State sensing method for accessing distributed new energy to active power distribution network

Abstract

The invention discloses a distributed new energy access active power distribution network state sensing method, and relates to the technical field of power grid state sensing, and the method comprises the steps: S1, data collection, S2, data preprocessing, S3, multi-time scale sliding window setting, S4, data feature extraction and preliminary analysis, S5, data fusion, and S6, state sensing. The sliding windows of different sizes are set to collect and analyze the data, rapid updating of the data is guaranteed in the short-time scale window to capture dynamic changes, the medium-time scale sliding window reflects running trend changes of the power distribution network in a time period, the long-time scale window mines trend characteristics of the data, and the real-time performance of the power distribution network is improved. And data of different windows are fused, so that the operation characteristics of the power distribution network in different time dimensions can be known comprehensively, and information loss caused by single time scale analysis is avoided.

Description

Technical Field

[0001] The present invention belongs to the technical field of power grid state perception, and particularly relates to a state perception method for distributed new energy access to an active distribution network. Background Art

[0002] With the development of renewable energy technologies, a large number of distributed energy sources (such as solar photovoltaics, small-scale wind power, etc.) are connected to the distribution network. These distributed energy sources are characterized by intermittency and volatility, and their output power is affected by natural conditions (such as light intensity, wind speed, etc.), and can change significantly within a short period of time.

[0003] In the existing state perception of active distribution networks, it is difficult to balance the real-time performance and accuracy of data, and the data characteristics at different time scales are not fully utilized. At only a short time scale, the power of distributed new energy may be affected by problems such as passing clouds, resulting in instantaneous fluctuations in photovoltaic power. If only based on this instantaneous data, it may misjudge that the distribution network has a power imbalance problem. And if too much attention is paid to data with high accuracy but poor real-time performance, when a fault such as a short circuit occurs, if the current and voltage data within a short time scale cannot be obtained and analyzed in time, it may delay the fault judgment and handling, leading to an expansion of the fault influence range. In view of the above problems, the following solutions are proposed. Summary of the Invention

[0004] The purpose of the present invention is to provide a state perception method for distributed new energy access to an active distribution network. By collecting and analyzing data through sliding windows of different sizes and then fusing the data of different windows, the state of the distribution network can be comprehensively and accurately perceived, solving the problems in the prior art that it is difficult to balance the real-time performance and accuracy of data, and the data characteristics at different time scales are not fully utilized.

[0005] To solve the above technical problems, the present invention is realized through the following technical solutions:

[0006] The present invention is a state perception method for distributed new energy access to an active distribution network, and the perception method includes the following steps:

[0007] Step S1, data collection: Install data collection devices at the output end of the distributed new energy generation unit, the key bus nodes of the distribution network, and the load access points to obtain electrical parameter information and environmental parameter data;

[0008] Step S2, data preprocessing: Preprocess the collected data, including data cleaning and data synchronization;

[0009] Step S3, Multi-time-scale Sliding Window Setting: Set the sliding windows for short-time scale, medium-time scale, and long-time scale respectively, which are used to collect and store data and capture the operation state information of the distribution network with different time characteristics;

[0010] Step S4, Data Feature Extraction and Preliminary Analysis: Extract features from the data under different time-scale windows to evaluate the operation state, trend, and regularity of the distribution network;

[0011] Step S5, Data Fusion: Establish a fusion model to fuse the data features of different time scales;

[0012] Step S6, State Perception: Based on the fused feature vector, perceive the state of the distribution network and provide results for decision support.

[0013] Preferably, in the step S3, Multi-time-scale Sliding Window Setting, set the short-time scale sliding window to capture the rapid change information of the operation state of the distribution network; set the medium-time scale sliding window to reflect the change of the operation trend of the distribution network within a time period; set the long-time scale sliding window to analyze the long-term operation characteristics and regularity of the distribution network.

[0014] Preferably, in the step S4, Data Feature Extraction and Preliminary Analysis, the data feature extraction method for the short-time scale is as follows:

[0015] Calculate the volatility of power:

[0016]

[0017] where σ P is the power volatility, P i is the power value at the i-th sampling point, is the average power, and n is the number of sampling points in a short time;

[0018] Monitor the distortion rates of voltage and current:

[0019]

[0020] where H V is the total harmonic distortion rate of voltage, H I is the total harmonic distortion rate of current, V h is the amplitude of the h-th harmonic of voltage, I h is the amplitude of the h-th harmonic of current, V 1 and I 1 are the amplitudes of fundamental voltage and current respectively;

[0021] Judge whether there are problems according to the volatility and distortion rate, count the number of problem events, and evaluate the short-term operation risk of the distribution network.

[0022] Preferably, in the step S4, data feature extraction and preliminary analysis, the data feature extraction method for medium time scale is as follows:

[0023] Analyze the slope change of distributed new energy power:

[0024]

[0025] In the formula, k is the power slope, P start is the power value at the start time of the medium time window, P end is the power value at the end time of the medium time window, t start is the start time, t end is the end time;

[0026] k is used to predict its short-term power generation trend and provide forward-looking information for power grid scheduling. A slope greater than 0 indicates an upward power trend, and a slope less than 0 indicates a downward trend;

[0027] Calculate the mean and median of voltage and current:

[0028]

[0029] In the formula, is the average value of voltage, is the average value of current, V j is the voltage value at the j-th sampling point within the medium time window, I j is the current value at the j-th sampling point within the medium time window, and m is the number of sampling points;

[0030] The median V med and I med are obtained by sorting the data and taking the middle value, and are used to evaluate the average operation level of the distribution network during the period. The mean and median can reflect the central tendency of the data;

[0031] Perform clustering analysis on the load data:

[0032] Randomly select K initial clustering centers μ k , and calculate the distance from each load data point x i to each clustering center:

[0033]

[0034] In the formula, p is the data dimension, x il is the l-th eigenvalue of the i-th sample, μ kl is the l-th eigenvalue of the k-th clustering center;

[0035] Assign the data points to the nearest cluster and update the clustering centers:

[0036]

[0037] Wherein, n k is the number of data points in cluster C k ;

[0038] Repeat the process of updating the cluster center until the cluster center no longer changes, and identify different types of load characteristics and their changing rules;

[0039] Preferably, in step S4, data feature extraction and preliminary analysis, the long-time scale data feature extraction method is as follows:

[0040] Fit the long-term relationship model between distributed new energy power generation and environmental factors:

[0041] Adopt the model:

[0042] D = a + bS + cW + ε,

[0043] Wherein, S is the light intensity, W is the wind speed, a, b, c are regression coefficients, and ε is the error term;

[0044] Solve the regression coefficients through the formula:

[0045]

[0046] Wherein, X is a matrix containing the constant term, light intensity, and wind speed data, and y is the power data vector;

[0047] Analyze the long-term statistical distribution characteristics of voltage and current: Calculate the mean values μ V , μ I and standard deviations σ V , σ I , and calculate the data proportion within the confidence interval to understand the long-term operation reliability of the distribution network;

[0048] Explore the seasonal and periodic laws of the load:

[0049] Adopt the formula Analyze the amplitudes and phases of different frequency components to master the daily cycle and weekly cycle characteristics of the load;

[0050] Wherein, L(t) is the function of the load changing with time t, a 0 is the constant term, T is the period, a n and b n are Fourier coefficients.

[0051] Preferably, the specific method of step S5, data fusion, is:

[0052] Build a data fusion model: Assign weights according to the importance of data on different time scales for the state perception of the distribution network. Let the weight of short-time scale data be ω 1 , the weight of medium-time scale data be ω 2 , and the weight of long-time scale data be ω 3 , and ω 1 +ω 2 +ω 3 = 1;

[0053] Fuse data features of different time scales:

[0054]

[0055] In the formula, F is the fused feature vector, F 1 , F 2 , F 3 are the feature vectors of short, medium, and long time scales respectively.

[0056] The present invention has the following beneficial effects:

[0057] 1. The present invention collects and analyzes data by setting sliding windows of different sizes. The short-time scale window ensures the rapid update of data to capture dynamic changes, the medium-time scale sliding window reflects the change of the operation trend of the distribution network within a time period, the long-time scale window mines the trend features of the data, and then fuses the data of different windows. Furthermore, it enables the timely monitoring of the sudden power change caused by weather mutations of distributed new energy at the short-time scale, the discovery of the voltage change trend brought by load transfer at the medium-time scale, and the grasp of seasonal features such as the peak period of air-conditioning load in summer at the long-time scale, so as to comprehensively understand the operation characteristics of the distribution network in different time dimensions and avoid information loss caused by single-time scale analysis.

[0058] 2. The present invention can effectively fuse multi-source data collected within different time scale windows, including electrical quantity data, meteorological data, equipment status data, etc., comprehensively consider the influence of various factors on the state of the distribution network, avoid the limitations of a single data source, improve the accuracy of perception. At the same time, according to the characteristics and change trends of data at different time scales, it can adaptively adjust the fusion weights and the parameters of the perception model, realize the dynamic perception and optimization of the state of the distribution network, and better adapt to the intermittency and volatility of distributed new energy.

[0059] Of course, any product implementing the present invention does not necessarily need to achieve all the above-mentioned advantages simultaneously. Brief Description of the Drawings

[0060] To more clearly illustrate the technical solutions of the embodiments of the present invention, the following will briefly introduce the drawings required for the description of the embodiments. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, without creative efforts, other drawings can be obtained based on these drawings.

[0061] Figure 1 It is a schematic flowchart of the state perception method for distributed new energy access to an active distribution network of the present invention. Specific embodiments

[0062] The following will clearly and completely describe the technical solutions in the embodiments of the present invention in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only some embodiments of the present invention, rather than all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those of ordinary skill in the art without creative efforts belong to the scope of protection of the present invention.

[0063] Please refer to Figure 1 As shown, the present invention is a state perception method for distributed new energy access to an active distribution network. The perception method includes the following steps:

[0064] Step S1, data acquisition: Install data acquisition devices at the output end of the distributed new energy generation unit, the key bus nodes of the distribution network, and the load access points to obtain electrical parameter information and environmental parameter data;

[0065] Step S2, data preprocessing: Preprocess the collected data, including data cleaning and data synchronization;

[0066] Step S3, multi-time scale sliding window setting: Set sliding windows for short time scale, medium time scale, and long time scale respectively to collect and store data and capture the operation state information of the distribution network with different time characteristics;

[0067] Step S4, data feature extraction and preliminary analysis: Extract features from the data under different time scale windows to evaluate the operation state, trend, and law of the distribution network;

[0068] Step S5, data fusion: Establish a fusion model to fuse the data features of different time scales;

[0069] Step S6, state perception: Based on the fused feature vector, perceive the state of the distribution network and provide results for decision support.

[0070] In step S3, in the multi-time scale sliding window setting, a short-time scale sliding window is set to capture the rapid change information of the distribution network operation status; a medium-time scale sliding window is set to reflect the operation trend change of the distribution network within a time period; a long-time scale sliding window is set to analyze the long-term operation characteristics and regularities of the distribution network.

[0071] In step S4, in the data feature extraction and preliminary analysis, the data feature extraction method for the short-time scale is as follows:

[0072] Calculate the volatility of power:

[0073]

[0074] where σ P is the power volatility, P i is the power value at the i-th sampling point, is the average power, and n is the number of sampling points in a short time;

[0075] Monitor the distortion rates of voltage and current:

[0076]

[0077] where H V is the total harmonic distortion rate of voltage, H I is the total harmonic distortion rate of current, V h is the amplitude of the h-th harmonic of voltage, I h is the amplitude of the h-th harmonic of current, V 1 and I 1 are the amplitudes of fundamental voltage and current respectively;

[0078] Judge whether there are problems according to the volatility and distortion rate, count the number of problem events, and evaluate the short-term operation risk of the distribution network.

[0079] In step S4, in the data feature extraction and preliminary analysis, the data feature extraction method for the medium-time scale is as follows:

[0080] Analyze the slope change of distributed new energy power:

[0081]

[0082] where k is the power slope, P start is the power value at the start time of the medium-time window, P end is the power value at the end time of the medium-time window, t start is the start time, t end is the end time;

[0083] k is used to predict its short-term power generation trend and provide forward-looking information for grid dispatching. A slope greater than 0 indicates an upward power trend, and a slope less than 0 indicates a downward trend;

[0084] Calculate the mean and median of voltage and current:

[0085]

[0086] In the formula, is the average value of voltage, is the average value of current, V j is the voltage value at the j-th sampling point within the medium time window, I j is the current value at the j-th sampling point within the medium time window, and m is the number of sampling points;

[0087] The median V med and I med are obtained by sorting the data and taking the middle value, and are used to evaluate the average operating level of the distribution network during the period. The mean and median can reflect the central tendency of the data;

[0088] Perform clustering analysis on the load data:

[0089] Randomly select K initial clustering centers μ k , and calculate the distance from each load data point x i to each clustering center:

[0090]

[0091] In the formula, p is the data dimension, x il is the l-th eigenvalue of the i-th sample, and μ kl is the l-th eigenvalue of the k-th clustering center;

[0092] Assign the data points to the nearest cluster and update the clustering centers:

[0093]

[0094] In the formula, n k is the number of data points in the cluster C k ;

[0095] Repeat the process of updating the clustering centers until the clustering centers no longer change, and identify different types of load characteristics and their variation laws;

[0096] In step S4, data feature extraction and preliminary analysis, the long-time scale data feature extraction method is as follows:

[0097] Fit the long-term relationship model between distributed new energy power generation and environmental factors:

[0098] Adopted model:

[0099] D = a + bS + cW + ε,

[0100] where S is the light intensity, W is the wind speed, a, b, and c are regression coefficients, and ε is the error term;

[0101] Solve for the regression coefficients through the formula:

[0102]

[0103] where, X is a matrix containing the constant term, light intensity, and wind speed data, and y is the power data vector;

[0104] Analyze the long-term statistical distribution characteristics of voltage and current: Calculate the means μ V , μ I and standard deviations σ V , σ I , and calculate the data proportion within the confidence interval to understand the long-term operation reliability of the distribution network;

[0105] Explore the seasonal and periodic patterns of the load:

[0106] Adopt the formula to analyze the amplitudes and phases of different frequency components and master the daily and weekly cycle characteristics of the load;

[0107] where L(t) is the function of the load varying with time t, a 0 is the constant term, T is the period, and a n and b n are Fourier coefficients.

[0108] Step S5. The specific method of data fusion is as follows:

[0109] Establish a data fusion model: Assign weights according to the importance of data on different time scales for the state perception of the distribution network. Let the weight of short-time scale data be ω 1 , the weight of medium-time scale data be ω 2 , and the weight of long-time scale data be ω 3 , and ω 1 + ω 2 + ω 3 = 1;

[0110] Fuse the data characteristics of different time scales:

[0111]

[0112] where F is the fused feature vector, F 1 , F 2, F 3 are the eigenvectors of short, medium, and long time scales respectively.

[0113] A specific application of this embodiment is:

[0114] Step S1, data acquisition: Install data acquisition devices such as smart meters and sensors at the output ends of distributed new energy power generation units (such as photovoltaic arrays, wind turbines), key bus nodes of the distribution network, and important load access points to ensure that electrical parameter information such as voltage, current, active power, reactive power, and frequency can be obtained, as well as environmental parameter data such as light intensity, wind speed, and temperature (for new energy power generation related);

[0115] Step S2, data preprocessing: Preprocess the collected data, including data cleaning, data synchronization, etc.;

[0116] Step S3, multi-time scale sliding window setting: Set sliding windows for short time scale, medium time scale, and long time scale respectively to collect and store data and capture the operating state information of different time characteristics of the distribution network. Specifically as follows:

[0117] Set a short time scale sliding window (such as a 1-minute window): This window is mainly used to capture the rapid change information of the distribution network operating state, such as the instantaneous power fluctuation of distributed new energy, the voltage sag caused by the sudden switching of loads, etc. Every 1 minute, the data collected during this time period is stored as a data segment, marked with a timestamp, and placed in the short time scale data buffer;

[0118] Set a medium time scale sliding window (such as a 10-minute window): This window can reflect the change trend of the distribution network operating in a certain time period, such as the average power change rate of distributed new energy within 10 minutes, the slow drift of voltage, etc. Every 10 minutes, generate the corresponding medium time scale data segment and store it in the corresponding buffer;

[0119] Set a long time scale sliding window (such as a 1-hour window): Used to analyze the long-term operating characteristics and regularity of the distribution network, such as the change trend of the total power generation of distributed new energy in different seasons and different time periods, the long-term impact of the daily cycle characteristics of the load on the voltage quality, etc. Generate a long time scale data segment every hour and store it;

[0120] Step S4, data feature extraction and preliminary analysis: Under different time scale windows, use corresponding formulas and methods to extract features from the data for evaluating the operating state, trend, and regularity of the distribution network. Specifically:

[0121] Data in the short time scale window (1-minute window):

[0122] Calculate the volatility of power:

[0123]

[0124] Among them, σ P is the power volatility, P i is the power value at the i-th sampling point, is the average power, and n is the number of sampling points within 1 minute; this index is used to measure the stability of the power output of distributed new energy. The greater the volatility, the more unstable the power output;

[0125] Monitor the distortion rates of voltage and current:

[0126]

[0127] In the formula, H V is the total harmonic distortion rate of voltage, H I is the total harmonic distortion rate of current, V h is the amplitude of the h-th harmonic of voltage, I h is the amplitude of the h-th harmonic of current, V 1 and I 1 are the amplitudes of fundamental voltage and current respectively;

[0128] Judge whether there are power quality problems. The higher the distortion rate, the worse the power quality;

[0129] Count the number of events occurring: such as the number of power overlimit times N P (count when P i > P max or P i < P min , P max and P min are the upper and lower power limits), the number of voltage sags N V (count when the voltage amplitude is lower than the set threshold N th and lasts for a certain time), etc., which are used to evaluate the short-term operation risk of the distribution network. The more the number of events occurring, the higher the risk;

[0130] Medium time-scale window data (10-minute window):

[0131] Analyze the slope change of distributed new energy power:

[0132]

[0133] In the formula, k is the power slope, P start is the power value at the start time of the medium time window, P end is the power value at the end time of the medium time window, t start is the start time, t endis the end time. The slope can predict its short-term power generation trend and provide forward-looking information for power grid dispatching. A slope greater than 0 indicates an upward power trend, and a slope less than 0 indicates a downward trend;

[0134] Calculate the mean and median of voltage and current:

[0135]

[0136] In the formula, is the average value of voltage, is the average value of current, V j is the voltage value at the j-th sampling point within the medium time window, I j is the current value at the j-th sampling point within the medium time window, and m is the number of sampling points; the median V med and I med can be obtained by sorting the data and taking the middle value (when m is even, take the average of the two middle numbers), which is used to evaluate the average operation level of the distribution network during this period. The mean and median can reflect the central tendency of the data;

[0137] First, randomly select K initial cluster centers μ k (k = 1, 2,..., K), and then calculate the distance from each load data point x i to each cluster center:

[0138]

[0139] In the formula, p is the data dimension, x il is the l-th eigenvalue of the i-th sample, and μ kl is the l-th eigenvalue of the k-th cluster center;

[0140] Assign the data points to the nearest cluster, and then update the cluster centers:

[0141]

[0142] In the formula, n k is the number of data points in cluster C k ;

[0143] Repeat the above process until the cluster centers no longer change, and identify different types of load characteristics and their changing rules to better carry out load management and power grid planning;

[0144] Long-time scale window data (1-hour window):

[0145] Fit the long-term relationship model between distributed new energy power generation and environmental factors (such as light, wind speed):

[0146] Adopt the formula:

[0147] D = a + bS + cW + ε,

[0148] where S is the light intensity, W is the wind speed, a, b, and c are regression coefficients, and ε is the error term;

[0149] By the formula the regression coefficients are solved to improve the accuracy of the model for predicting the power generation, and the power change is predicted according to the environmental factors;

[0150] where X is a matrix containing the constant term, light intensity, and wind speed data, and y is the power data vector;

[0151] Analyze the long-term statistical distribution characteristics of voltage and current: Calculate the means μ V 、μ I and standard deviations σ V 、σ I , and assume that they approximately follow a normal distribution and Understand the long-term operation reliability of the distribution network by calculating the data ratio within a certain confidence interval. The more concentrated the data is around the mean and the smaller the standard deviation, the higher the reliability. For example:

[0152] P(μV - z α / 2 σ V <V<μ V + z α / 2 σ V ), z α / 2 is the quantile of the standard normal distribution;

[0153] Explore the seasonal and periodic patterns of the load:

[0154] Use the formula to analyze the amplitudes and phases of different frequency components, master the daily cycle, weekly cycle, etc. characteristics of the load, and optimize the long-term planning and resource allocation of the distribution network. For example, dispatch the power generation resources in advance before the load peak;

[0155] where L(t) is the function of the load varying with time t, a 0 is the constant term, T is the period, such as the daily period T = 24 hours, a n and b n are Fourier coefficients, which can be obtained by integration;

[0156] Step S5, Data fusion: The data characteristics at different time scales are fused by establishing a fusion model. Specifically:

[0157] Establish a data fusion model: Using the weighted average method, weights are assigned according to the importance of data at different time scales for the state perception of the distribution network. Let the weight of short-time scale data be ω 1 , the weight of medium-time scale data be ω 2 , and the weight of long-time scale data be ω 3 , and ω 1 + ω 2 + ω 3 = 1. For example, for rapid fault diagnosis, ω 1 = 0.6, ω 2 = 0.3, ω 3 = 0.1. For long-term operation planning, ω 1 = 0.1, ω 2 = 0.3, ω 3 = 0.6;

[0158] Fuse the data features of different time scales:

[0159]

[0160] In the formula, F is the fused feature vector, F 1 , F 2 , F 3 are the feature vectors of short, medium, and long time scales respectively

[0161] Step S6, State perception: Based on the fused feature vector, through preset threshold judgment, pattern recognition (such as classification algorithms like support vector machines and neural networks), or a rule-based inference system, comprehensively perceive the operating state of the distribution network, including but not limited to power balance state, voltage stability, power quality level, operation risk degree, etc., and display the perception results on the human-machine interface of the distribution network monitoring center in real time to provide decision support for operators, and at the same time can also be used as the basis for subsequent grid optimization control and fault warning;

[0162] In the description of this specification, the descriptions referring to terms such as "one embodiment", "example", "specific example", etc. mean that the specific features, structures, materials, or characteristics described in connection with the embodiment or example are included in at least one embodiment or example of the present invention. In this specification, the schematic representations of the above terms do not necessarily refer to the same embodiment or example. Moreover, the specific features, structures, materials, or characteristics described can be combined in a suitable manner in any one or more embodiments or examples.

[0163] The preferred embodiments of the present invention disclosed above are only used to help illustrate the present invention. The preferred embodiments do not describe all the details in detail, nor do they limit the invention to the specific embodiments described. Obviously, many modifications and variations can be made according to the content of this specification. These embodiments are selected and specifically described in this specification to better explain the principles and practical applications of the present invention, so that those skilled in the art can well understand and utilize the present invention. The present invention is only limited by the claims and their full scope and equivalents.

Claims

1. A state perception method for accessing a distributed renewable energy source to an active distribution network, characterized in that: The sensing method comprises the following steps: Step S1, data acquisition: installing data acquisition equipment at the output end of the distributed renewable energy power generation unit, the key bus node of the distribution network, and the load access point to obtain electrical parameter information and environmental parameter data; Step S2, data preprocessing: preprocessing the collected data, including data cleaning and data synchronization; Step S3, multi-time scale sliding window setting: setting sliding windows of short time scale, medium time scale and long time scale respectively, for collecting and storing data, and capturing the operation status information of distribution network with different time characteristics; Step S4, data feature extraction and preliminary analysis: extract features from the data in different time scale windows to evaluate the operating status, trends and laws of the distribution network; Step S5, data fusion: fusing data features of different time scales by establishing a fusion model; Step S6, state perception: Based on the fused feature vector, the state of the distribution network is perceived and the results are provided for decision support.

2. The state perception method for accessing a distributed renewable energy source to an active power distribution network according to claim 1 is characterized in that: In the step S3, in the multi-time scale sliding window setting, a short time scale sliding window is set to capture the rapid change information of the distribution network operation status; a medium time scale sliding window is set to reflect the operation trend changes of the distribution network within a time period; and a long time scale sliding window is set to analyze the long-term operation characteristics and regularity of the distribution network.

3. The state perception method for accessing a distributed renewable energy source to an active power distribution network according to claim 1 is characterized in that: In the step S4, data feature extraction and preliminary analysis, the method for extracting data features on a short time scale is as follows: Calculate the fluctuation rate of power: In the formula, σ P is the power fluctuation rate, P i is the power value of the i-th sampling point, is the average power value, n is the number of sampling points in a short time; Monitor the distortion rate of voltage and current: In the formula, H V is the total harmonic distortion rate of voltage, H I is the total harmonic distortion rate of the current, V h is the hth harmonic amplitude of the voltage, I h is the hth harmonic amplitude of the current, V1 and I1 are the fundamental voltage and current amplitudes respectively; Determine whether there is a problem based on the volatility and distortion rate, count the number of problem events, and evaluate the short-term operation risk of the distribution network.

4. The state perception method for accessing a distributed renewable energy source to an active power distribution network according to claim 1, characterized in that: In the step S4, data feature extraction and preliminary analysis, the method for extracting data features at a medium time scale is as follows: Analyze the slope change of distributed new energy power: Where k is the power slope, P start is the power value at the start of the time window, P end is the power value at the end of the middle time window, t start is the starting time, t end For the end moment; k is used to predict its short-term power generation trend and provide forward-looking information for power grid dispatch. A slope greater than 0 indicates an upward trend in power, and a slope less than 0 indicates a downward trend. Calculate the mean and median of voltage and current: In the formula, is the average value of the voltage, is the average value of the current, V j is the voltage value of the jth sampling point in the time window, I j is the current value of the jth sampling point in the time window, and m is the number of sampling points; MedianV med and I med It is obtained by sorting the data and taking the middle value, which is used to evaluate the average operation level of the distribution network during the period. The mean and median can reflect the central trend of the data; Perform cluster analysis on load data: Randomly select K initial cluster centers μ k , calculate each load data point x i Distance to each cluster center: In the formula, p is the data dimension, x il is the lth eigenvalue of the ith sample, μ kl is the lth eigenvalue of the kth cluster center; Assign data points to the closest cluster and update the cluster center: Where n k is cluster C k The number of data points in ; Repeat the process of updating the cluster center until the cluster center no longer changes, and identify different types of load characteristics and their changing laws.

5. The state perception method for accessing a distributed renewable energy source to an active power distribution network according to claim 1, characterized in that: In the step S4, data feature extraction and preliminary analysis, the long-time scale data feature extraction method is as follows: Fitting the long-term relationship model between distributed renewable energy power generation and environmental factors: Adopting model: D=a+bS+cW+ε, In the formula, S is the light intensity, W is the wind speed, a, b, c are regression coefficients, and ε is the error term; Solve the regression coefficient using the formula: In the formula, X is a matrix containing constant terms, light intensity and wind speed data, and y is a power data vector; Analyze the long-term statistical distribution characteristics of voltage and current: Calculate the mean value μ of voltage and current V , μ I and standard deviation σ V , σ I , calculate the proportion of data within the confidence interval to understand the long-term operation reliability of the distribution network; Seasonal and cyclical patterns of excavation loads: Using formula Analyze the amplitude and phase of different frequency components; In the formula, L(t) is the function of load changing with time t, a0 is the constant term, T is the period, a n and b n are the Fourier coefficients.

6. The state perception method for connecting distributed renewable energy to an active power distribution network according to claim 1, characterized in that: The specific method of step S5, data fusion, is as follows: Establish a data fusion model: assign weights according to the importance of data at different time scales to the distribution network state perception, assuming that the weight of short-time scale data is ω1, the weight of medium-time scale data is ω2, the weight of long-time scale data is ω3, and ω1+ω2+ω3=1; Fusion of data features at different time scales: Where F is the fused feature vector, and F1, F2, and F3 are the feature vectors of short, medium, and long time scales, respectively.

Images