Feeder adjustment characteristic modeling and parameter identification method, system and device considering voltage sensitivity and medium

Abstract

The invention relates to the technical field of power system operation and control, and discloses a voltage sensitivity-considered feeder adjustment characteristic modeling and parameter identification method, system and device and a medium, and the method comprises the following steps: fully considering the complex dynamic characteristics of feeder load through multi-step data processing and an advanced parameter identification method; in the aspect of data processing, environmental factor trend term power data are firstly removed, voltage power data strongly related to voltage fluctuation caused by load change are accurately extracted, and a high-quality data basis is provided for subsequent modeling. In the aspect of modeling, a step-down energy-saving coefficient model is established, and the adjustment potential of the feeder load can be accurately represented. And in a parameter identification link, a robust least square identification method based on random sampling consistency is combined, an optimal model is screened out through multi-round iteration, and model parameters are dynamically adjusted by adopting a sliding time window mechanism, so that the model can adapt to changes of load characteristics and operation conditions.

Description

Technical Field

[0001] This invention relates to the field of power system operation and control technology, and in particular to a method, system, device and medium for modeling and identifying feeder regulation characteristics and parameters that take into account voltage sensitivity. Background Technology

[0002] With the deepening of the construction of new power systems, the scale of distributed energy generation, represented by photovoltaic and wind power, has experienced explosive growth. However, new energy sources are characterized by significant intermittency and volatility, and their large-scale grid connection will increase the uncertainty on both the power system source and load sides, posing difficulties for the safe and stable operation of the power grid.

[0003] Against this backdrop, fully leveraging the flexible regulation resources on the load side can enhance the grid's regulation and control capabilities and the capacity for renewable energy absorption to a certain extent. As a key controllable resource in the distribution network, feeder loads, relying on their voltage sensitivity characteristics, can achieve precise control of active power based on Conservation Voltage Reduction (CVR) technology, thereby effectively improving the power system's source-load synergy and interaction capabilities. However, the composition of feeder loads is highly complex, including various load types such as industrial motors, commercial air conditioners, and residential appliances, with significant differences in the voltage sensitivity characteristics of these types of loads. Secondly, the system's operating state exhibits strong time-varying characteristics; the dynamic changes in load composition ratios, operating conditions, and environmental factors lead to time-varying overall load regulation characteristics.

[0004] This complex dynamic characteristic makes it difficult for traditional modeling methods based on static assumptions to accurately characterize the actual regulation behavior of feeder loads. There is an urgent need to explore dynamic parameter identification methods that can take into account the coupling effects of multiple factors. Summary of the Invention

[0005] In view of the aforementioned existing problems, the present invention is proposed.

[0006] Therefore, the present invention provides a method, system, device and medium for modeling and parameter identification of feeder regulation characteristics that takes into account voltage sensitivity, which can solve the problem that traditional modeling methods based on static assumptions in the prior art are difficult to accurately characterize the actual regulation behavior of feeder loads.

[0007] To solve the above-mentioned technical problems, the present invention provides the following technical solution:

[0008] In a first aspect, the present invention provides a method for modeling and identifying parameters of feeder regulation characteristics that take into account voltage sensitivity, including:

[0009] Obtain feeder load-related data of the target distribution network and perform a rejection operation on the feeder load-related data;

[0010] The removal operation is used to remove power data from the environmental factor trend item in the feeder load related data;

[0011] Extract voltage and power data that are strongly correlated with the voltage fluctuation process caused by load changes from the feeder load-related data after the removal operation, and use them as power identification data.

[0012] Obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and establish a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads;

[0013] Based on the aforementioned voltage reduction energy-saving coefficient model and power identification data, parameter identification is performed using a robust least squares identification method based on random sampling consistency.

[0014] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, the rejection operation includes:

[0015] Filtering methods are used to process feeder load-related data of the target distribution network;

[0016] The voltage-sensitive power component is extracted by separating the high-frequency power component of voltage fluctuation from the low-frequency trend term of environmental factors through sliding window polynomial fitting.

[0017] Subtract the fitted data of the non-voltage trend term from the original power data to obtain the voltage-sensitive power data sequence, and use the voltage-sensitive power data sequence as the voltage power data.

[0018] This preferred scheme can more accurately focus on power data closely related to voltage fluctuations. Traditional modeling methods often ignore the interference of environmental factors on feeder load data, making it difficult for the model to accurately reflect the actual regulation behavior of the feeder load. This preferred scheme, however, effectively removes some noise interference and improves data quality by using filtering methods to process the data. The sliding window polynomial fitting cleverly separates the high-frequency power components of voltage fluctuations from the low-frequency trend terms of environmental factors, allowing us to extract the power components truly related to voltage sensitivity. By subtracting the fitted data for non-voltage trend terms from the original power data, the resulting voltage-sensitive power data sequence more purely reflects the process of voltage fluctuations caused by load changes. This provides more reliable data support for the subsequent voltage reduction energy-saving coefficient model. When using this model in conjunction with a robust least squares identification method based on random sampling consistency for parameter identification, more accurate parameter results can be obtained. This makes the modeling of the entire feeder regulation characteristics considering voltage sensitivity more consistent with reality, providing strong support for solving the problem that traditional modeling methods based on static assumptions cannot accurately characterize the actual regulation behavior of feeder loads, and contributing to improving the operating efficiency and energy-saving effect of the distribution network.

[0019] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, wherein: the voltage power data strongly correlated with the process of voltage fluctuation caused by load change in the feeder load-related data after the removal operation is extracted, and the power identification data includes:

[0020] A correlation acquisition model is established, and the correlation of the voltage and power data is calculated based on the correlation acquisition model.

[0021] A pre-defined correlation judgment strategy is used to filter voltage and power data that are strongly correlated with the voltage fluctuations caused by load changes.

[0022] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, the step of obtaining the voltage-power coupling characteristics of the feeder load of the target distribution network and establishing a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder load includes:

[0023] The voltage reduction energy saving coefficient model is used to characterize the regulation potential of the feeder load. The larger the value, the more significant the power response of the feeder load when the voltage changes, and the stronger the regulation capability.

[0024] The voltage reduction energy saving coefficient model is obtained through the change in active power of the feeder load, the initial active power of the feeder load, the change in feeder voltage, and the initial voltage of the feeder.

[0025] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, the parameter identification based on the step-down energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, includes:

[0026] The power identification data is subjected to time-series differential processing to extract the load power response features corresponding to the voltage change range;

[0027] Based on the voltage variation range and power response characteristics, a sequence of parameters to be identified is constructed to characterize the regulation capability;

[0028] The random sampling consensus method is used to initially screen the parameter sequence to be identified, removing data points that are affected by abnormal disturbances or noise, and retaining the set of interior points that conform to the model trend;

[0029] Perform a least-squares fitting operation on the set of interior points to obtain the estimated values ​​of stable parameters reflecting load regulation characteristics in the voltage reduction energy-saving coefficient model.

[0030] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, the parameter identification based on the step-down energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, further includes:

[0031] In each parameter identification process, a minimum number of data samples and a maximum number of iterations are set;

[0032] In each iteration, a number of data points are randomly selected to construct a temporary parametric model, and the degree of matching between the remaining data points and the model is evaluated.

[0033] Data points with a matching degree higher than a preset threshold are marked as inliers, and the model with the most inliers in each round of iteration is the optimal model.

[0034] Re-execute least squares fitting using the set of interior points corresponding to the optimal model, and output the final optimized model parameters.

[0035] As a preferred embodiment of the feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity described in this invention, the parameter identification based on the step-down energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, further includes:

[0036] A sliding time window mechanism is adopted to perform periodic rolling updates, so that the model parameters are dynamically adjusted as load characteristics and operating conditions change;

[0037] After each window slide, the entire process of data culling, feature extraction, model building, and parameter identification is re-executed.

[0038] Secondly, the present invention provides a feeder regulation characteristic modeling and parameter identification system considering voltage sensitivity, comprising:

[0039] The data acquisition and processing module is used to acquire feeder load-related data of the target distribution network and to perform a rejection operation on the feeder load-related data.

[0040] The removal operation is used to remove power data from the environmental factor trend item in the feeder load related data;

[0041] The extraction module is used to extract voltage and power data that are strongly correlated with the voltage fluctuation caused by load changes from the feeder load-related data after the removal operation, and use it as power identification data.

[0042] The model building module is used to obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and to build a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads.

[0043] The solution module is used to identify parameters based on the voltage reduction energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency.

[0044] Thirdly, the present invention provides an electronic device including a memory and a processor, wherein the memory stores a computer program, and the processor executes the computer program to implement the steps of the method described above.

[0045] Fourthly, the present invention provides a computer-readable storage medium having a computer program stored thereon, which, when executed by a processor, implements the steps of the method described above.

[0046] Compared with existing technologies, the beneficial effects of this invention are as follows: This invention proposes a method for modeling and identifying feeder regulation characteristics that considers voltage sensitivity. Through multi-step data processing and advanced parameter identification methods, it fully considers the complex dynamic characteristics of feeder loads. In data processing, power data with environmental factor trends are first removed, and voltage and power data strongly correlated with voltage fluctuations caused by load changes are accurately extracted, providing a high-quality data foundation for subsequent modeling. In modeling, a voltage reduction energy-saving coefficient model is established, which can accurately characterize the regulation potential of feeder loads. In the parameter identification stage, a robust least squares identification method based on random sampling consistency is combined, and the optimal model is selected through multiple rounds of iteration. A sliding time window mechanism is used to dynamically adjust the model parameters, enabling the model to adapt to changes in load characteristics and operating conditions.

[0047] This invention effectively solves the problem that traditional modeling methods based on static assumptions cannot accurately characterize the actual regulation behavior of feeder loads. Through this invention, the regulation characteristics of feeder loads can be more accurately grasped, enabling precise control of the active power of feeder loads. This, in turn, effectively improves the power system's source-load synergy and interaction capabilities, enhances grid regulation and control capabilities and the capacity for renewable energy absorption, and provides strong support for the safe and stable operation of new power systems. Attached Figure Description

[0048] To more clearly illustrate the technical solutions of the embodiments of the present invention, the drawings used in the following description of the embodiments will be briefly introduced. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0049] Figure 1 The flowchart illustrates a method for modeling and identifying the parameters of a feeder regulation characteristic that takes into account voltage sensitivity, as provided in one embodiment of the present invention.

[0050] Figure 2The identification results of three strategies for a feeder regulation characteristic modeling and parameter identification method considering voltage sensitivity provided in an embodiment of the present invention are as follows: (a) CVR coefficient identification results of each strategy from 15 to 30 minutes, and (b) CVR coefficient identification results of each strategy from 30 to 45 minutes.

[0051] Figure 3 This is a comparison chart of power changes of various strategies under voltage disturbances, provided by a method for modeling and identifying feeder regulation characteristics and parameters considering voltage sensitivity in an embodiment of the present invention.

[0052] Figure 4 This is an internal structure diagram of an electronic device that includes a method for modeling and identifying the parameters of a feeder regulation characteristic that takes into account voltage sensitivity, as provided in an embodiment of the present invention. Detailed Implementation

[0053] To make the above-mentioned objects, features, and advantages of the present invention more apparent and understandable, specific embodiments of the present invention will be described in detail below with reference to the accompanying drawings. Obviously, the described embodiments are only a part of the embodiments of the present invention, and not all of them. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort should fall within the protection scope of the present invention.

[0054] Example 1, referring to Figure 1 This is the first embodiment of the present invention, which provides a method for modeling and identifying the parameters of feeder regulation characteristics that take into account voltage sensitivity, including:

[0055] Existing technologies have several drawbacks. For example, traditional feeder load modeling methods often fail to adequately consider the complex dynamic characteristics of the load, relying solely on static assumptions, leading to significant discrepancies between the model and actual conditions. Some methods do not effectively eliminate environmental interference with feeder load data during processing, making it difficult for the model to accurately reflect the actual regulation behavior of the feeder load. Some parameter identification methods lack the ability to handle abnormal disturbances and noise, making them susceptible to the influence of anomalous data and resulting in inaccurate identification results. Furthermore, existing models cannot adequately adapt to dynamic changes in load characteristics and operating conditions, hindering precise control of the active power of the feeder load.

[0056] This invention provides a method that can effectively solve the problems mentioned above. The following will describe in detail how to implement the method for modeling and identifying the feeder regulation characteristics that take into account voltage sensitivity, using multiple embodiments.

[0057] Figure 1 A flowchart illustrating a method for modeling and identifying the parameters of feeder regulation characteristics that take into account voltage sensitivity is shown, including:

[0058] S101, Obtain feeder load-related data of the target distribution network, and perform a data removal operation on the feeder load-related data, wherein:

[0059] It should be noted that in order to model and identify the feeder regulation characteristics taking voltage sensitivity into account, it is necessary to first obtain feeder load-related data of the target distribution network. This data contains various information about the feeder load during operation and forms the basis for subsequent analysis and modeling. The reason for the removal operation is that the original data may contain power data of environmental factor trend terms, which can interfere with the accurate analysis of the actual regulation behavior of the feeder load.

[0060] It should be noted that changes in feeder load power are not only affected by voltage fluctuations, but also closely related to slow time-varying environmental factors such as ambient temperature and relative humidity. These environmental parameters exhibit significant low-frequency variation characteristics, with time scales typically on the order of hours. In contrast, grid voltage fluctuations exhibit relatively high-frequency dynamic characteristics, mostly on the order of minutes or even seconds.

[0061] Therefore, it is necessary to effectively separate the high-frequency power component of voltage fluctuations from the low-frequency trend term of environmental factors.

[0062] In this embodiment of the invention, the elimination operation is used to eliminate power data of the environmental factor trend item in the feeder load related data.

[0063] In one alternative implementation, the elimination operation can employ a combination of multiple filtering algorithms. For example, a median filter algorithm can be used first to pre-process the feeder load-related data of the target distribution network. Median filtering effectively removes impulse noise from the data, making it smoother. Then, a Kalman filter algorithm is used to further optimize the data. The Kalman filter can update and predict in real time based on the dynamic characteristics of the data, thereby more accurately estimating the true value of the data. By combining these two filtering algorithms, noise interference in the data can be removed more comprehensively, improving the data quality.

[0064] In an optional implementation, an adaptive sliding window technique can be used when separating the high-frequency power component of voltage fluctuations from the low-frequency trend term of environmental factors using a sliding window polynomial fitting. The size of the sliding window is dynamically adjusted based on real-time data changes. When data changes drastically, the window size is appropriately reduced to more sensitively capture the high-frequency power component of voltage fluctuations; when data changes are relatively stable, the window size is increased to more accurately separate the low-frequency trend term of environmental factors. This improves the accuracy of the fitting and more precisely extracts the voltage-sensitive power component.

[0065] In one optional implementation, after obtaining the voltage-sensitive power data sequence, it can be normalized. Each data point in the voltage-sensitive power data sequence is divided by the maximum value of the sequence, mapping the data to the interval [0,1]. Normalization eliminates the influence of data dimensions, making data of different scales comparable, which is beneficial for subsequent establishment of a voltage reduction energy-saving coefficient model and parameter identification. Simultaneously, the normalized data is more stable during calculations, improving the accuracy and reliability of the model.

[0066] In this embodiment of the invention, the rejection operation includes:

[0067] Filtering methods are used to process feeder load-related data of the target distribution network;

[0068] The voltage-sensitive power component is extracted by separating the high-frequency power component of voltage fluctuation from the low-frequency trend term of environmental factors through sliding window polynomial fitting.

[0069] Subtract the fitted data of the non-voltage trend term from the original power data to obtain the voltage-sensitive power data sequence, and use the voltage-sensitive power data sequence as the voltage power data.

[0070] It should be noted that filtering methods refer to the use of mathematical algorithms or signal processing techniques to remove noise or extract specific components from data. Here, it refers to processing data from the load cells of a distribution network feeder, which may be used to eliminate interference or extract key information.

[0071] For example, suppose the load data of a distribution network feeder contains random fluctuations and periodic interference. Using a low-pass filter can remove high-frequency noise and retain low-frequency trends, thus obtaining a smoother load curve.

[0072] It should be noted that a sliding window is an analysis method based on local data segments, while polynomial fitting uses a polynomial function to approximate the trend of data changes. This sentence refers to using polynomial fitting within a sliding window to separate rapid changes (high-frequency components) and slow changes (low-frequency trend terms, such as the influence of environmental factors like temperature and humidity) in voltage fluctuations.

[0073] For example, when a sliding window is used to fit a cubic polynomial to voltage data collected over a certain period of time, it is found that the high-frequency components exhibit small-amplitude rapid oscillations, while the low-frequency trend term shows a slowly rising trend curve.

[0074] It should be noted that voltage-sensitive power components refer to the power portion that is closely related to voltage changes and is usually directly affected by voltage fluctuations. This sentence refers to separating this part of the power data from the raw data.

[0075] For example, if the power of a device increases significantly with increasing voltage, this power is the voltage-sensitive power component, which can be extracted by calculating the region with a large slope of the power-voltage relationship curve.

[0076] It should be noted that the voltage-sensitive power data obtained after the above processing is regarded as the main power data reflecting voltage changes and used for subsequent analysis or modeling.

[0077] For example, after processing, the resulting voltage-sensitive power data sequence is [10,12,15,18], which can be directly used to analyze the impact of voltage fluctuations on the system.

[0078] Specifically, this invention uses the Savitzky-Golay (SG) filtering method to process measurement data. By using a sliding window polynomial fitting, it effectively separates the high-frequency power component of voltage fluctuations from the low-frequency trend term of environmental factors, and accurately extracts the voltage-sensitive power component.

[0079] Furthermore, the filter window width is set to 2m+1, and a k-1 degree polynomial is used to fit the data points within the window, resulting in the following polynomial fitting model:

[0080] y = a0 + a1x + a2x 2 +...+a k-1 x k-1

[0081] In the formula, x represents the non-voltage trend term data; y represents the fitted output data; and a represents the parameter to be solved.

[0082] Furthermore, the system of k equations consisting of 2m+1 equations is written in matrix form, and the parameters A = (a0,...,a1) are determined using the least squares method. k-1 ) T Find the least squares solution and obtain the model filter value:

[0083] Y = XA + B

[0084]

[0085] Furthermore, by subtracting the fitted data of the non-voltage trend terms from the original power data, the voltage-sensitive power data sequence can be obtained.

[0086] It should be noted that acquiring feeder load-related data from the target distribution network and then removing this data provides a high-quality and accurate data foundation for subsequent voltage and power data extraction and the establishment of the step-down energy-saving coefficient model. After removing power data related to environmental factors, the extracted power identification data can more purely reflect the process of voltage fluctuations caused by load changes, enabling the step-down energy-saving coefficient model to more accurately characterize the voltage-power coupling characteristics of the feeder load.

[0087] S102, Extract voltage and power data strongly correlated with the voltage fluctuation process caused by load changes from the feeder load-related data after the removal operation, and use it as power identification data, wherein:

[0088] It should be noted that after obtaining the feeder load-related data after the removal process, it is necessary to further filter out the voltage and power data that are strongly correlated with the voltage fluctuations caused by load changes. This is because after removing the power data related to environmental factors, there may still be some data in the data that are not closely related to the voltage fluctuations caused by load changes. These data will affect the accuracy of subsequent parameter identification.

[0089] In one alternative implementation, correlation analysis can be used to extract power identification data. By calculating the correlation coefficient between each data point and the voltage fluctuation process caused by load changes, a suitable threshold is set, and data with correlation coefficients higher than this threshold are filtered out. For example, the Pearson correlation coefficient can be used to measure the linear correlation between data points; when the correlation coefficient is greater than 0.8, the data is considered to be strongly correlated with the voltage fluctuation process caused by load changes and is thus used as power identification data.

[0090] In another alternative implementation, feature selection algorithms from machine learning can be incorporated. For example, a random forest algorithm can be used to evaluate the feature importance of feeder load-related data after the culling operation. The random forest algorithm determines the importance of each feature based on its contribution to the decision tree construction process. Selecting the data corresponding to the features with the highest importance ranking as power identification data can more effectively filter out data strongly correlated with voltage fluctuations caused by load changes.

[0091] It should be noted that the extracted power identification data is a crucial input for subsequent parameter identification. This data accurately reflects the actual changes in feeder load under voltage fluctuations, providing reliable data support for establishing a more accurate step-down energy-saving coefficient model and conducting parameter identification. Only based on high-quality power identification data can the model more accurately characterize the regulation characteristics of the feeder load, thereby achieving precise control of the active power of the feeder load.

[0092] In this embodiment of the invention, voltage and power data strongly correlated with the voltage fluctuation process caused by load changes are extracted from the feeder load-related data after the removal operation and used as power identification data, including:

[0093] Establish a correlation acquisition model, and calculate the correlation of voltage and power data based on the correlation acquisition model;

[0094] A pre-defined correlation judgment strategy is used to filter voltage and power data that are strongly correlated with the voltage fluctuations caused by load changes.

[0095] It should be noted that establishing a relevance acquisition model refers to constructing a computational model that can quantify the strength of the association between different data using mathematical or statistical methods. This model can be based on regression analysis, machine learning algorithms (such as decision trees and neural networks), or other data analysis techniques.

[0096] For example, a model is built using the Pearson correlation coefficient formula to calculate the linear correlation between voltage and power data.

[0097] It should be noted that the collected voltage and power data are input into the established correlation acquisition model, which then outputs the correlation values ​​between each set of data. These values ​​are typically represented as scores between 0 and 1, with the closer to 1 indicating a higher correlation.

[0098] For example, suppose there is a set of voltage data [220V, 230V, 215V] and a set of power data [5kW, 6kW, 4.8kW]. The correlation between them is calculated to be 0.92 by the model.

[0099] It should be noted that a pre-defined relevance assessment strategy can establish a set of rules or standards to determine which data meet specific relevance requirements. Such strategies may include setting thresholds (e.g., a relevance greater than 0.8 is considered a strong correlation), category labels, or other filtering conditions.

[0100] For example, it is stipulated that only when the correlation calculation result exceeds 0.8 are the two sets of data considered to be strongly correlated, and then these data are retained for subsequent analysis.

[0101] It should be noted that data points that significantly reflect voltage fluctuations when the load changes are selected from all voltage and power data. This requires consideration of the actual application scenario to ensure that the selected data effectively describes the target phenomenon.

[0102] For example, in one experiment, multiple sets of voltage and power data were recorded. Some data points showed that when the load increased from 5kW to 10kW, the voltage dropped from 220V to 210V. Based on a correlation judgment strategy, these data were identified as strongly correlated and selected for further investigation into the causes of voltage fluctuations.

[0103] Specifically, in actual power systems, the coupling relationship between voltage and power includes the VTL (voltage-to-load) process and the LTV (load-to-voltage) process. The VTL process reflects voltage fluctuations caused by load changes through line voltage drops, with power and voltage exhibiting a negative correlation. The LTV process, on the other hand, reflects the impact of active voltage regulation on load power, showing a positive correlation. It is important to note that only data related to the LTV process is suitable for load characteristic parameter identification. To achieve effective data selection, Pearson correlation coefficients are used for quantitative analysis, extracting data relevant to the LTV process for parameter identification.

[0104]

[0105] In the formula, r is the Pearson coefficient, which measures the strength of the linear correlation between variables X and Y. i With Y i These are the data points for these two variables, and These are the average values ​​of the two variables, and n is the data sample size.

[0106] In one optional implementation, the preset relevance judgment strategy can be designed as follows:

[0107] When -0.3≤r≤0.3, it indicates that there is no linear correlation between power and voltage during this period, mainly caused by the transient process of starting and stopping electrical equipment and the measurement noise of related data;

[0108] When r < -0.3, power and voltage have a certain negative correlation, indicating that the data in this period is related to the VTL process;

[0109] When r>0.3, it indicates that the data during this period is related to the LTV process associated with the feeder load regulation characteristics, and can be used for feeder load model parameter identification.

[0110] It should be noted that extracting voltage and power data strongly correlated with the voltage fluctuations caused by load changes from the feeder load-related data after the removal process, and using this as power identification data, allows the subsequently established step-down energy-saving coefficient model to more accurately reflect the actual situation of the feeder load. By selecting data strongly correlated with voltage fluctuations caused by load changes, the model can avoid interference from irrelevant data, thereby more accurately characterizing the voltage-power coupling characteristics of the feeder load.

[0111] S103, obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and establish a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads, wherein:

[0112] It is important to note that after obtaining the power identification data, it is necessary to conduct in-depth analysis of the voltage-power coupling characteristics of the feeder loads in the target distribution network. This characteristic reflects the intrinsic relationship between voltage and power changes, and is crucial for understanding the operating patterns of the feeder loads. The characteristics of voltage-power coupling can be comprehensively grasped through multi-dimensional analysis of the power identification data, such as time-domain analysis and frequency-domain analysis.

[0113] In one alternative implementation, the voltage-power coupling characteristics of the feeder load can be obtained by constructing a mathematical model based on historical power identification data. Regression analysis is used, with voltage as the independent variable and power as the dependent variable, to establish a functional relationship between the two. For example, a linear regression model can be used, and the least squares method can be used to fit a linear equation between voltage and power; the coefficients of the equation reflect the degree of influence of voltage on power.

[0114] In an alternative implementation, machine learning algorithms, such as neural network models, can also be incorporated. Power identification data is input into the neural network as a training set, and through multiple iterations of training, the network learns the complex nonlinear relationship between voltage and power. The trained neural network model can predict the corresponding power value based on the input voltage data, thus reflecting the voltage-power coupling characteristics of the feeder load.

[0115] It should be noted that after obtaining the voltage-power coupling characteristics of the feeder load, a step-down energy-saving coefficient model can be established based on this. The core of this model is to quantify the relationship between voltage drop and power reduction, that is, to calculate the proportion of power reduction in the feeder load under different voltage drop magnitudes.

[0116] In this embodiment of the invention, obtaining the voltage-power coupling characteristics of the feeder loads of the target distribution network and establishing a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads includes:

[0117] The step-down energy-saving coefficient model is used to characterize the regulation potential of feeder loads. The larger the value, the more significant the power response of the feeder load when the voltage changes, and the stronger the regulation capability.

[0118] The step-down energy-saving coefficient model is obtained by measuring the change in active power of the feeder load, the initial active power of the feeder load, the change in feeder voltage, and the initial voltage of the feeder.

[0119] It should be noted that voltage-power coupling characteristics describe the relationship between voltage and power in the load of a distribution network feeder. Under this relationship, changes in voltage will cause corresponding changes in power (such as active power), and these changes follow certain patterns.

[0120] For example, in a certain distribution network feeder, when the feeder voltage drops from 10kV to 9.5kV, the active power of the feeder load decreases from 500kW to 480kW. This phenomenon, where a voltage change leads to a change in active power, reflects the voltage-power coupling characteristics of the feeder load.

[0121] It should be noted that the voltage reduction energy-saving coefficient model is a mathematical model used to quantify the change in active power of the feeder load when the voltage decreases, thereby reflecting the energy-saving potential of the feeder load under voltage regulation. This coefficient can represent the sensitivity of the feeder load to voltage changes and the potential energy-saving effect.

[0122] For example, suppose a feeder load has an initial active power of 1000kW and an initial voltage of 10kV. When the voltage drops by 0.5kV, the active power decreases by 50kW. Based on the construction method of the voltage reduction energy-saving coefficient model, using these data (active power change of 50kW, initial active power of 1000kW, voltage change of 0.5kV, and initial voltage of 10kV), a voltage reduction energy-saving coefficient value can be calculated. If this value is large, it indicates that the feeder load has a significant power response when the voltage decreases, and possesses a strong ability to achieve energy savings through voltage reduction.

[0123] Specifically, the voltage-power coupling characteristics of feeder loads are a crucial foundation for distribution network feeder regulation. Although feeder loads are complex and diverse (covering various types including residential, commercial, industrial, and agricultural loads), they all exhibit certain voltage-power coupling characteristics. Typically, traditional constant-impedance loads (such as electric heating equipment and incandescent lamps) exhibit typical quadratic voltage-power characteristics, showing high sensitivity to voltage changes; while constant-power loads (such as frequency converters and switching power supplies) maintain constant power through internal control, exhibiting weaker voltage dependence. To accurately characterize this characteristic, the ZIP load model is commonly used in engineering.

[0124]

[0125] In the formula P u V0 is the active power of the load; V0 is the initial voltage, and P0 is the load power corresponding to V0. The three terms in the formula represent the constant impedance, constant current, and constant power load, respectively, A. p B p C p The proportions of the three types of loads.

[0126] Furthermore, traditionally, feeder load models are generally represented by the voltage reduction factor (CVR coefficient) model, which is defined as the ratio of the percentage change in load active power to the percentage change in voltage. This model characterizes the steady-state regulation characteristics of the load, and the magnitude of the CVR coefficient directly represents the regulation potential of the feeder load. The larger the value, the more significant the power response of the feeder load to voltage changes, and the stronger its regulation capability.

[0127]

[0128] Furthermore, CVR is the CVR coefficient of the feeder load, ΔP% is the percentage change in active power of the feeder load, ΔP% is the percentage change in feeder voltage, ΔP is the change in active power of the feeder load, P0 is the initial active power of the feeder load, ΔV is the change in feeder voltage, and V0 is the initial voltage of the feeder.

[0129] It should be noted that, on the one hand, the differences in the composition and proportion of feeder load at different times will cause the CVR coefficient to have significant time-varying characteristics;

[0130] Furthermore, on the other hand, there is a coupling relationship between the CVR coefficient and the load voltage level, and the feeder load regulation characteristics also differ significantly under different operating conditions. Based on the traditional ZIP load model, the derivative can be obtained as follows:

[0131] CVR f =ΔP% / ΔU%=2A P U%+B P

[0132] Furthermore, once the load type is determined, both AP and BP are constants, and U% is the load voltage percentage, which is the ratio of the actual voltage to the rated voltage.

[0133] It should be noted that, as shown in the above formula, the CVR coefficient of a certain static load is a linear function related to voltage, and this needs to be considered in the subsequent identification process.

[0134] S104, based on the step-down energy-saving coefficient model and power identification data, combines a robust least squares identification method based on random sample consistency to identify parameters, wherein:

[0135] It should be noted that once the step-down energy-saving coefficient model and power identification data are obtained, parameter identification can be performed.

[0136] In some alternative implementations, a robust least squares identification method based on random sample consensus can be employed. This method combines the robustness of the random sample consensus algorithm with the accuracy of least squares estimation, effectively handling data containing noise and outliers. The random sample consensus algorithm fits a model by randomly selecting a subset of data points, then calculates the error of other data points relative to the model. Data points with errors less than a certain threshold are identified as interior points. This process is repeated multiple times, and the model with the most interior points is selected as the final model.

[0137] In some alternative implementations, other advanced parameter identification algorithms, such as genetic algorithms and particle swarm optimization, can also be employed. A genetic algorithm is an optimization algorithm that simulates natural selection and genetic mechanisms. It searches for the optimal solution in the solution space by simulating selection, crossover, and mutation operations in biological evolution. Using the parameters of the voltage reduction energy-saving coefficient model as variables to be optimized, the genetic algorithm iteratively updates these parameters to minimize the error between the model's output and the power identification data.

[0138] Particle swarm optimization (PSO) seeks optimal solutions by simulating the collective behavior of flocks of birds or schools of fish. Each particle represents a possible solution, flying through the solution space and adjusting its flight direction and speed based on its own experience and the experience of the group. During parameter identification, the position of each particle is mapped to a set of parameters in the voltage reduction energy-saving coefficient model. By continuously updating the particle positions, the optimal parameter values ​​are gradually approximated.

[0139] In this embodiment of the invention, parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency, includes:

[0140] Time-series differential processing is performed on the power identification data to extract the load power response characteristics corresponding to the voltage variation range;

[0141] Based on the voltage variation range and power response characteristics, a sequence of parameters to be identified is constructed to characterize the regulation capability;

[0142] The random sampling consistency method is used to initially screen the parameter sequence to be identified, removing data points that are affected by abnormal disturbances or noise, and retaining the set of interior points that conform to the model trend;

[0143] Perform a least-squares fitting operation on the set of interior points to obtain the estimated values ​​of stable parameters reflecting load regulation characteristics in the voltage reduction energy-saving coefficient model.

[0144] In this embodiment of the invention, parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency, further includes:

[0145] In each parameter identification process, a minimum number of data samples and a maximum number of iterations are set;

[0146] In each iteration, a number of data points are randomly selected to construct a temporary parametric model, and the degree of matching between the remaining data points and the model is evaluated.

[0147] Data points with a matching degree higher than a preset threshold are marked as inliers, and the model with the most inliers in each round of iteration is the optimal model.

[0148] Re-execute least squares fitting using the set of interior points corresponding to the optimal model, and output the final optimized model parameters.

[0149] In this embodiment of the invention, parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency, further includes:

[0150] A sliding time window mechanism is adopted to perform periodic rolling updates, so that the model parameters are dynamically adjusted as load characteristics and operating conditions change;

[0151] After each window slide, the entire process of data culling, feature extraction, model building, and parameter identification is re-executed.

[0152] It should be noted that time-series differential processing is used to perform point-by-point differential calculations on power identification data in chronological order in order to analyze the changing trends between adjacent time points.

[0153] For example, assuming the power identification data is [P1, P2, P3, P4], the timing differential processing will generate [(P2-P1), (P3-P2), (P4-P3)], thereby extracting the power response characteristics within the voltage variation range.

[0154] It should be noted that the load power response characteristics are used to describe the dynamic response of load power to voltage fluctuations within a specific voltage variation range.

[0155] For example, if the load power increases from 5kW to 5.5kW when the voltage drops from 220V to 200V, the response characteristic can be expressed as "power increases by 10% when the voltage drops by 10%".

[0156] It should be noted that the sequence of parameters to be identified is a set of parameters that need to be further optimized and fitted based on voltage change and power response characteristics, used to characterize the system's regulation capability.

[0157] For example, if the system regulation capability is related to the rate of change of voltage and the power response rate, the sequence of parameters to be identified may include parameters such as [rate of change of voltage, rate of change of power, regulation sensitivity].

[0158] It should be noted that periodic rolling updates are used to repeatedly execute the entire process at certain time intervals to maintain the real-time performance and accuracy of the model parameters.

[0159] For example, the data culling, feature extraction, model building, and parameter identification processes are rerun every hour to reflect the latest load characteristics and operating conditions.

[0160] Specifically, in the data preprocessing stage, based on N sets of voltage-power time-series observation data (Vi, Pi), N-1 sets of CVR coefficient estimates are obtained by calculating the difference between adjacent data points. Simultaneously, the average value of adjacent voltage data is taken as the characteristic voltage value of the corresponding CVR coefficient, thus constructing a data sequence for subsequent modeling. Specifically, for the i-th data point, its characteristic voltage V... avg,i =(V i +V i+1 The corresponding CVR coefficients are obtained through differential calculation, thus forming a CVR coefficient sequence that matches the voltage level, providing regular input data for subsequent parameter identification.

[0161] Furthermore, during the data acquisition process, some outliers will inevitably be obtained due to factors such as equipment measurement errors. Therefore, the Random Sample Consensus Algorithm (RANSCA) is first used to preprocess the data to remove outliers, and then the least squares method is used for identification.

[0162] Furthermore, the RANSAC (Random Sample Consensus) algorithm is a robust regression method whose core idea is to identify the optimal set of interior points from data containing outliers through iterative random sampling and model validation. First, given a dataset D = {(x...} i ,y i Here, the parameters x and y correspond to the voltage U and CVR coefficient, respectively, and may contain some outliers. The goal of RANSAC is to estimate the optimal linear model y = ax + b. At the start of the algorithm, three key parameters need to be set: minimum number of sampling points k = 2, interior point decision threshold δ (usually 1.5-3 times the standard deviation of the data), and maximum number of iterations T (determined by a probability formula).

[0163] In each iteration, the algorithm performs the following steps:

[0164] (1) Random sampling: k points are uniformly drawn from D to form a sample set St

[0165] (2) Model estimation: Calculate the temporary straight line parameters using two points (x1, y1) and (x2, y2) in St:

[0166] a t = (y2-y1) / (x2-x1)

[0167] b t =y1-a t ×x1

[0168] (3) Interior point identification: Calculate the distance d from all points to the line. i , satisfying d i The points <δ form the interior point set It

[0169]

[0170] (4) Model evaluation: Record the current number of interior points |I t |

[0171] Furthermore, iteration stops when a sufficiently large set of interior points is found or the maximum number of iterations is reached. The model with the largest set of interior points, Imax, is selected, and the least squares estimate is recalculated using all interior points. The least squares method obtains the optimal matching function by minimizing the sum of squared residuals, and its basic form is as follows:

[0172]

[0173] In the formula, N is the length of the data sequence, J is the objective function, (x i y i ) represents a pair of observations, L i (x) is called the residual function, w i These are parameters to be determined. In this invention, the observable quantity is (U... i CVR fi The residual function is CVR. fi -f(U i From the load model, we know that f is a linear function. Assuming f(X) = θ1 + θ2X, then the objective function is:

[0174]

[0175] Furthermore, the best fit for parameter θ is achieved when the objective function J is minimized. Therefore, the problem is transformed into finding the minimum value of the objective function. By taking the partial derivatives of the objective function with respect to parameters θ1 and θ2 and setting them equal to zero, the following calculations can be obtained:

[0176]

[0177] This establishes a linear relationship between the feeder load CVR coefficient and the voltage level. The data window length used for identification is set to 1 hour, and the relationship is updated every 15 minutes.

[0178] In summary, this invention proposes a method for modeling and identifying feeder regulation characteristics that considers voltage sensitivity. Through multi-step data processing and advanced parameter identification methods, it fully considers the complex dynamic characteristics of feeder loads. In data processing, power data representing environmental factors are first removed, and voltage and power data strongly correlated with voltage fluctuations caused by load changes are accurately extracted, providing a high-quality data foundation for subsequent modeling. In modeling, a voltage reduction energy-saving coefficient model is established, which accurately characterizes the regulation potential of the feeder load. In the parameter identification stage, a robust least squares identification method based on random sampling consistency is used to select the optimal model through multiple iterations. A sliding time window mechanism is employed to dynamically adjust the model parameters, enabling the model to adapt to changes in load characteristics and operating conditions.

[0179] This invention effectively solves the problem that traditional modeling methods based on static assumptions cannot accurately characterize the actual regulation behavior of feeder loads. Through this invention, the regulation characteristics of feeder loads can be more accurately grasped, enabling precise control of the active power of feeder loads. This, in turn, effectively improves the power system's source-load synergy and interaction capabilities, enhances grid regulation and control capabilities and the capacity for renewable energy absorption, and provides strong support for the safe and stable operation of new power systems.

[0180] Example 2, refer to Figures 2-3 In a preferred embodiment, an improved IEEE 33-node distribution network model is built based on the MATLAB simulation platform. An OLTC is installed at the system headend, with each adjustment level being 0.0125 pu, for a total of 11 levels. The total system load is 3715 kW + j2300 kvar. It includes three distributed photovoltaic (PV) systems (PV1, PV2, PV3), connected to nodes 8, 23, and 32 respectively, each with a rated capacity of 500 kW. In addition, the system is equipped with three reactive power compensation devices (SVC1, SVC2, SVC3), connected to nodes 8, 17, and 25 respectively, with an adjustable range of [-250, 250] kvar. In this scenario, controlling the active power of the PV systems is not considered; both the PV systems and the SVC devices are used as parallel reactive power voltage regulators.

[0181] The CVR coefficients of each node are shown in Table 1. The original node admittance matrix and node injected power matrix are corrected.

[0182] Table 1 CVR coefficients for each node

[0183]

[0184] This example simulates grid voltage fluctuations and feeder load power changes, with system status data collected every 5 minutes. Table 1 records the voltage and power changes over one hour (12.66kV as the reference voltage). Parameter identification was performed in two time periods: 15-30 minutes and 30-45 minutes, with the data window set to 15 minutes. To eliminate the impact of temperature changes on load power, an SG filter (window size 5) was used to process the power data and extract voltage-sensitive parameters. Power data before and after filtering are shown in Table 2.

[0185] Table 2 shows the system operating status data before and after SG filtering.

[0186]

[0187]

[0188] Based on voltage and power data, the CVR coefficient within the two cycles is identified. This invention sets up three identification strategies, where strategies one and two are comparative examples, and strategy three is the improved identification strategy proposed in this invention, as detailed below:

[0189] Identification Strategy 1: Use the power data after SG filtering for identification, but ignore the linear relationship between CVR coefficient and node voltage.

[0190] Identification Strategy 2: Consider the linear relationship between CVR coefficient and node voltage, but ignore the coupling relationship between temperature and feeder load power, and use the data before filtering for identification.

[0191] Identification Strategy 3: Simultaneously considering the time-varying characteristics and temperature effect of the CVR coefficient, the weighted least squares method is used to identify the filtered data.

[0192] The identification results of the three strategies are as follows Figure 2 (a) and Figure 2 As shown in (b):

[0193] The blue, orange, and yellow curves in the graph correspond to the results of strategies one, two, and three, respectively. In strategy one, the CVR coefficient is a constant, calculated from the last two sets of data in the identification window. Strategy two uses the data before filtering, and the CVR coefficient is negatively correlated with the slack line voltage, contrary to theoretical expectations. The results of strategy three are more reasonable, with the CVR coefficient showing a positive correlation with the voltage level, consistent with the actual model of feeder load.

[0194] To compare the advantages and disadvantages of the three identification strategies, a comparative analysis is conducted based on MATPOWER. Perturbations were applied to the root node voltage at the 18th and 42nd minutes, respectively, lowering and raising it by one level. The CVR coefficients calculated based on the identification curves are shown in Table 3. The actual feeder power change obtained through MATPOWER power flow analysis was used as a reference value and compared with the power change calculated by the CVR coefficient model to evaluate the accuracy of the three strategies. The results are as follows: Figure 3 As shown.

[0195] Table 3. CVR coefficient identification results for each strategy at two time points.

[0196]

[0197] Figure 3 The results show that the calculation results based on the CVR coefficient of Strategy 3 are highly consistent with the actual system active power changes at both disturbance times. The accuracy of the other two strategies is relatively low. Strategy 3 can more accurately quantify the impact of voltage disturbance on feeder load power, verifying the effectiveness of the CVR coefficient model and the accuracy of the parameter identification strategy.

[0198] Example 3, referring to Figure 4 This embodiment also provides a feeder regulation characteristic modeling and parameter identification system that takes voltage sensitivity into account, including:

[0199] The data acquisition and processing module is used to acquire feeder load-related data of the target distribution network and to perform a data rejection operation on the feeder load-related data.

[0200] The removal operation is used to remove power data from the environmental factor trend item in feeder load-related data;

[0201] The extraction module is used to extract voltage and power data that are strongly correlated with the voltage fluctuation caused by load changes from the feeder load-related data after the removal operation, and use it as power identification data.

[0202] The model building module is used to obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network and to build a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads.

[0203] The solution module is used to identify parameters based on the step-down energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency.

[0204] The above-mentioned unit modules can be embedded in the processor of the electronic device in hardware form or independent of it, or they can be stored in the memory of the electronic device in software form, so that the processor can call and execute the corresponding operations of the above modules.

[0205] This embodiment also provides an electronic device, which can be a terminal, and its internal structure diagram can be as follows: Figure 4 As shown, the electronic device includes a processor, memory, communication interface, display screen, and input device connected via a system bus. The processor provides computing and control capabilities. The memory includes a non-volatile storage medium and internal memory. The non-volatile storage medium stores the operating system and computer programs. The internal memory provides an environment for the operation of the operating system and computer programs in the non-volatile storage medium. The communication interface is used for wired or wireless communication with external terminals; wireless communication can be achieved through Wi-Fi, carrier networks, NFC (Near Field Communication), or other technologies. When the computer program is executed by the processor, it implements a method for modeling and identifying parameters of feeder regulation characteristics that take voltage sensitivity into account. The display screen can be an LCD screen or an e-ink screen. The input device can be a touch layer covering the display screen, buttons, a trackball, or a touchpad mounted on the device's casing, or an external keyboard, touchpad, or mouse.

[0206] This embodiment also provides a computer-readable storage medium on which a computer program is stored, and when the computer program is executed by a processor, it performs the following steps:

[0207] Obtain feeder load-related data of the target distribution network and perform a data rejection operation on the feeder load-related data;

[0208] The removal operation is used to remove power data from the environmental factor trend item in feeder load-related data;

[0209] Extract voltage and power data that are strongly correlated with the voltage fluctuation caused by load changes from the feeder load-related data after the removal operation, and use them as power identification data.

[0210] Obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and establish a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads;

[0211] Based on the step-down energy-saving coefficient model and power identification data, parameter identification is performed using a robust least squares identification method based on random sampling consistency.

[0212] It should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical solutions of the present invention without departing from the spirit and scope of the technical solutions of the present invention, and all such modifications or substitutions should be covered within the scope of the claims of the present invention.

[0213] Although preferred embodiments of the invention have been described, those skilled in the art, upon learning the basic inventive concept, can make other changes and modifications to these embodiments. Therefore, the appended claims are intended to be interpreted as including both the preferred embodiments and all changes and modifications falling within the scope of the invention.

[0214] Obviously, those skilled in the art can make various modifications and variations to this invention without departing from its spirit and scope. Therefore, if these modifications and variations fall within the scope of the claims of this invention and their equivalents, this invention also intends to include these modifications and variations.

Claims

1. A method for modeling and identifying parameters of feeder regulation characteristics considering voltage sensitivity, characterized in that, include: Obtain feeder load-related data of the target distribution network and perform a rejection operation on the feeder load-related data; The removal operation is used to remove power data from the environmental factor trend item in the feeder load related data; Extract voltage and power data that are strongly correlated with the voltage fluctuation process caused by load changes from the feeder load-related data after the removal operation, and use them as power identification data. Obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and establish a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads; Based on the aforementioned voltage reduction energy-saving coefficient model and power identification data, parameter identification is performed using a robust least squares identification method based on random sampling consistency.

2. The method for modeling and identifying parameters of feeder regulation characteristics considering voltage sensitivity as described in claim 1, characterized in that, The rejection operation includes: Filtering methods are used to process feeder load-related data of the target distribution network; The voltage-sensitive power component is extracted by separating the high-frequency power component of voltage fluctuation from the low-frequency trend term of environmental factors through sliding window polynomial fitting. Subtract the fitted data of the non-voltage trend term from the original power data to obtain the voltage-sensitive power data sequence, and use the voltage-sensitive power data sequence as the voltage power data.

3. The method for modeling and identifying feeder regulation characteristics considering voltage sensitivity as described in claim 2, characterized in that, The voltage and power data strongly correlated with the voltage fluctuation process caused by load changes in the feeder load-related data after the extraction and removal operation are included as power identification data: A correlation acquisition model is established, and the correlation of the voltage and power data is calculated based on the correlation acquisition model. A pre-defined correlation judgment strategy is used to filter voltage and power data that are strongly correlated with the voltage fluctuations caused by load changes.

4. The method for modeling and identifying feeder regulation characteristics considering voltage sensitivity as described in claim 3, characterized in that, The process of obtaining the voltage-power coupling characteristics of the feeder loads of the target distribution network and establishing a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads includes: The voltage reduction energy saving coefficient model is used to characterize the regulation potential of the feeder load. The larger the value, the more significant the power response of the feeder load when the voltage changes, and the stronger the regulation capability. The voltage reduction energy saving coefficient model is obtained through the change in active power of the feeder load, the initial active power of the feeder load, the change in feeder voltage, and the initial voltage of the feeder.

5. The method for modeling and identifying feeder regulation characteristics considering voltage sensitivity as described in claim 4, characterized in that, The parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, includes: The power identification data is subjected to time-series differential processing to extract the load power response features corresponding to the voltage change range; Based on the voltage variation range and power response characteristics, a sequence of parameters to be identified is constructed to characterize the regulation capability; The random sampling consensus method is used to initially screen the parameter sequence to be identified, removing data points that are affected by abnormal disturbances or noise, and retaining the set of interior points that conform to the model trend; Perform a least-squares fitting operation on the set of interior points to obtain the estimated values ​​of stable parameters reflecting load regulation characteristics in the voltage reduction energy-saving coefficient model.

6. The method for modeling and identifying feeder regulation characteristics considering voltage sensitivity as described in claim 5, characterized in that, The parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, also includes: In each parameter identification process, a minimum number of data samples and a maximum number of iterations are set; In each iteration, a number of data points are randomly selected to construct a temporary parametric model, and the degree of matching between the remaining data points and the model is evaluated. Data points with a matching degree higher than a preset threshold are marked as inliers, and the model with the most inliers in each round of iteration is the optimal model. Re-execute least squares fitting using the set of interior points corresponding to the optimal model, and output the final optimized model parameters.

7. The method for modeling and identifying feeder regulation characteristics considering voltage sensitivity as described in claim 6, characterized in that, The parameter identification based on the voltage reduction energy-saving coefficient model and power identification data, combined with the robust least squares identification method based on random sampling consistency, also includes: A sliding time window mechanism is adopted to perform periodic rolling updates, so that the model parameters are dynamically adjusted as load characteristics and operating conditions change; After each window slide, the entire process of data culling, feature extraction, model building, and parameter identification is re-executed.

8. A feeder regulation characteristic modeling and parameter identification system considering voltage sensitivity, using the method described in any one of claims 1 to 7, characterized in that, include: The data acquisition and processing module is used to acquire feeder load-related data of the target distribution network and to perform a rejection operation on the feeder load-related data. The removal operation is used to remove power data from the environmental factor trend item in the feeder load related data; The extraction module is used to extract voltage and power data that are strongly correlated with the voltage fluctuation caused by load changes from the feeder load-related data after the removal operation, and use it as power identification data. The model building module is used to obtain the voltage-power coupling characteristics of the feeder loads of the target distribution network, and to build a step-down energy-saving coefficient model based on the voltage-power coupling characteristics of the feeder loads. The solution module is used to identify parameters based on the voltage reduction energy-saving coefficient model and power identification data, combined with a robust least squares identification method based on random sampling consistency.

9. An electronic device comprising a memory and a processor, wherein the memory stores a computer program, characterized in that, When the processor executes the computer program, it implements the steps of the method for modeling and identifying the feeder regulation characteristics and parameters taking into account voltage sensitivity, as described in any one of claims 1 to 7.

10. A computer-readable storage medium having a computer program stored thereon, characterized in that, When the computer program is executed by the processor, it implements the steps of the method for modeling and identifying the feeder regulation characteristics and parameters taking into account voltage sensitivity, as described in any one of claims 1 to 7.

Images