset abnormality judgment standard, the index abnormal terminal is screened out from the electrical partition power quality data set, and the mean value and standard deviation of each power quality data set are calculated according to the terminal data in the electrical partition, and when the absolute difference value of each power quality data set and the mean value is greater than 3 times the standard deviation, the characteristic value is determined as an abnormal value, and is removed; the power quality data set is normalized by maximum and minimum, and the dimension difference is eliminated.

[0014] In a second aspect, the power quality comprehensive monitoring system based on the intelligent fusion terminal comprises a processor and a memory, and the memory stores computer program instructions.

[0015] The present application has the following effects: 1、The present application can more accurately identify and distinguish multiple disturbance sources by combining power quality characteristics and electrical distance for adaptive clustering, and further analyzing and re-clustering the mixed cluster, thereby improving the accuracy of disturbance source positioning. Avoids the positioning error caused by the difference in electrical characteristics across regions and mixed clusters in the traditional method, and ensures the accurate reflection of the actual operation of the power grid.

[0016] 2、The present application can automatically adapt to the data distribution of different density regions by adjusting the adaptive truncation distance and the low-density region adjustment factor, thereby improving the adaptability and reliability of the clustering algorithm, reducing the misclustering phenomenon, improving the accuracy of the clustering result, and enhancing the robustness of the clustering algorithm in different scenarios, thereby ensuring the accurate reflection of the actual operation of the power grid and the effective identification of multiple disturbance sources. BRIEF DESCRIPTION OF DRAWINGS

[0017] Figure 1 is a method flowchart of steps S1-S4 in the power quality comprehensive monitoring method based on the intelligent fusion terminal of the embodiment of the present application.

[0018] Figure 2 is a structural block diagram of the power quality comprehensive monitoring system based on the intelligent fusion terminal of the embodiment of the present application. DETAILED DESCRIPTION

[0019] The technical solutions in the embodiments of the present application will be described clearly and completely below with reference to the drawings in the embodiments of the present application. Obviously, the described embodiments are part of the embodiments of the present application, not all the embodiments.

[0020] Referring to Figure 1 The power quality comprehensive monitoring method based on the intelligent fusion terminal comprises steps S1-S4, and specifically as follows: S1: Obtain the power quality dataset of the intelligent fusion terminal and preprocess it, divide the electrical partition, and determine the dominant harmonic component.

[0021] The power quality dataset includes: voltage deviation, frequency deviation, total harmonic distortion, voltage amplitude, instantaneous voltage value, voltage sag duration, and voltage sag amplitude. It should be noted that the dominant harmonic component refers to the one or several harmonics in the power system that have the greatest impact on power quality. Harmonics are periodic components of voltage or current waveforms in a power system that deviate from sinusoidal waves. They can be integer multiples of the fundamental frequency. The dominant harmonic component usually has the largest amplitude and has the most significant impact on the operation of the power grid and the performance of the equipment.

[0022] According to the topology of the power grid, the electrical partition is divided into three levels: substation, feeder, and distribution transformer. For example, the terminal code can be represented as substation ID, feeder ID, and distribution transformer ID. Only terminals within the same partition are analyzed subsequently to avoid interference from cross-area disturbances.

[0023] Based on the power quality dataset after electrical partitioning, the abnormal terminals are selected according to the preset abnormality judgment criteria. For example, the voltage deviation threshold is set to , the frequency deviation threshold is set to , the total harmonic distortion threshold is set to , the voltage amplitude of each harmonic is set to , the voltage sag amplitude threshold is set to , and the voltage sag duration threshold is set to . The abnormal terminals selected are preprocessed, including outlier removal and standardization. The mean and standard deviation of each power quality dataset are calculated with the terminal data within the electrical partition. When the absolute difference between each power quality dataset and the mean is greater than 3 times the standard deviation, the feature value is determined as an outlier and is removed to eliminate invalid data caused by terminal failure or communication interference. The power quality dataset is normalized by maximum and minimum to eliminate dimension differences.

[0024] In the data preprocessing stage of the power quality monitoring system, the terminals with abnormal indicators are first selected according to the preset abnormality judgment criteria, focusing on the terminals that need to be paid attention to and excluding the terminals with normal data. Subsequently, invalid data caused by terminal failure or communication interference is further removed for these abnormal terminals to ensure the reliability of the data for subsequent analysis. Then, the preprocessed data is standardized to eliminate the dimension differences of different indicators, providing reliable data for subsequent operations. Through electrical partitioning, the analysis focuses on terminals with close electrical connections, which conforms to the physical boundaries of disturbance propagation. The reliability and consistency of the data are improved, ensuring the accuracy and effectiveness of subsequent analysis.

[0025] In power quality monitoring, accurately identifying the main harmonic component is crucial for diagnosing power grid problems. The present application calculates the energy proportion of each harmonic component by intelligently fusing the harmonic voltage amplitude data collected by the terminal, based on the principle that the square of the harmonic voltage amplitude is proportional to the harmonic power contribution.

[0026] Specifically, the square of each harmonic voltage amplitude is divided by the cumulative sum of the squares of all harmonic voltage amplitudes to obtain the energy proportion of each harmonic component, and the order of the harmonic component with the largest energy proportion is selected as the dominant harmonic order. By retaining the dominant harmonic component, false positives caused by interference from secondary harmonic components can be reduced, thereby improving the overall monitoring accuracy.

[0027] It should be noted that the voltage amplitude is a key feature that typically includes the amplitudes of the dominant harmonics and sub-harmonics. The one or several harmonic components with the largest amplitudes among all harmonic components. These components have the most significant impact on power quality, hence the term dominant. However, when performing density peak clustering, particular attention is paid to the voltage amplitude of the dominant harmonic, as the dominant harmonic has the most significant impact on power quality.

[0028] The dominant harmonic order refers to the frequency multiple of the harmonic with the largest amplitude in the harmonic spectrum in a power system. For example, if the fundamental frequency is 50Hz or 60Hz, the dominant harmonic order could be the 3rd, 5th, or 7th harmonic, etc., depending on which harmonic has the largest amplitude. Therefore, the dominant harmonic order corresponding to the highest energy proportion in the harmonic voltage amplitude is obtained, and the specific steps are as follows: Specifically, the dominant harmonic order satisfies the following relationship: ; In the formula, represents the dominant harmonic order, represents the harmonic voltage amplitude, the square term reflects the power contribution of the harmonic, and the harmonic with the highest energy proportion is the component that has the most significant impact on propagation, represents the maximum harmonic component order that the terminal can collect, represents the minimum harmonic component order that the terminal can collect, represents the order of the harmonic component with the highest energy proportion among all harmonic voltage amplitudes.

[0029] In power quality monitoring systems, topological information of terminals (such as their associated lines and transformer connections) is obtained from databases like Geographic Information Systems (GIS). This aims to accurately grasp the structure of the power grid and the connections between equipment, providing the monitoring system with a detailed view of the grid layout and enhancing monitoring accuracy. It also clarifies disturbance propagation paths and facilitates faster fault location. Furthermore, topological information-based zoned monitoring strategies can optimize monitoring resource allocation and improve monitoring efficiency. In addition, topological information provides crucial data analysis and intelligent decision-making, helping to implement preventative measures and optimize maintenance plans, thereby significantly improving the reliability and operational efficiency of the power grid.

[0030] S2: Calculate the electrical distance between smart fusion terminals within the same partition by combining the dominant harmonic components in the preprocessed power quality dataset with the electrical paths in the power grid topology.

[0031] Taking any terminal as the target terminal, obtain multiple electrical paths between the target terminal and other terminals, wherein each electrical path includes multiple wire lines and multiple transformers; The resistive and inductive reactance components of all conductors in the electrical path are summed separately, and the sums are added together to obtain the total line impedance of the electrical path. The short-circuit impedances of all transformers in the electrical path are summed to obtain the total transformer impedance of the electrical path. The total line impedance and the total transformer impedance are summed, and the complex modulus is taken to obtain the electrical distance between the target terminal and other terminals.

[0032] Specifically, electrical distance satisfies the following relationship: ; In the formula, Indicates terminal With terminal Electrical distance between them Indicates terminal To the terminal There are a total of Section of conductor line, Represents the resistance component of a wire. This represents the inductive reactance component of the conductor. Indicates terminal With terminal The number of transformers between Indicates the first in the electrical path The short-circuit impedance (Ω) of a transformer, where the short-circuit impedance of the transformer includes the equivalent resistance. and equivalent leakage reactance , , Indicates taking the modulus of a complex number. This reflects the transformer's impediment to current. By taking a complex modulus, the vector sum of resistance and inductive reactance is converted into a scalar value, ensuring that the physical meaning of the electrical distance is clear.

[0033] Taking any conductor as the target conductor, calculate the complex modulus of the resistance per unit length of the target conductor's line and the length of the target conductor's line to obtain the resistance component of the target conductor; convert the imaginary unit in the complex number... The inductive reactance component of the target conductor is obtained by taking the complex modulus between the dominant harmonic component frequency, the corresponding unit length inductance of the conductor, and the length of the line.

[0034] Specifically, the resistive component and the inductive reactance component satisfy the following relationships: ; ; In the formula, Represents the resistance component of a wire. This represents the inductive reactance component of the conductor. This represents the resistance per unit length of a conductor (Ω / km). Indicates the first The length of a conductor line (km). The imaginary unit in a complex number is used to represent the inductive reactance component. Indicates the frequency of the dominant harmonic component. Inductance per unit length of a conductor (Ω / km).

[0035] To further explain, This reflects the energy loss caused by resistance when current flows through a wire. This reflects the resistance of inductance to current in AC circuits. The higher the frequency (such as higher harmonics), the greater the inductive reactance, the stronger the resistance to current, and the faster the harmonics attenuate during propagation.

[0036] Based on physical meaning, it comprehensively considers the resistance and inductive reactance components of the line, as well as the short-circuit impedance of the transformer. By quantifying the equivalent impedance loss of disturbance propagation, it provides a physically meaningful distance metric for cluster analysis. It quantifies the electrical correlation between nodes by considering line impedance and transformer impedance.

[0037] In electrical engineering, the imaginary unit in complex numbers This is used to represent impedance in complex form, where the real part represents resistance and the imaginary part represents inductive reactance. The formula multiplies by the imaginary unit. is to combine the inductive component and the resistive component to form a complex impedance. Not only can it comprehensively consider the impedance of the line to the alternating current, including resistive loss and inductive loss, but also can calculate the total impedance size through the modulus of the complex number, thereby providing a physical meaning clear measurement for the calculation of electrical distance.

[0038] The electrical distance can quantify the physical loss of the disturbance in the process of propagation in the power grid. Compared with the pure spatial distance and other measurement methods, it is more in line with the actual law of disturbance attenuation. In addition, by calculating only the electrical distance between terminals in the same electrical partition, the distance calculation error caused by the topological barrier across the partition can be effectively avoided.

[0039] S3: clustering the intelligent fusion terminals in the same partition based on the density peak value, calculating the feature difference and the electrical distance between each two intelligent fusion terminals, weighting and fusing the feature difference and the electrical distance according to the preset weight distribution rule to obtain a similarity distance, determining an adaptive cut-off distance combined with an adjustment factor of a low-density area, calculating the local density of each intelligent fusion terminal and the relative distance between each intelligent fusion terminal and other intelligent fusion terminals with higher density, drawing a decision graph based on the local density and the relative distance, identifying a density peak point to determine a clustering center, and obtaining a preliminary clustering result in the same partition.

[0040] An electric energy quality feature vector is constructed for each terminal in the same partition, wherein the electric energy quality feature vector includes: voltage deviation, frequency deviation, total harmonic distortion, voltage amplitude of dominant harmonic component, voltage sag amplitude and sag duration; In the clustering analysis, the voltage amplitude is replaced by the voltage amplitude of the dominant harmonic. Since the voltage amplitude of the dominant harmonic can more accurately reflect the main source of harmonic pollution in the power grid, the terminals affected by the same disturbance source can be more effectively identified during clustering. The electric energy quality disturbance source can be more accurately located, and the accuracy and efficiency of clustering can be improved.

[0041] The calculation method of the similarity distance includes: Taking any intelligent fusion terminal as a target terminal, the Euclidean distance between the electric energy quality feature vectors of each two intelligent fusion terminals in the same partition is calculated to obtain a feature difference; The weight coefficient of the similarity distance is obtained by dividing the standard deviation of the electric energy quality feature difference of all terminals in the same partition by the sum of the standard deviation of the electric energy quality feature difference and the standard deviation of the electrical distance, and the weight coefficient of the electrical distance is obtained by subtracting the weight coefficient of the similarity distance from 1; The similarity distance and the electrical distance are weighted and fused based on the weight coefficient of the similarity distance and the weight coefficient of the electrical distance to obtain the similarity distance between the target terminal and other terminals.

[0042] The similarity matrix is constructed according to the similarity distance, and is sorted in descending order. The adaptive truncation distance is fused according to the target truncation ratio and the adjustment factor of the low-density area. In the existing density peak value clustering method, the truncation distance is determined according to the target truncation ratio. Compared with the prior art, by introducing the adjustment factor of the low-density area, the clustering centers in different density areas can be more accurately identified and distinguished. The effect is to improve the adaptability and accuracy of the clustering analysis, so as to more effectively identify the disturbance source in the power grid and optimize the power quality monitoring and maintenance strategy of the power grid.

[0043] For example, the target truncation ratio is 2%, which is set according to the prior art and will not be described in detail.

[0044] The calculation method of the adjustment factor of the low-density area includes: The average distance of the electrical distance between all terminal pairs in the partition is calculated, and the distance of the nearest neighbor point pair for density calculation is determined according to the target truncation ratio and the number of sample pairs. The difference between the average distance and the distance of the nearest neighbor point pair is normalized and added by 1, as the adjustment factor of the low-density area.

[0045] Specifically, the adjustment factor of the low-density area satisfies the following relationship: ; In the formula, the adjustment factor of the low-density area, the average value of the electrical distance between all intelligent fusion terminal pairs in the same partition, the target truncation ratio, the selected nearest neighbor distance, the number of sample pairs, the upward rounding symbol.

[0046] Specifically, the adaptive truncation distance satisfies the following relationship: ; In the formula, the adaptive truncation distance, the target truncation ratio, the selected nearest neighbor distance, the adjustment factor of the low-density area.

[0047] The difference between this embodiment and existing technologies lies in the fact that the influence weight of low-density regions during partitioning is analyzed based on electrical distance. By introducing an adjustment factor for low-density regions, the adaptive cutoff distance can adapt to power grid regions of different densities, thereby improving the accuracy of clustering analysis and the robustness of the algorithm. Specifically, the influence weight of electrical distance on low-density region partitioning is considered, optimizing the clustering results. This makes terminals within the same cluster more similar in power quality and electrical connectivity, facilitating more accurate identification and location of disturbance sources in the power grid, and enhancing the efficiency and reliability of power grid operation.

[0048] The remaining sample points are sorted in descending order of local density and then assigned to the clusters containing the nearest points with local densities greater than their own. This step is an existing technique.

[0049] Specifically, the local density satisfies the following relationship: ; In the formula, Indicates terminal Local density, Indicates terminal With terminal Similarity distance, Indicates the adaptive cutoff distance. Represented by natural numbers An exponential function with base 0.

[0050] Specifically, the relative distance satisfies the following relationship: terminal The formula for the relative distance to high-density terminals is: ; In the formula, Indicates terminal The minimum distance between it and other terminals with a higher density. Indicates terminal With terminal Similarity distance, Indicates terminal Local density, Indicates terminal Local density, This indicates taking the minimum value in the set.

[0051] For local density It is the largest terminal: By constructing " "Decision graph" adaptive selection and The terminals with significantly larger values ​​are used as cluster centers.

[0052] S4: filtering the clustering results, analyzing the dominance of each power quality feature, and determining whether it is a mixed cluster, re-clustering the mixed cluster based on the corresponding power quality feature to obtain significant clustering results; combining the significant clustering results and electrical distance, using the power quality attenuation law to obtain the location of the disturbance source.

[0053] In power quality analysis, the features are divided into two categories: harmonics and voltage sags, and the dominant degree is calculated respectively because these two features represent different physical phenomena, have different influences, consequences, monitoring methods and treatment measures.

[0054] Harmonics are mainly related to nonlinear loads, affecting power grid efficiency and equipment life, while voltage sags are usually caused by short circuits or large equipment start-up, which may cause sensitive equipment to malfunction. Distinguishing between these two features helps to more accurately identify power quality problems in the power grid, so that more effective and targeted treatment measures can be taken to improve the efficiency and reliability of the power grid.

[0055] In addition, this distinction is particularly important in clustering analysis, as it can identify and locate different types of disturbance sources, providing more targeted data support for the maintenance and optimization of the power grid.

[0056] The total number of intelligent fusion terminals contained in the statistical cluster is counted, and the number of harmonic features and voltage sag features exceeding the standard value of the corresponding feature is filtered out respectively, and the ratio to the total number is calculated to obtain the feature dominant degree.

[0057] If the feature dominant degree is less than or equal to the preset threshold, it is a single cluster of harmonic disturbance / voltage sag disturbance, otherwise, it is a mixed cluster of harmonic disturbance / voltage sag disturbance.

[0058] Specifically, the preset threshold is 80%, which can be adjusted according to specific circumstances.

[0059] If it is a mixed cluster, secondary clustering is needed, the power quality feature corresponding to the dominant disturbance type is selected, the irrelevant features are removed, and a specific distance function is applied for secondary clustering, so as to split the mixed cluster into multiple single sub-clusters. Not only improves the accuracy of disturbance source positioning, but also provides more targeted data support for power quality monitoring and maintenance of the power grid, especially suitable for handling complex power grid environments with multiple disturbance types.

[0060] Mixed clusters refer to clusters containing the influence of different disturbance sources, in which case the terminals in a cluster can be affected by multiple different disturbance sources. To improve the accuracy and reliability of the clustering results, further analysis and splitting of these mixed clusters is needed. By secondary clustering or other methods, mixed clusters are split into clusters of single disturbance sources, ensuring that each cluster corresponds to a single disturbance type. Accurate clustering results are crucial for subsequent disturbance source localization, and the presence of mixed clusters can reduce the accuracy of localization.

[0061] First, a harmonic disturbance event library and a voltage sag event library are established based on historical data and clustering results, recording the location of the disturbance source, the power grid topology, and the terminal's electrical energy characteristic values. Then, the electrical distance from each terminal to the disturbance source point is determined using an electrical distance calculation method, and the attenuation curve fitting is performed for harmonic and voltage sag disturbances respectively, obtaining their respective attenuation coefficients and attenuation models.

[0062] By selecting the terminal most affected by the disturbance as the backtracking point, and substituting the amplitude of the terminal into the fitted attenuation function, the electrical distance from the disturbance source to the backtracking point is calculated, and finally the location of the disturbance source is determined in combination with the power grid topology. This significantly improves the accuracy and efficiency of disturbance source localization, reduces the scope of field investigation, and achieves the core goal of the present application, which is to accurately locate the source of power quality disturbances based on intelligent fusion terminal data.

[0063] By fitting the attenuation curve and combining historical data and power grid topology, the approximate location of the disturbance source can be accurately determined. By analyzing the clustering results and electrical distance, the location of the disturbance source can be inferred, providing a scientific basis for the maintenance and optimization of the power grid.

[0064] The present application also provides an intelligent fusion terminal-based comprehensive power quality monitoring system. As shown in Figure 2 The system includes a processor and a memory, and the memory stores computer program instructions that, when executed by the processor, implement the intelligent fusion terminal-based comprehensive power quality monitoring method according to the first aspect of the present application. The system also includes a communication bus and a communication interface, as well as other components known to those skilled in the art, the settings and functions of which are known in the art, and therefore will not be described here.

[0065] It should be noted that for those skilled in the art, without departing from the concept of the present application, several modifications and improvements can be made, which are within the scope of the present application. Therefore, the scope of protection of the present application patent should be subject to the appended claims.

Claims

1. A power quality comprehensive monitoring method based on an intelligent fusion terminal, characterized in that, The method comprises the following steps: obtaining power quality data sets of intelligent fusion terminals and preprocessing the power quality data sets, dividing electrical partitions, and determining dominant harmonic components; calculating electrical distances between the intelligent fusion terminals in the same partition based on the dominant harmonic components in the preprocessed power quality data sets and electrical paths in the topology of the power grid; performing clustering on the intelligent fusion terminals in the same partition based on density peaks, calculating feature differences and electrical distances between the intelligent fusion terminals, weighting and fusing the feature differences and the electrical distances according to a preset weight distribution rule to obtain a similarity distance, determining an adaptive cut-off distance in combination with an adjustment factor of a low-density area, calculating local densities of the intelligent fusion terminals and relative distances between the intelligent fusion terminals and other intelligent fusion terminals with higher densities, drawing a decision graph based on the local densities and the relative distances, identifying density peak points to determine clustering centers, and obtaining a preliminary clustering result in the same partition; screening the clustering result, analyzing dominant degrees of each power quality feature, and determining whether it is a mixed cluster, re-clustering the mixed cluster based on corresponding power quality features to obtain a significant clustering result; obtaining a position of a disturbance source in combination with the significant clustering result and the electrical distances and using a power quality attenuation rule. The preset weight distribution rule is that a weight coefficient of the similarity distance is obtained by dividing a standard deviation of the feature differences of the power quality of all terminals in the same partition by a sum of the standard deviation of the feature differences of the power quality and a standard deviation of the electrical distances, and a weight coefficient of the electrical distances is obtained by subtracting the weight coefficient of the similarity distance from 1.

2. The method according to claim 1, wherein, The dominant harmonic component comprises the following steps: calculating energy proportions of each harmonic component by dividing squares of voltage amplitudes of each harmonic by an accumulated sum of squares of voltage amplitudes of all harmonics, selecting a harmonic component with the largest energy proportion as a dominant harmonic component, and identifying the dominant harmonic component according to the dominant harmonic component. 3.The method of claim 1, wherein, The calculation of the electrical distance comprises the following steps: taking any terminal as a target terminal, obtaining a plurality of electrical paths between the target terminal and other terminals, wherein each electrical path comprises a plurality of conductor lines and a plurality of transformers; summing resistive components and inductive components of all conductor lines in the electrical path, adding the summing results to obtain a total line impedance of the electrical path, summing short-circuit impedances of all transformers in the electrical path to obtain a total transformer impedance of the electrical path, summing the total line impedance and the total transformer impedance, and taking a complex modulus value to obtain the electrical distance between the target terminal and other terminals.

4. The method according to claim 3, wherein, The steps of obtaining the resistive components and the inductive components comprise the following steps: Taking any conductor as a target conductor, a complex modulus value between a unit length resistance of a conductor line of the target conductor and a length of the conductor line of the target conductor is calculated to obtain a resistance component of the target conductor; an imaginary unit in the complex number is replaced by a real number, A complex modulus value between a corresponding unit length inductance and a length of the line of the main harmonic component frequency multiplied by two of the conductor line is obtained to obtain a reactance component of the target conductor.

5. The method according to claim 1, wherein, The steps of obtaining the feature differences comprise the following steps: constructing a power quality feature vector for each terminal in the same partition, wherein the power quality feature vector comprises a voltage deviation, a frequency deviation, a total harmonic distortion rate, a voltage amplitude of the dominant harmonic component, a voltage sag amplitude, and a sag duration; taking any intelligent fusion terminal as a target terminal, calculating Euclidean distances between power quality feature vectors of two intelligent fusion terminals in the same partition to obtain feature differences.

6. The method for monitoring power quality according to claim 1, wherein, The calculation method of the adjustment factor of the low-density area comprises the following steps: The average distance of electrical distance between all terminal pairs in the partition is calculated, and the distance of the nearest neighbor pair for density calculation is determined according to the target truncation ratio and the number of sample pairs; the difference between the average distance and the distance of the nearest neighbor pair is normalized and added by 1 as the adjustment factor of the low-density area.

7. The method according to claim 1, wherein the method further comprises: The dominance degree of the power quality feature includes: The total number of intelligent fusion terminals contained in the cluster is counted, the number of terminals with the preset power quality feature exceeding the standard value of the corresponding feature is screened out, and the ratio to the total number is calculated to obtain the feature dominance degree. 8.The method of claim 1, wherein, The position of the disturbance source is obtained by using the power quality attenuation law, including: A harmonic disturbance event library and a voltage sag event library are established according to historical data and clustering results, and the position of the disturbance source, the power grid topology structure and the terminal power feature value are recorded; The electrical distance from each terminal to the disturbance source point is determined by using the electrical distance calculation method, and the attenuation curve fitting is performed on the harmonic and voltage sag disturbances respectively to obtain the respective attenuation coefficients and attenuation models; The intelligent fusion terminal most affected by the disturbance is selected as the backtracking point, and the amplitude of the intelligent fusion terminal is substituted into the fitted attenuation model to calculate the electrical distance from the disturbance source to the backtracking point, and finally the position of the disturbance source is determined combined with the power grid topology structure. 9.The smart fusion terminal based power quality comprehensive monitoring method of claim 1, wherein, The power quality data set includes: voltage deviation, frequency deviation, total harmonic distortion, voltage amplitude, instantaneous voltage value, voltage sag duration and voltage sag amplitude; According to the power grid topology structure, the electrical partition is divided, and for the power quality data set after the electrical partition, according to the preset abnormality judgment standard, the index abnormal terminal is screened out, and according to the terminal data in the same electrical partition, the mean value and standard deviation of each power quality data set are calculated, when the absolute difference value of each power quality data set and the mean value is greater than 3 times the standard deviation, it is determined that the feature value is an abnormal value, and is rejected; the power quality data set is normalized by maximum and minimum, and the dimension difference is eliminated.

10. The power quality comprehensive monitoring system based on intelligent fusion terminal, characterized in that, It includes: A processor and a memory, the memory stores computer program instructions, when the computer program instructions are executed by the processor, the intelligent fusion terminal based power quality comprehensive monitoring method according to any one of claims 1-9 is realized.

Images