Time-sharing segmented high and low voltage evaluation method based on data distribution

Through the time-sharing and segmented voltage evaluation method based on data distribution, the problems of single-point sampling, single deviation, instantaneousness and randomness in voltage quality evaluation are solved, and dynamic and comprehensive evaluation of voltage quality is achieved, providing More accurate and comprehensive voltage evaluation results.

CN119087091APending Publication Date: 2024-12-06STATE GRID XINJIANG ELECTRIC POWER CO URUMQI ELECTRIC POWER SUPPLY CO
View PDF 0 Cites 0 Cited by

Patent Information

Application Number
CN202411251718.8
Authority / Receiving Office
CN · China
Patent Type
Applications(China)
Current Assignee / Owner
Filing Date
2024-09-09
Publication Date
2024-12-06

AI Technical Summary

Technical Problem

The existing voltage quality evaluation methods have problems such as limited single point sampling, single voltage deviation, difficult to reflect instantaneousness and randomness, lack of dynamic evaluation and insufficient data utilization, and lack of comprehensive evaluation indicators, resulting in inaccurate and comprehensive evaluation results.

Method used

Using the time-sharing and segmented voltage evaluation method based on data distribution, the time series-based adaptive data filtering algorithm is used to establish reasonable evaluation indicators and models, dynamically Evaluate voltage quality.

Benefits of technology

It achieves a more accurate, comprehensive and dynamic assessment of voltage quality, which can comprehensively consider the frequency, duration and impact range of voltage fluctuations, provide real-time decision support, and overcome the shortcomings of traditional methods.

✦ Generated by Eureka AI based on patent content.

Smart Images

  • Figure CN119087091A_ABST
    Figure CN119087091A_ABST
Patent Text Reader

Abstract

The invention discloses a time-sharing segmented high and low voltage evaluation method based on data distribution, and the method comprises the following steps: 1) setting evaluation indexes: carrying out the precise quantification of a voltage level, and building a reasonable evaluation index; 2) data preprocessing: adopting a self-adaptive data filtering algorithm based on a time sequence to filter abnormal values and / or noise; 3) frequency spectrum center estimation: analyzing a voltage data frequency spectrum center, evaluating the intensity of voltage fluctuation, and identifying dominant frequency components of the voltage fluctuation; 4) trend evaluation: analyzing the long-term trend of the voltage level through trend evaluation based on a time sequence; and 5) spatial analysis: performing spatial analysis of voltage data considering the feeder topology, and evaluating the feeder head end voltage level by analyzing the propagation characteristics of voltage abnormity in the feeder. According to the method, time and space are comprehensively considered, existing data are fully utilized to evaluate the high voltage and the low voltage of the transformer area in a multi-dimensional mode, and an effective auxiliary decision-making basis can be provided for treatment of the low voltage of the transformer area.
Need to check novelty before this filing date? Find Prior Art

Description

Technical Field

[0001] The invention relates to the technical field of power system voltage judgment, and in particular to a method for evaluating the voltage quality of a distribution network by using data distribution analysis. Background Art

[0002] Voltage quality is a key indicator of power system performance, which directly affects the stability of the power grid and the user's power experience. In the power system, voltage quality problems are mainly manifested as low voltage and high voltage phenomena, both of which may have adverse effects on the operation of the power grid and user equipment. Low voltage may cause the equipment to fail to start properly or operate inefficiently, while high voltage may cause insulation damage or premature aging of equipment. Therefore, accurate assessment of voltage quality is crucial to prevent and solve these problems.

[0003] Existing voltage quality evaluation methods are mainly based on two parameters: voltage deviation and voltage qualification rate. Voltage deviation refers to the difference between the actual voltage value and the rated value, while the voltage qualification rate is usually calculated based on the measurement data at a certain time of the day.

[0004] These methods have obvious shortcomings:

[0005] 1. Limitations of single-point sampling: The calculation of voltage qualification rate is usually based on a single measurement, which cannot fully reflect the actual situation of voltage fluctuations within a day, and may cause the evaluation results to not accurately represent the voltage quality.

[0006] 2. The singleness of voltage deviation: Although voltage deviation can reflect the degree of deviation between the voltage and the rated value, this indicator is single and cannot comprehensively consider the frequency, duration and impact range of voltage fluctuations.

[0007] 3. Instantaneity and randomness issues: Existing evaluation methods are difficult to accurately reflect the instantaneity and randomness of voltage fluctuations. Instantaneous voltage drops or surges may be ignored in a single measurement, while these events may have a very serious impact on grid stability and electrical equipment.

[0008] 4. Lack of dynamic evaluation: Existing evaluation methods often adopt static, offline evaluation methods and lack the ability to conduct real-time, dynamic monitoring and evaluation of voltage quality.

[0009] 5. Insufficient data utilization: With the development of smart grid technology, a large amount of voltage data can be obtained through smart meters and sensors. However, existing evaluation methods fail to make full use of this data to provide a deeper and more detailed voltage quality assessment.

[0010] 6. Lack of comprehensive evaluation indicators: Traditional voltage evaluation methods often consider low voltage and high voltage problems in isolation, and lack an evaluation system that comprehensively considers multiple voltage quality issues.

[0011] In view of the above problems, there is an urgent need for a new voltage evaluation method that can overcome the shortcomings of existing methods and provide a more accurate, comprehensive and dynamic voltage quality assessment. Summary of the invention

[0012] In order to solve the above problems, the present invention proposes a time-segmented voltage evaluation method based on data distribution, aiming to provide more effective decision support for voltage quality management of power system through a dynamic and comprehensive evaluation method.

[0013] As a technical solution to the above technical problems, the inventors disclose a time-sharing and segmented high and low voltage evaluation method based on data distribution, comprising the following steps:

[0014] S1: Accurately quantify the voltage level and establish reasonable evaluation indicators;

[0015] S2: Adopting adaptive data filtering algorithm based on time series to filter outliers or noise;

[0016] S3: Analyze the center of the voltage data spectrum to identify voltage spikes or abnormal peaks, evaluate the intensity of voltage fluctuations, identify the dominant frequency components of voltage fluctuations, and make a preliminary evaluation of the voltage;

[0017] S4: Analyze the long-term trend of voltage levels through time series-based trend assessment to help predict and identify potential voltage problems, such as persistently high or low voltage, thereby evaluating the voltage qualification rate;

[0018] S5: Perform spatial analysis of voltage data taking into account feeder topology, and evaluate the voltage level at the feeder head end by analyzing the propagation characteristics of voltage anomalies in the feeder.

[0019] As a further improvement of the above method, S1 includes: establishing reasonable evaluation indicators through accurate quantification of voltage levels, and strictly dividing the voltage into five interval levels according to the provisions of the national standard GB / T 12325-2008 "Power Quality Power Supply Voltage Deviation" that "the voltage deviation of three-phase power supply of 20kV and below is +7% of the nominal voltage; the voltage deviation of 220V single-phase power supply is +7% and -10% of the nominal voltage", and each level is equipped with strictly defined evaluation indicators:

[0020] Severe overvoltage: A voltage value exceeding 110% of the rated value is defined as severe overvoltage;

[0021] Overvoltage: Voltage values ​​within the range of 107%-110% of the rated value are defined as overvoltage;

[0022] Voltage qualified: When the voltage value is within 90%-107% of the rated value, it is defined as voltage qualified;

[0023] Low voltage: The voltage value is defined as low voltage within the range of 85%-90% of the rated value;

[0024] Severe low voltage: A voltage value lower than 85% of the rated voltage is defined as severe low voltage.

[0025] As a further improvement of the above method, S2 includes: establishing a baseline model: using historical voltage data to train the baseline model to capture the statistical characteristics of the voltage data, including but not limited to the mean, variance, etc. Anomaly detection: for newly collected voltage data points, use the baseline model to predict their expected values ​​and confidence intervals to identify abnormal voltage data. Adaptive confidence interval adjustment: according to the volatility of voltage data, dynamically adjust the confidence interval parameter value to adapt to the fluctuation range of voltages of different levels. Identification and processing of abnormal points: if the voltage data point exceeds the confidence interval, it is marked as abnormal and the abnormal value is filtered.

[0026] As a further improvement of the above method, the abnormality detection is specifically: the voltage prediction value is: in, is the predicted voltage value at time t, V t-1 ,V t-2 ,V t-3 ,…,V t-n is the historical voltage value; the confidence interval is expressed as: Where σ is the standard deviation of the predicted value and k is the confidence level coefficient. The real-time voltage data point is compared with the predicted expected value. If the data point V t Beyond the confidence interval CT t , marking it as a potential outlier.

[0027] As a further improvement of the above method, the adaptive confidence interval adjustment is specifically: dynamically adjusting the k value according to the volatility of the voltage data to adapt to the fluctuation range of voltages of different levels.

[0028] As a further improvement of the above method, S3 includes: calculating energy distribution: quantifying the intensity of voltage fluctuations in different time periods by calculating energy distribution. Estimating frequency center: performing Fourier transform on the voltage signal to obtain the spectrum representation of the voltage signal, and identifying the peaks in the spectrum after obtaining the spectrum representation of the voltage signal. By identifying these peaks, the dominant frequency of the signal can be determined. Then, the frequency center is estimated using a weighted average method to determine in which level interval the voltage is.

[0029] As a further improvement of the above method, the energy distribution is calculated by applying energy calculation to the sliding time window of the voltage signal. The calculation formula of energy E(t) at time t is:

[0030]

[0031] Where T is the time window length and ν(t) is the voltage value at time τ.

[0032] As a further improvement of the above method, the frequency center estimation is specifically: performing Fourier transform on the voltage signal to obtain the frequency spectrum representation of the voltage signal. The transformation formula is:

[0033] V(f)=FFT(v(t))

[0034] Where v(t) is the voltage signal in the time domain, V(f) is the corresponding frequency domain representation, and f is the frequency. After obtaining the spectrum representation of the voltage signal, the peaks in the spectrum are identified. By identifying these peaks, the dominant frequency of the signal can be determined. Then the frequency center is estimated using the weighted average method, and the formula is:

[0035]

[0036] Among them, f i is the frequency point, V(f i ) is the corresponding spectrum amplitude. If the voltage frequency center can be obtained, the voltage level is determined to determine which level interval the voltage is in. If the frequency center does not exist, the voltage is evaluated based on the time series trend.

[0037] As a further improvement of the above method, S4 includes: establishing an ARIMA model:

[0038]

[0039] Use ACF and PACF plots to identify the autoregressive (AR) parameter (p), difference (I) order (d), and moving average (MA) parameter (q) of the appropriate ARIMA model, and use the Bayesian method to estimate the autoregressive parameter (AR parameter) φ of the ARIMA model i , Moving average parameter (MA parameter) θ j , the variance of the error term σ 2 The estimated parameters are applied to the ARIMA model and fitted to the voltage data:

[0040]

[0041] where c is a constant term. Evaluate the model parameters, especially the autoregressive term φ i and the moving average term θ j , to determine whether there is a trend component. If the differential order d is greater than 0, it may indicate that there is a trend. After obtaining the voltage trend, the voltage evaluation is performed based on the voltage trend.

[0042] As a further improvement of the above method, S5 includes: evaluating the voltage level at the head end by analyzing the propagation characteristics of the voltage anomaly in the feeder, and the evaluation formula is:

[0043]

[0044] Among them, V start is the voltage prediction value at the feeder head end, V en d is the known voltage at the end of the feeder, n is the number of segments on the feeder, R i is the resistance of the i-th segment, X i is the reactance of the i-th segment, α is a proportional coefficient determined according to the line length and system characteristics, used to adjust the effect of resistance and reactance on voltage drop. L is the length of the feeder, and Z

[0045] is the total impedance of the feeder.

[0046] The beneficial effects of the present invention are:

[0047] The voltage level of the substation is accurately quantified, and the voltage level is strictly divided into five intervals, providing a basis for the evaluation of high and low voltages. By filtering outliers and / or noise through an adaptive data filtering algorithm based on time series, the present invention can be adapted to the evaluation of voltages of different levels. The preliminary evaluation of the substation voltage is achieved by calculating the voltage energy distribution and the frequency center. If the frequency center does not exist, the voltage trend is evaluated based on the time series to obtain the evaluation result, taking into account the instantaneous and random nature of voltage fluctuations, and overcoming the problem that traditional methods cannot comprehensively consider the frequency, duration and impact range of voltage fluctuations. Finally, based on the feeder topology, the topological correlation of abnormal distribution transformers on the feeder is analyzed to obtain the voltage evaluation result of the feeder head end. Compared with the current evaluation method, the present invention comprehensively considers time and space, and makes full use of the existing data to evaluate the high and low voltages in the substation in multiple dimensions, which can provide an effective auxiliary decision-making basis for the management of low voltage in the substation. BRIEF DESCRIPTION OF THE DRAWINGS

[0048] Figure 1 Evaluation index diagram for accurate quantitative definition of voltage levels. DETAILED DESCRIPTION

[0049] The technical solution of the present invention is further described in detail below through specific implementation methods. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0050] The present invention provides a time-sharing and segmented high and low voltage evaluation method based on data distribution, through a dynamic and comprehensive evaluation method, referring to Figure 1As shown, the time-sharing and segmented high and low voltage evaluation method based on data distribution includes the following five steps:

[0051] In a specific example, S1 includes: accurately quantifying the voltage level and establishing reasonable evaluation indicators: by accurately quantifying the voltage level, the voltage level is strictly divided into five intervals, and each level is equipped with strictly defined evaluation indicators, namely:

[0052] Severe overvoltage: A voltage value exceeding 110% of the rated value is defined as severe overvoltage;

[0053] Overvoltage: Voltage values ​​within the range of 107%-110% of the rated value are defined as overvoltage;

[0054] Voltage qualified: When the voltage value is within 95%-107% of the rated value, it is defined as voltage qualified;

[0055] Low voltage: The voltage value within the range of 90%-95% of the rated value is defined as low voltage;

[0056] Severe low voltage: A voltage value lower than 90% of the rated voltage is defined as severe low voltage.

[0057] S2 data preprocessing: using a time series-based adaptive data filtering algorithm to filter outliers and / or noise while retaining the natural fluctuation characteristics of the data; in a specific example, S2 further includes:

[0058] Establish a baseline model: Use historical voltage data to train a baseline model to capture the statistical characteristics of the voltage data, including but not limited to the mean, variance, etc.

[0059] Anomaly detection: For newly collected voltage data points, use the baseline model to predict their expected value and confidence interval. The formula for the predicted value is: The confidence interval is: in, is the predicted voltage value at time t, V t-1 ,V t-2 ,V t-3 ,…,V t-n is the historical voltage value, σ is the standard deviation of the predicted value, and k is the confidence level coefficient. The real-time voltage data point is compared with the predicted expected value. If the data point V t Beyond the confidence interval CT t , marking it as a potential outlier.

[0060] Adaptive confidence interval adjustment: According to the volatility of voltage data, the k value is dynamically adjusted to adapt to the fluctuation range of voltages of different levels.

[0061] Identification and processing of outliers: If a voltage data point is outside the confidence interval, it is marked as an anomaly and the outliers are filtered.

[0062] S3 spectrum center estimation: Combine mathematical analysis and signal processing methods to analyze the spectrum center of voltage data, evaluate the intensity of voltage fluctuations, identify the dominant frequency components of voltage fluctuations, and make a preliminary evaluation of voltage.

[0063] In a specific example, the S3 further includes:

[0064] Energy distribution calculation: Through energy distribution calculation, the intensity of voltage fluctuations in different time periods is quantified. Energy calculation is applied to the sliding time window of the voltage signal. The calculation formula of energy E(t) at time t is:

[0065]

[0066] Where T is the time window length and ν(t) is the voltage value at time τ.

[0067] Frequency center estimation: Perform Fourier transform on the voltage signal to obtain the frequency spectrum representation of the voltage signal. The transformation formula is:

[0068] V(f)=FFT(v(t))

[0069] Where v(t) is the voltage signal in the time domain, V(f) is the corresponding frequency domain representation, and f is the frequency. After obtaining the spectrum representation of the voltage signal, the peaks in the spectrum are identified. By identifying these peaks, the dominant frequency of the signal can be determined. Then the frequency center is estimated using the weighted average method, and the formula is:

[0070]

[0071] Among them, f i is the frequency point, V(f i ) is the corresponding spectrum amplitude. If the voltage frequency center can be obtained, the voltage level is determined to determine which level interval the voltage is in. If the frequency center does not exist, the voltage is evaluated based on the time series trend.

[0072] S4 Trend Assessment: Analyze the long-term trend of voltage levels through time series-based trend assessment to help predict and identify potential voltage problems, such as persistently high or low voltage, thereby evaluating the voltage qualification rate.

[0073] In a specific example, the S4 further includes:

[0074] Build an ARIMA model:

[0075]

[0076] Use ACF and PACF plots to identify the autoregressive (AR) parameter (p), difference (I) order (d), and moving average (MA) parameter (q) of the appropriate ARIMA model, and use the Bayesian method to estimate the autoregressive parameter (AR parameter) φ of the ARIMA model i , Moving average parameter (MA parameter) θ j , the variance of the error term σ 2 The estimated parameters are applied to the ARIMA model and fitted to the voltage data:

[0077]

[0078] where c is a constant term. Evaluate the model parameters, especially the autoregressive term φ i and the moving average term θ j , to determine whether there is a trend component. If the differential order d is greater than 0, it may indicate that there is a trend. After obtaining the voltage trend, the voltage evaluation is performed based on the voltage trend.

[0079] S5 spatial analysis: Based on the voltage evaluation results of the low-voltage substation at the end of the feeder, combined with the topological structure and electrical parameters of the feeder, the propagation characteristics of voltage anomalies in the feeder are analyzed to evaluate the voltage level at the head end.

[0080] In a specific example, the S5 further includes:

[0081] The evaluation formula is defined as:

[0082]

[0083] Among them, V start is the voltage prediction value at the feeder head end, V end is the known voltage at the end of the feeder, n is the number of segments on the feeder, R i is the resistance of the i-th segment, X i is the reactance of the i-th segment, α is a proportional coefficient determined according to the line length and system characteristics, used to adjust the effect of resistance and reactance on voltage drop. L is the length of the feeder, and Z

[0084] is the total impedance of the feeder.

[0085] Specifically for this technical solution, the voltage level is accurately quantified in combination with the actual situation of the substation, and the voltage level is strictly divided into five intervals, providing a basis for the evaluation of high and low voltages. By filtering outliers and / or noise through an adaptive data filtering algorithm based on time series, the present invention can be adapted to the evaluation of voltages of different levels. The preliminary evaluation of the substation voltage is achieved by calculating the voltage energy distribution and the frequency center. If the frequency center does not exist, the voltage trend is evaluated based on the time series to obtain the evaluation result, taking into account the instantaneous and random nature of voltage fluctuations, and overcoming the problem that traditional methods cannot comprehensively consider the frequency, duration and impact range of voltage fluctuations. Finally, based on the feeder topology, the topological correlation of abnormal distribution transformers on the feeder is analyzed to obtain the voltage evaluation result at the head end of the feeder. Compared with the current evaluation method, the present invention comprehensively considers time and space, and makes full use of the existing data to evaluate the high and low voltages in the substation in multiple dimensions, which can provide an effective auxiliary decision-making basis for the management of low voltage in the substation.

[0086] The above description is only a preferred embodiment of the present invention and is not intended to limit the present invention. For those skilled in the art, the present invention may have various modifications and variations. Any modification, equivalent replacement, improvement, etc. made within the spirit and principle of the present invention shall be included in the protection scope of the present invention.

Claims

1. Time- and segment-based high and low voltage evaluation method based on data distribution, It is characterized in that The following steps are involved: S1: Accurately quantify the voltage level and establish reasonable evaluation indicators; S2: Use time series-based adaptive data filtering algorithms to filter outliers and / or noise; S3: Analyze the center of the voltage data spectrum, identify voltage spikes or abnormal peaks, evaluate the intensity of voltage fluctuations, identify the dominant frequency components of voltage fluctuations, and make a preliminary evaluation of the voltage; S4: Analyze the long-term trend of voltage level through time series-based trend assessment, assist in predicting and identifying potential voltage problems, and evaluate the voltage qualification rate; S5: Perform spatial analysis of voltage data taking into account feeder topology, and evaluate the voltage level at the feeder head end by analyzing the propagation characteristics of voltage anomalies in the feeder.

2. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 1, It is characterized in that In step S1: Reasonable evaluation indicators include: The voltage is divided into five interval levels, each with strictly defined evaluation indicators: Severe overvoltage: A voltage value exceeding 110% of the rated value is defined as severe overvoltage; Overvoltage: Voltage values ​​within the range of 107%-110% of the rated value are defined as overvoltage; Voltage qualified: When the voltage value is within 90%-107% of the rated value, it is defined as voltage qualified; Low voltage: The voltage value is defined as low voltage within the range of 85%-90% of the rated value; Severe low voltage: A voltage value lower than 85% of the rated voltage is defined as severe low voltage.

3. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 1, It is characterized in that The step S2 also includes: Build a baseline model: Using historical voltage data to train a baseline model to capture statistical characteristics of voltage data, including anomaly detection, wherein the anomaly detection includes anomaly detection of mean and variance; for newly collected voltage data points, using the baseline model to predict their expected values ​​and confidence intervals, and identifying abnormal voltage data; Adaptive confidence interval adjustment: According to the volatility of voltage data, the confidence interval parameter value is dynamically adjusted to adapt to the fluctuation range of voltage at different levels; Identification and processing of outliers: If a voltage data point is outside the confidence interval, it is marked as an anomaly and the outliers are filtered.

4. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 3, It is characterized in that The anomaly detection is specifically as follows: The voltage prediction value is defined as: in, is the predicted voltage value at time t, V t-1 ,V t-2 ,V t-3 ,…,V t-n is the historical voltage value; Express the confidence interval as: Where σ is the standard deviation of the predicted value and k is the confidence level coefficient; Compare the real-time voltage data point with the predicted expected value. If the data point V t Beyond the confidence interval CT t , marking it as a potential outlier.

5. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 3, It is characterized in that The adaptive confidence interval adjustment is specifically as follows: According to the volatility of voltage data, the k value is dynamically adjusted to adapt to the fluctuation range of voltages of different levels.

6. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 1, It is characterized in that The step S3 further comprises: Calculate the energy distribution: Quantify the intensity of voltage fluctuations in different time periods through energy distribution calculation; Frequency center estimation: Perform Fourier transform on the voltage signal to obtain the spectrum representation of the voltage signal. After obtaining the spectrum representation of the voltage signal, the peaks in the spectrum are identified. By identifying these peaks, the dominant frequency of the signal can be determined. The frequency center is then estimated using a weighted average method to determine which level interval the voltage falls within.

7. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 6, It is characterized in that The calculation energy distribution is specifically: Energy calculation is applied to the sliding time window of the voltage signal; the energy E(t) at time t is calculated as: Where T is the time window length and ν(t) is the voltage value at time τ.

8. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 7, It is characterized in that The frequency center estimation is specifically: performing Fourier transform on the voltage signal to obtain the frequency spectrum representation of the voltage signal. The transformation formula is: V(f)=FFT(v(t)) Where v(t) is the voltage signal in the time domain, V(f) is the corresponding frequency domain representation, and f is the frequency. After obtaining the spectrum representation of the voltage signal, the peaks in the spectrum are identified. By identifying these peaks, the dominant frequency of the signal can be determined, and then the frequency center is estimated using the weighted average method. The formula is: Among them, f i is the frequency point, V(f i ) is the corresponding spectrum amplitude; If the voltage frequency center can be obtained, the voltage level is determined to determine which level range the voltage is in. If the frequency center does not exist, the voltage trend is evaluated based on the time series.

9. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 1, It is characterized in that The step S4 further comprises: Build an ARIMA model: Use ACF and PACF plots to identify the autoregressive (AR) parameter (p), difference (I) order (d), and moving average (MA) parameter (q) of the appropriate ARIMA model, and use the Bayesian method to estimate the autoregressive parameter (AR parameter) φ of the ARIMA model i , Moving average parameter (MA parameter) θ j , the variance of the error term σ 2 ; Apply the estimated parameters to the ARIMA model and fit it to the voltage data: Among them, c is the constant term, the evaluation model parameter, and the autoregressive term φ i and the moving average term θ j Determine whether there is a trend component. If the differential order d is greater than 0, it may indicate that there is a trend. After obtaining the voltage trend, perform voltage evaluation based on the voltage trend.

10. The time-sharing and segmented high and low voltage evaluation method based on data distribution according to claim 1, It is characterized in that The step S5 further comprises: By analyzing the propagation characteristics of voltage anomalies in the feeder, the voltage level at the head end is evaluated. The evaluation formula is: Among them, V start is the voltage prediction value at the feeder head end, V end is the known voltage at the end of the feeder, n is the number of segments on the feeder, R i is the resistance of the i-th segment, X i is the reactance of the i-th segment, α is a proportional coefficient determined according to the line length and system characteristics, used to adjust the effect of resistance and reactance on voltage drop. L is the length of the feeder, and Z is the total impedance of the feeder.