Photovoltaic power supply default power generation identification method based on generating capacity characteristic curve

Abstract

The invention provides a photovoltaic power supply default power generation identification method based on a power generation capacity characteristic curve, and belongs to the technical field of intelligent power grids, and the method comprises the following steps: S1, selecting power consumption data of nearly N days based on an identified power consumption user, averagely dividing a day into a plurality of time points, and obtaining currents corresponding to different time points of the power consumption user; s2, determining a current curve by taking the time point as an abscissa and the current as an ordinate; s3, performing fitting based on the determined current curve, and judging whether the current curve is close to a cosine function or not; and S4, evaluating goodness of fit, calculating a fitting residual error, evaluating a distribution condition of the residual error, if the residual error is close to normal distribution, judging that the user is a photovoltaic power generation user, and further judging whether the photovoltaic power generation user is a default power generation user or not. According to the method, illegal power generation behaviors can be accurately identified, illegal occupation and abuse of power resources are effectively restrained, the power data management efficiency is improved, and powerful guarantee is provided for fine management of power enterprises.

Description

Technical Field

[0001] The present invention belongs to the technical field of smart grids, and in particular relates to a method for identifying photovoltaic power supply default power generation based on a power generation characteristic curve. Background Art

[0002] Distributed photovoltaic refers to a form of photovoltaic power generation in which photovoltaic power generation systems are installed on the user side, such as residential roofs, commercial buildings, factories, etc. The electricity generated is mainly consumed locally, and the excess electricity can be connected to the grid and transmitted to the public power grid. At present, in the monitoring of distributed photovoltaics, the spatial distribution of distributed photovoltaic panels on aerospace satellite maps is used to identify distributed photovoltaic systems using intelligent algorithms. However, this method is difficult to implement in practical applications and is not very applicable. There are also methods that obtain the power generation data of benchmark photovoltaic users and other photovoltaic users in the same area during the same period, and use the original data to train BP neural networks to screen users with a high correlation with the power generation data of benchmark photovoltaic users. However, such methods are only in the simulation experiment stage.

[0003] For example, the patent with publication number CN 117173586A discloses a distributed photovoltaic monitoring and identification method, device, equipment and storage medium. By obtaining the geographical coordinates of distributed photovoltaics, the remote sensing image of the distributed photovoltaics and the scene in which they are located are obtained according to the geographical coordinates, and training samples of different scenes are produced based on the remote sensing images. The samples are cut and the scenes are spliced ​​to obtain a training sample set, and the initial photovoltaic recognition model is trained according to the training sample set to obtain a photovoltaic recognition model. Photovoltaic recognition is performed on large images through the photovoltaic recognition model, and photovoltaic recognition results are output. However, the above method relies on geographical coordinates, remote sensing images, etc., which is difficult to implement in practical applications and has low applicability. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, a method for identifying photovoltaic power supply default power generation based on power generation characteristic curve is provided.

[0005] The technical solution adopted by the present invention to solve its technical problem is:

[0006] This technical solution proposes a photovoltaic power generation default identification method based on a power generation characteristic curve, comprising the following steps:

[0007] S1: Based on the identified electricity users, select the electricity consumption data of the past N days, divide a day into several time points on average, and obtain the current corresponding to the electricity users at different time points;

[0008] S2: Determine the current curve with the time point as the horizontal axis and the current as the vertical axis;

[0009] S3: Based on the determined current curve, fitting is performed to determine whether the current curve is close to the cosine function;

[0010] S4: Evaluate the goodness of fit and calculate the residuals of the fit, and calculate the residuals at each time point Calculate the mean square error, which must be less than the threshold, evaluate the distribution of the residuals, and use the Shapiro-Wilk test to test normality. If the residuals are close to a normal distribution, it is judged to be a photovoltaic power generation user, and then determine whether the photovoltaic power generation user is in default of power generation.

[0011] Preferably, the photovoltaic current formula is expressed as:

[0012] I = Jsc*A (1);

[0013] Where I is the photovoltaic current, Jsc is the photovoltaic cell current density under light intensity, and A is the area of ​​the solar panel;

[0014] The formula for the intensity of solar radiation on a sunny day is:

[0015]

[0016] Where I is the horizontal intensity of solar radiation, n is the refractive index of the atmosphere, and z is the solar altitude angle;

[0017] The photovoltaic current curve reaches its peak at noon, and the photovoltaic current curve is close to the cosine function curve.

[0018] Preferably, the power consumption data is processed by sliding average filtering, including:

[0019] S11: sorting the electricity consumption data in chronological order and determining the window size;

[0020] S12: Create a sliding window and slide the window along the data sequence, moving one data point at a time;

[0021] S13: Calculate the average value, at each window position, calculate the average value of the data points in the window;

[0022] S14: Generate smooth data, and use the average value calculated at each window position as the smoothed data point;

[0023] S15: Repeat S12-S14 until all data are processed.

[0024] Preferably, the cosine function formula is expressed as:

[0025]

[0026] Where A is the amplitude, ω is the angular frequency, is the phase, k is the vertical offset;

[0027] When generating electricity, the current is positive, and when not generating electricity, the current is 0. Equation (3) can be simplified as follows:

[0028]

[0029] During the fitting process, the current value corresponds to y, and the time point at the same moment corresponds to x.

[0030] Preferably, the geographical location of the users in the station area is determined, the time zone corresponding to the geographical location is determined, the offset of the time zone relative to the standard time zone is calculated, and the position of the peak segment is adjusted based on the offset of the time zone.

[0031] Preferably, the maximum value in the adjusted peak segment is selected as the amplitude, and the time point corresponding to the amplitude is recorded as p0, so as to obtain the maximum interval in which the current is continuously greater than 0;

[0032] The minimum point is denoted as p min , the maximum point p max , get max((p0-p min ),(p max -p0)) is denoted as A, ω is for Use the daily current curve data and the corresponding point of each point to perform fitting to determine whether the current curve is close to the cosine function.

[0033] Preferably, fitting is performed by the least squares method, which finds the best function match for the data by minimizing the sum of squares of errors, selects the cosine function formula as the fitting function, constructs a loss function, and defines the sum of squares of errors between observed values ​​and fitted values. The fitting function parameters that minimize the loss function are found by derivation.

[0034] Preferably, the Shapiro-Wilk test method includes the following statistic W formula:

[0035]

[0036] In the formula, n is the sample size, a i is the coefficient related to the sample size n, x (i) is the ith observation after sorting, is the sample mean.

[0037] Preferably, it also includes:

[0038] Hypothesis test: Null hypothesis H0: sample data follows normal distribution, alternative hypothesis H1: sample data does not follow normal distribution;

[0039] Calculate the test statistic W, determine the critical value or use the distribution to calculate the value;

[0040] Make a decision: If the p-value is less than the significance level α, reject the null hypothesis and assume that the sample data does not follow a normal distribution. If the p-value is greater than or equal to the significance level α, there is insufficient evidence to reject the null hypothesis and assume that the sample data follows a normal distribution.

[0041] Preferably, electricity consumption data of the past N days are selected for analysis, and if the number of days exceeding N / 2 fits the judgment criteria, the user is considered to be a photovoltaic power generation user.

[0042] Compared with the prior art, the present invention has the following advantages:

[0043] 1. This application realizes accurate analysis of electricity consumption curves, can automatically compare the characteristics of electricity consumption curves, realizes accurate identification of illegal power generation behaviors, effectively identifies and distinguishes whether users have legally connected to distributed photovoltaic power generation equipment, can accurately capture the behavior of privately connecting photovoltaic power generation in violation of regulations, and promptly curb illegal electricity consumption, ensure the safety of the power grid and the fair order of the power market, provide strong technical support for power regulatory departments, and effectively curb the illegal occupation and abuse of power resources.

[0044] 2. This application can automatically correct erroneous records about the category of photovoltaic power generation equipment in user files, ensure the accuracy of power management data, avoid the tediousness and inefficiency of manual verification, and combine with business scenarios to effectively improve the efficiency of power data management. It significantly improves the management efficiency and accuracy of power data, and provides a strong guarantee for the refined management of power companies. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] The above and / or additional aspects and advantages of the present invention will become apparent and easily understood from the description of the embodiments in conjunction with the following drawings, in which:

[0046] Figure 1 It is the overall flow chart of the present invention;

[0047] Figure 2 This is the current curve for a sunny day in winter;

[0048] Figure 3 This is the current curve for a clear autumn day;

[0049] Figure 4 This is a current curve diagram in winter with clear mornings and occasional clouds in the afternoons;

[0050] Figure 5 This is a graph of current with occasional clouds throughout the day in winter. DETAILED DESCRIPTION

[0051] Embodiments of the present invention are described in detail below, examples of which are shown in the accompanying drawings, wherein the same or similar reference numerals throughout represent the same or similar elements or elements having the same or similar functions. The embodiments described below with reference to the accompanying drawings are exemplary and are only used to explain the present invention, and cannot be understood as limiting the present invention.

[0052] like Figure 1-Figure 5 As shown, this embodiment proposes a photovoltaic power supply default power generation identification method based on a power generation characteristic curve, comprising the following steps:

[0053] S1: Based on the identified electricity users, select the electricity consumption data of the past N days, divide a day into several time points on average, and obtain the current corresponding to the electricity users at different time points;

[0054] S2: Determine the current curve with the time point as the horizontal axis and the current as the vertical axis;

[0055] S3: Based on the determined current curve, fitting is performed to determine whether the current curve is close to the cosine function;

[0056] S4: Evaluate the goodness of fit and calculate the residuals of the fit, and calculate the residuals at each time point Calculate the mean square error, which must be less than the threshold, evaluate the distribution of the residuals, and use the Shapiro-Wilk test to test normality. If the residuals are close to a normal distribution, it is judged to be a photovoltaic power generation user, and then determine whether the photovoltaic power generation user is in default of power generation.

[0057] The photovoltaic current formula is expressed as:

[0058] I = Jsc*A (1);

[0059] Where I is the photovoltaic current in amperes (A), and Jsc is the photovoltaic cell current density under light intensity in amperes / square meter (A / m 2 ), this parameter is usually related to the type and quality of photovoltaic cells and the intensity of light. A is the area of ​​the solar panel in square meters (m 2 ), which is the effective area of ​​the solar panel that receives light;

[0060] For a single fixed photovoltaic device, the area of ​​the solar panel is fixed. If the peak current is not reached throughout the day, the photovoltaic current is positively correlated with the light intensity. The light intensity depends on many factors, including the solar altitude angle, atmospheric conditions, ground reflectivity, etc. Therefore, it is necessary to estimate the direct solar radiation intensity on a sunny day. The formula for the sunny day solar radiation intensity is expressed as:

[0061]

[0062] In the formula, I is the horizontal intensity of solar radiation, n is the refractive index of the atmosphere, which is a parameter related to the atmospheric composition, temperature, pressure, etc., and z is the solar altitude angle, that is, the angle between the center of the sun and the horizon;

[0063] When the weather is stable, the atmospheric refractive index in the same area is relatively stable, so the solar radiation intensity curve is close to the cosine function curve, and it can be inferred that the photovoltaic current curve is close to the cosine function curve. Through the analysis of a large amount of actual current curve data, it is found that the photovoltaic current curve generally reaches a peak at noon, while the current curve of conventional electricity users is often not at a peak at noon, and may even be at a trough. Therefore, it is possible to judge whether the current curve may belong to the current curve of a photovoltaic power generation household by fitting whether the user's current curve is close to the cosine function curve.

[0064] The ideal photovoltaic current curve is close to the cosine function curve. However, in some cases, the data may deviate from the offline condition due to weather problems such as cloudy weather within a day. The data needs to be processed by sliding average filtering to achieve the purpose of smoothing the data and removing noise, so as to avoid the influence of individual extreme jump values ​​on subsequent analysis. Subsequent analysis uses the data processed by sliding average. The power consumption data is processed by sliding average filtering, including:

[0065] S11: sorting the electricity consumption data in chronological order and determining the window size;

[0066] S12: Create a sliding window and slide the window along the data sequence, moving one data point at a time;

[0067] S13: Calculate the average value, at each window position, calculate the average value of the data points in the window;

[0068] S14: Generate smooth data, and use the average value calculated at each window position as the smoothed data point;

[0069] S15: Repeat S12-S14 until all data are processed.

[0070] The cosine function formula is expressed as:

[0071]

[0072] Where A is the amplitude, ω is the angular frequency, is the phase, k is the vertical offset;

[0073] When generating electricity, the current is positive, and when not generating electricity, the current is 0, so there is no vertical offset involved. Equation (3) is simplified as:

[0074]

[0075] In the fitting process, the current value corresponds to y, and the time point at the same moment corresponds to x. Other parameters can also be indirectly obtained through the current curve. Generally speaking, the light intensity is the highest at noon, so the peak occurs at noon and does not exceed half an hour before and after.

[0076] Determine the geographical location of the users in the station area, determine the time zone corresponding to the geographical location, calculate the offset of the time zone relative to the standard time zone, and adjust the position of the peak segment based on the offset of the time zone.

[0077] Take the 96-point current curve as an example. That is, there are no more than 2 points before and after the 49th point. The maximum value from the 47th to the 51st point is selected as the amplitude. However, since it spans five time zones and all use the Eastern Time Zone 8 as the standard time, the difference between the time zone in which the time is located and the Eastern Time Zone 8 is considered when selecting points. For example, a point around 53 o'clock is required for noon in the Eastern Time Zone 7. After other regions obtain the basic time zone information, they can achieve adaptive adjustment by configuring the number of offset Eastern Time Zone 8 time points parameter.

[0078] The maximum value in the adjusted peak segment is selected as the amplitude, and the time point corresponding to the amplitude is recorded as p0, and the maximum interval in which the current is continuously greater than 0 is obtained;

[0079] The minimum point is denoted as p min , the maximum point p max , get max((p0-p min ),(p max -p0)) is denoted as A, ω is for At this time, A, ω, It is known that the current curve data of each day and the point position corresponding to each point are used for fitting to determine whether the current curve is close to the cosine function.

[0080] The least squares method can be used for fitting. The least squares method finds the best function matching the data by minimizing the sum of squares of errors. The cosine function formula is selected as the fitting function, and the loss function is constructed. The loss function is usually defined as the sum of squares of errors between the observed value and the fitted value. The fitting function parameters that minimize the loss function are found by derivation.

[0081] After the fitting is completed, the goodness of the fitting needs to be evaluated by calculating the residual of the fitting and checking the parameter distribution. Ideally, the residual should be close to the normal distribution. The electricity consumption data of the past N days are selected for analysis. If the fitting meets the judgment criteria for more than N / 2 days, the user is considered to be a photovoltaic power generation user.

[0082] The Shapiro-Wilk test was used to test normality. The Shapiro-Wilk test includes the following statistic W formula:

[0083]

[0084] In the formula, n is the sample size, a i is the coefficient related to the sample size n, x (i) is the ith observation after sorting, is the sample mean;

[0085] Hypothesis test: Null hypothesis H0: sample data follows normal distribution, alternative hypothesis H1: sample data does not follow normal distribution;

[0086] Calculate the test statistic W, determine the critical value or use the distribution to calculate the value;

[0087] Make a decision: If the p-value is less than the significance level α, reject the null hypothesis and assume that the sample data does not follow a normal distribution. If the p-value is greater than or equal to the significance level α, there is insufficient evidence to reject the null hypothesis and assume that the sample data follows a normal distribution.

[0088] In addition, power curve data can be used instead of current curve data. Since power is affected by current, voltage, and power factor, and for stable power generation users, voltage and power factor are relatively stable within a day, the fluctuations of power curve and current curve are generally similar.

[0089] Although the embodiments of the present invention have been shown and described, it will be appreciated by those skilled in the art that various changes, modifications, substitutions and variations may be made to the embodiments without departing from the principles and spirit of the present invention, and that the scope of the present invention is defined by the claims and their equivalents.

Claims

1. A photovoltaic power source default power generation identification method based on power generation characteristic curve, characterized in that: The following steps are involved: S1: Based on the identified electricity users, select the electricity consumption data of the past N days, divide a day into several time points on average, and obtain the current corresponding to the electricity users at different time points; S2: Determine the current curve with the time point as the horizontal axis and the current as the vertical axis; S3: Based on the determined current curve, fitting is performed to determine whether the current curve is close to the cosine function; S4: Evaluate the goodness of fit by calculating the residuals of the fit and the residuals at each time point Calculate the mean square error, which must be less than the threshold, evaluate the distribution of the residuals, and use the Shapiro-Wilk test to test normality. If the residuals are close to a normal distribution, it is judged to be a photovoltaic power generation user, and then determine whether the photovoltaic power generation user is in default of power generation.

2. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 1 is characterized in that: The photovoltaic current formula is expressed as: I = Jsc*A (1); Where I is the photovoltaic current, Jsc is the photovoltaic cell current density under light intensity, and A is the area of ​​the solar panel; The formula for the intensity of solar radiation on a sunny day is: Where I is the horizontal intensity of solar radiation, n is the refractive index of the atmosphere, and z is the solar altitude angle; The photovoltaic current curve reaches its peak at noon, and the photovoltaic current curve is close to the cosine function curve.

3. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 1 is characterized in that: The power consumption data is processed through sliding average filtering, including: S11: sorting the electricity consumption data in chronological order and determining the window size; S12: Create a sliding window and slide the window along the data sequence, moving one data point at a time; S13: Calculate the average value, at each window position, calculate the average value of the data points in the window; S14: Generate smooth data, and use the average value calculated at each window position as the smoothed data point; S15: Repeat S12-S14 until all data are processed.

4. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 1 is characterized in that: The cosine function formula is expressed as: Where A is the amplitude, ω is the angular frequency, is the phase, k is the vertical offset; When generating electricity, the current is positive, and when not generating electricity, the current is 0. Equation (3) can be simplified as follows: During the fitting process, the current value corresponds to y, and the time point at the same moment corresponds to x.

5. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 4 is characterized in that: Determine the geographical location of the users in the station area, determine the time zone corresponding to the geographical location, calculate the offset of the time zone relative to the standard time zone, and adjust the position of the peak segment based on the offset of the time zone.

6. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 5 is characterized in that: The maximum value in the adjusted peak segment is selected as the amplitude, and the time point corresponding to the amplitude is recorded as p0, and the maximum interval in which the current is continuously greater than 0 is obtained; The minimum point is denoted as p min , the maximum point p max , get max((p0-p min ),(p max -p0)) is denoted as A, ω is for Use the daily current curve data and the corresponding point of each point to perform fitting to determine whether the current curve is close to the cosine function.

7. The photovoltaic power source default power generation identification method according to the power generation characteristic curve of claim 1 is characterized in that: The least squares method is used for fitting. The least squares method finds the best function matching the data by minimizing the sum of squares of errors. The cosine function formula is selected as the fitting function, and the loss function is constructed. The loss function is defined as the sum of squares of errors between the observed value and the fitted value. The fitting function parameters that minimize the loss function are found by derivation.

8. The photovoltaic power source default power generation identification method based on power generation characteristic curve according to claim 1 is characterized in that: The Shapiro-Wilk test method includes the following formula for the statistic W: In the formula, n is the sample size, a i is the coefficient related to the sample size n, x (i) is the ith observation after sorting, and x is the sample mean.

9. A photovoltaic power source default power generation identification method based on power generation characteristic curve according to claim 8, characterized in that: Also includes: Hypothesis test: Null hypothesis H0: sample data follows normal distribution, alternative hypothesis H1: sample data does not follow normal distribution; Calculate the test statistic W, determine the critical value or use the distribution to calculate the value; Make a decision: If the p-value is less than the significance level α, reject the null hypothesis and assume that the sample data does not follow a normal distribution. If the p-value is greater than or equal to the significance level α, there is insufficient evidence to reject the null hypothesis and assume that the sample data follows a normal distribution.

10. The photovoltaic power source default power generation identification method based on power generation characteristic curve according to claim 1, characterized in that: The electricity consumption data of the past N days are selected for analysis. If the number of days exceeding N / 2 fits the judgment criteria, the user is considered to be a photovoltaic power generation user.

Images