Virtual power plant data tracking detection method and system

Abstract

The invention relates to the technical field of electric digital data processing, in particular to a virtual power plant data tracking detection method and system, and the method comprises the steps: collecting a wind speed and a power time sequence in real time; analyzing and confirming the wind speed and power time sequence to obtain a wind speed-generated power function model; according to the wind speed-generated power function model, the wind speed and the power time sequence, determining to obtain the authenticity evaluation value of each power data; according to a vector weighting algorithm and the authenticity evaluation value of each piece of power data, different weights are given, and prediction is carried out to obtain predicted power data; and comparing the predicted power data with the power data corresponding to the predicted time, and performing tracking detection on the virtual power plant based on a comparison result. According to the method, the predicted power data is compared with the real power corresponding to the moment, so that the unsteady state change of the wind driven generator can be quickly confirmed and obtained, and the continuous tracking detection abnormity of the wind driven generator of the virtual power plant can be quickly confirmed and obtained.

Description

Technical Field

[0001] The present invention relates to the technical field of electric digital data processing, and particularly relates to a method and system for tracking and detecting virtual power plant data. Background Art

[0002] A virtual power plant is a conceptual model or entity that realizes power supply by integrating various distributed energy resources (such as solar energy, wind energy, energy storage systems, etc.) and adopting flexible scheduling strategies. Among these resources, wind energy, as a low-cost and renewable power source, occupies an important position. Utilizing wind energy not only helps reduce dependence on traditional fossil fuels and lower the carbon footprint, but also conforms to environmental protection policies and sustainable development goals.

[0003] The predictability of wind energy is crucial for the stability of the power system. Effectively tracking and managing the wind power generation can help the virtual power plant better predict and adjust the operation of the power grid, ensuring the stability and reliability of power supply. However, when predicting wind power generation, the existing technology usually adopts the Long short-term memory (LSTW) algorithm to predict the current data through historical power generation data. Although this method has certain effects, its prediction accuracy still has room for improvement. Especially when dealing with wind changes, the model cannot fully focus on the real time points.

[0004] Therefore, there is an urgent need for a method that can effectively improve the data utilization efficiency of virtual power plants to better continuously track and detect abnormal data. Summary of the Invention

[0005] In order to solve the above technical problems, the purpose of the present invention is to provide a method and system for tracking and detecting virtual power plant data to solve the above problems. To achieve the above purpose, the technical solutions adopted by the present invention are as follows:

[0006] In the first aspect, the present application provides a method for tracking and detecting virtual power plant data, and the method includes:

[0007] Real-time collect wind speed and power generation to obtain a wind speed time series and a power time series;

[0008] Analyze and confirm the wind speed - power generation function model from the wind speed time series and the power time series;

[0009] Confirm the authenticity evaluation value of each power data according to the wind speed - power generation function model, the wind speed time series, and the power time series;

[0010] Divide the wind speed time series into a training set and a prediction set;

[0011] Assign different weights to each power data in the training set according to the vector weighting algorithm and the authenticity evaluation value of each power data, and train a preset neural network model to obtain a prediction model;

[0012] Assign different weights to each power data in the prediction set according to the vector weighting algorithm and the authenticity evaluation value of each power data, and input it into the prediction model for prediction to obtain predicted power data;

[0013] Compare the predicted power data with the power data corresponding to the prediction time, and perform tracking detection on the virtual power plant based on the comparison result, where the prediction time is the time corresponding to the predicted power data.

[0014] Combined with the first aspect, in a possible implementation manner, analyzing and confirming the wind speed-time series and the power-time series to obtain a wind speed-power generation power function model, including:

[0015] Analyze and confirm the authenticity evaluation value of each wind speed data based on the Weibull distribution for the wind speed time series;

[0016] Construct a wind speed-power generation power function model based on the authenticity evaluation value of each wind speed data and the power time series.

[0017] Combined with the first aspect, in a possible implementation manner, analyzing and confirming the authenticity evaluation value of each wind speed data based on the Weibull distribution for the wind speed time series, including:

[0018] Calculate the probability that the time corresponding to each wind speed data is the Weibull center based on the wind speed time series;

[0019] Judge the total number of Weibull centers according to the probability corresponding to each time based on a preset probability threshold, where the probability corresponding to the Weibull center is greater than the preset probability threshold;

[0020] Construct an unsolved wind speed function based on the total number of Weibull centers and a preset Weibull function, where the unsolved wind speed function includes at least one wind speed parameter;

[0021] Perform wind speed parameter optimization processing on all the wind speed parameters based on the least squares method to obtain a wind speed function;

[0022] Calculate an effective wind speed time series based on the wind speed function;

[0023] Calculate the authenticity evaluation value of each wind speed based on the effective wind speed time series and the wind speed time series.

[0024] Combined with the first aspect, in a possible implementation manner, the wind speed parameter optimization processing includes:

[0025] An error desirability evaluation function is constructed based on each wind speed data in the wind speed time series and the data on both sides of each wind speed data;

[0026] A wind speed error sum of squares function is constructed based on the error desirability evaluation function and the wind speed function to be solved;

[0027] Using the wind speed error sum of squares function as the fitting function of the least squares method to fit the wind speed time series and the power time series, and obtaining the solution of each wind speed parameter.

[0028] Combined with the first aspect, in a possible implementation, the preset probability threshold is 0.7.

[0029] Combined with the first aspect, in a possible implementation, a wind speed - power generation power function model is constructed based on the authenticity evaluation value of each wind speed data and the power time series, including:

[0030] Based on the authenticity evaluation value of each wind speed data, the wind speed time series and the power time series, a wind speed - power sequence is statistically obtained;

[0031] Based on the wind speed - power sequence, the wind speed corresponding to the maximum power generation is analyzed;

[0032] Based on the wind speed corresponding to the maximum power generation, a wind speed - power generation power function model to be solved is constructed, and at least one power parameter is included in the wind speed - power generation power function model to be solved;

[0033] Based on the least squares method, power parameter optimization processing is performed on all the power parameters to obtain the wind speed - power generation power function model.

[0034] Combined with the first aspect, in a possible implementation, the power parameter optimization processing includes:

[0035] An error contribution function is constructed based on the authenticity evaluation value of each wind speed data;

[0036] Based on the wind speed - power generation power function model to be solved and the error contribution function, a power error sum of squares function is constructed;

[0037] Using the power error sum of squares function as the fitting function of the least squares method to fit the wind speed - power sequence, and obtaining the solution of the power parameter.

[0038] Combined with the first aspect, in a possible implementation, based on the wind speed - power sequence, analyzing the wind speed corresponding to the maximum power generation includes:

[0039] Calculate the standard deviations on both sides of each wind speed in the wind speed-power sequence respectively to obtain the first standard deviation and the second standard deviation corresponding to each wind speed. The first standard deviation is calculated by performing standard deviation calculation on the power generations corresponding to a number of first wind speeds that increase successively, and the second standard deviation is calculated by performing standard deviation calculation on the power generations corresponding to a number of second wind speeds that increase successively. The first wind speed is less than the second wind speed;

[0040] Calculate the maximum effective wind speed probability corresponding to each wind speed based on the first standard deviation and the second standard deviation corresponding to each wind speed. The maximum effective wind speed probability is used to characterize the probability that the wind speed just reaches the maximum power generation;

[0041] Screen the wind speeds corresponding to the maximum power generation based on the maximum effective wind speed probability.

[0042] In a second aspect, the present application also provides a virtual power plant data tracking and detection system, including:

[0043] A data acquisition module, configured to collect the wind speed and the power generation in real time to obtain a wind speed time series and a power time series;

[0044] A model confirmation module, configured to analyze and confirm the wind speed time series and the power time series to obtain a wind speed-power generation function model;

[0045] A power evaluation module, configured to confirm the authenticity evaluation value of each power data according to the wind speed-power generation function model, the wind speed time series and the power time series;

[0046] A partitioning module, configured to partition the wind speed time series into a training set and a prediction set;

[0047] A training module, configured to assign different weights to each power data in the training set according to the vector weighting algorithm and the authenticity evaluation value of each power data, and train a preset neural network model to obtain a prediction model;

[0048] A prediction module, configured to assign different weights to each power data in the prediction set according to the vector weighting algorithm and the authenticity evaluation value of each power data and input the data into the prediction model for prediction to obtain prediction power data;

[0049] A comparison module, configured to compare the prediction power data with the power data corresponding to the prediction moment, and perform tracking and detection on the virtual power plant based on the comparison result. The prediction moment is the moment corresponding to the prediction power data.

[0050] Combined with the second aspect, in a possible implementation manner, the model confirmation module includes:

[0051] A wind speed module for analyzing and validating the authenticity evaluation value of each wind speed data based on the Weibull distribution for the wind speed time series;

[0052] A model construction module for constructing a wind speed - power generation power function model based on the authenticity evaluation value of each wind speed data and the power time series.

[0053] The present invention has the following beneficial effects:

[0054] In the present invention, by mining the correlation between wind speed and power generation power, a wind speed - power generation power function model is confirmed; and based on the wind speed - power generation power function model, the authenticity evaluation of each power data is confirmed, and different weights are assigned to each power data using the authenticity evaluation value, enabling the prediction model to better focus on more important time points, thereby improving the prediction accuracy. Moreover, based on the relatively accurate prediction results, the non - steady - state changes of the wind turbine can be directly confirmed by comparison, which can effectively improve the data utilization efficiency of the virtual power plant and more conveniently perform continuous tracking and detecting anomalies of the wind turbines in the virtual power plant. BRIEF DESCRIPTION OF THE DRAWINGS

[0055] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the following will briefly introduce the drawings required for use in the description of the embodiments or the prior art. Obviously, the drawings in the following description are only some embodiments of the present invention. For those of ordinary skill in the art, other drawings can be obtained based on these drawings without creative efforts.

[0056] Figure 1 It is a schematic flowchart of a method for tracking and detecting virtual power plant data provided in Embodiment 1 of the present invention;

[0057] Figure 2 It is a wind speed diagram in Embodiment 1 of the present invention;

[0058] Figure 3 It is a schematic flowchart of step S2 provided in Embodiment 1 of the present invention;

[0059] Figure 4 It is a schematic structural diagram of a virtual power plant data tracking and detecting system described in Embodiment 2 of the present invention;

[0060] Figure 5 It is a schematic structural diagram of the model confirmation module described in Embodiment 2 of the present invention. DETAILED DESCRIPTION OF THE EMBODIMENTS

[0061] In order to further elaborate on the technical means and effects adopted by the present invention to achieve the intended invention purpose, the following, in combination with the accompanying drawings and preferred embodiments, details a virtual power plant data tracking and detection method and system proposed according to the present invention, including its specific implementation manner, structure, features, and effects, as follows. In the following description, different "one embodiment" or "another embodiment" do not necessarily refer to the same embodiment. In addition, specific features, structures, or characteristics in one or more embodiments can be combined in any suitable form.

[0062] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as commonly understood by those skilled in the technical field to which this invention belongs.

[0063] Embodiment 1:

[0064] The following specifically describes the specific solution of a virtual power plant data tracking and detection method provided by the present invention in combination with the accompanying drawings.

[0065] Please refer to Figure 1 , which shows a schematic flow chart of the virtual power plant data tracking and detection method provided by an embodiment of the present invention. Specifically, this method includes steps S1, S2, S3, S4, S5, S6, S7, and S8.

[0066] S1. Real-time collect wind speed and power generation to obtain a wind speed time series and a power time series.

[0067] It should be noted that in this embodiment, the wind speed data is collected by an anemometer installed on the wind turbine (collected once per second), and each anemometer is connected to the data acquisition system for real-time data transmission. At the same time, the power generation mentioned in the embodiment is collected on the monitoring system of each wind turbine (collected once per second). And it is divided into a wind speed time series and a power time series according to the number of days. Through the above two data sources, in this embodiment, a wind speed data set in time series and a power generation data set in time series can be obtained. For the convenience of subsequent explanation, therefore, in this embodiment, the wind speed data set in time series is referred to as the wind speed time series, and the power generation data set in time series is referred to as the power time series.

[0068] In this embodiment, considering that the wind speed data collected on the wind turbine may be the result of the combined action of multiple winds, refer to Figure 2The wind speed diagram shows the change of wind speed magnitude over time. That is, a wind speed data may be the result of the combined action of multiple winds. Correspondingly, the power generation of a wind turbine is also generated by the combined push of multiple winds. Therefore, in this embodiment, first, a wind speed-power generation power function model is confirmed according to the wind speed time series and the power time series, and then the authenticity of the power generation power corresponding to each time is determined through the wind speed-power generation power function model. Finally, the authenticity is weighted and combined with the long short-term neural network model to accurately predict the power generation power at the next moment. Therefore, see steps S2 - S6 in detail.

[0069] S2. Analyze and confirm the wind speed-power generation power function model for the wind speed time series and the power time series.

[0070] In this embodiment, considering that a wind speed data may be the result of the combined action of multiple winds, and combining the characteristic that the wind speed presents a Weibull distribution to determine the authenticity of the wind speed data at each time in the time dimension. Then, the authenticity of the wind speed data is used to determine the unique power generation power data corresponding to each wind speed data, and this is used to construct the wind speed-power generation power function model. Based on this, for the convenience of those skilled in the art to understand, the wind speed data mentioned in this embodiment refers to the magnitude of the wind speed corresponding to a time, and similarly, the power generation power data refers to the magnitude of the power generation power corresponding to a time; and the wind speed refers to the wind speed magnitude itself, and at the same time, the power generation power refers to the power generation power magnitude itself.

[0071] Specifically, see Figure 3 , Figure 3 which shows that step S2 includes step S21 and step S22.

[0072] S21. Analyze the wind speed time series based on the Weibull distribution to confirm the authenticity evaluation value of each wind speed data.

[0073] In this embodiment, first, according to the characteristic that the wind speed presents a Weibull distribution, the authenticity of the wind speed is determined in the time dimension. For the wind speed data of the time series collected on the same wind turbine, each wind speed data is the result of the combined push of multiple winds. That is, in this embodiment, first determine how many Weibull centers there are, then use the number of Weibull centers to construct a wind speed function, and then determine the acceptability of each error by comparing with the surrounding data in space and time. Use the minimized constructed error sum function to determine the wind speed function, and determine the authenticity of the wind speed through the wind speed function. Specifically, see steps S211 - S2116.

[0074] S211. Calculate the probability that the moment corresponding to each wind speed data is a Weibull center based on the wind speed time series.

[0075] In this embodiment, the probability corresponding to each moment is confirmed by constructing a probability calculation formula. Among them, the probability calculation formula is:

[0076]

[0077] Among them, W A represents the probability that the moment corresponding to the A-th wind speed data in the wind speed time series is the Weibull center; exp represents the exponential function with the natural constant e as the base, e +[A,A ′ 1...A′q] represents the error of using a straight line with a positive slope to optimally fit the A-th and the q wind speed data on its left side in time series; A ′ 1 represents the first wind speed data on the left side of the A-th wind speed data in time series; A ′ q represents the q-th wind speed data on the left side of the A-th wind speed data in time series; e -[A,A ′ 1 ′ ...A′q′] represents the error of using a straight line with a negative slope to optimally fit the A-th and the q wind speed data on its right side in time series; A ′ 1 ′ represents the first wind speed data on the right side of the A-th wind speed data in time series; A ′ q ′ represents the q-th wind speed data on the right side of the A-th wind speed data in time series; q = 600.

[0078] In this embodiment, considering that for the Weibull distribution formed by the wind, the Weibull center is the moment corresponding to the highest point, so the slope of the straight line formed by fitting with the right-side data should be less than 0, and the slope of the straight line formed by fitting with the left-side data should be greater than 0. And because the change of the wind speed magnitude in time series is continuous, the smoothness of the fitted straight lines is relatively high (the error is small). Therefore, in this technology, the above formula is constructed by fitting the left q data with a straight line with a slope greater than 0 and the right q data with a straight line with a slope less than 0, and the smaller the fitting error, the better, to represent the probability that each data is the Weibull center. At the same time, in this embodiment, considering that the shortest interval time between each gust of wind may be between a few minutes and dozens of minutes, therefore, in the present invention, the value of q is 600. At the same time, those skilled in the art can also select other values of q, and this embodiment does not make specific restrictions.

[0079] S212. Judging the total number of Weibull centers according to the probability corresponding to each moment, and the probability corresponding to the Weibull center is greater than the preset probability threshold.

[0080] Specifically, in this embodiment, the preset probability threshold is 0.7. If the probability corresponding to each moment is greater than 0.7, it proves that the wind generated at this moment is a Weibull center of the wind.

[0081] S213. Construct a wind speed function to be solved based on the total number of Weibull centers and a preset Weibull function. The wind speed function to be solved includes at least one wind speed parameter.

[0082] That is, according to the above analysis results, it can be known that the wind speed function is the result of the cumulative performance of a series of Weibull distributions. Therefore, in this embodiment, the following functional expression is constructed to represent the wind speed function in time series:

[0083]

[0084] Among them, Z(t) represents the wind speed function showing the relationship between the wind speed magnitude and time; m represents the total number of Weibull centers; x represents the wind speed; e represents the natural constant; a i represents the i-th shape parameter. Since the shape parameter is used to describe the arc state of Weibull, so a i is not zero; b i represents the i-th scale parameter; T i represents the moment of the Weibull center corresponding to the i-th wind speed, and T i is not zero.

[0085] It should be noted that the calculation methods for the shape parameter and the scale parameter are prior arts and will not be elaborated in this embodiment; in addition, one Weibull center represents one wind, and its corresponding one shape parameter and one scale parameter. Therefore, the number of shape parameters and scale parameters is equal to the number of Weibull centers.

[0086] In the above calculation formula, represents the Weibull function corresponding to the i-th wind speed; by performing cumulative summation on the Weibull functions corresponding to individual wind speeds, the wind speed function expected in this embodiment can be obtained. And each Weibull function has only one unknown parameter. Therefore, in this embodiment, it is assumed that it is confirmed that there are a total of 20 Weibull centers within 1 hour, then the wind speed function has a total of m = 20 unknown parameters.

[0087] S214. Perform optimization processing on all the wind speed parameters based on the least squares method to obtain the wind speed function.

[0088] To clarify the process of the wind speed parameter optimization processing, refer specifically to steps S2141 - S2143.

[0089] S2141. Construct an error desirability evaluation function based on each wind speed data in the wind speed time series and the data on both sides of each wind speed data.

[0090] In this embodiment, for the unknown wind speed parameter, the least squares method is used to minimize the sum of errors to solve the wind speed parameter. For some points with too large differences, their influence on minimizing the sum of errors is very significant. Therefore, in this embodiment, the following error desirability evaluation function formula is first constructed to determine the error desirability evaluation value of each wind speed data on the wind speed time series:

[0091]

[0092] Among them, P A represents the error desirability evaluation value of the A-th wind speed data on the wind speed time series; exp represents the exponential function with the natural constant e as the base; n represents that there are n wind speed data in the wind speed time series. If the wind speed time series is one-day data, then n = 86400; Z A represents the A-th wind speed data on the wind speed time series, Z A′ represents the wind speed data on the left side of the A-th wind speed data in the wind speed time series, Z A″ represents the wind speed data on the right side of the A-th wind speed data in the wind speed time series in terms of time sequence; Z p represents the p-th wind speed data on the wind speed time series; Z p+1 represents the (p + 1)-th wind speed data on the wind speed time series; || represents the absolute value function.

[0093] In the above calculation formula, |Z A -Z A′ | represents the volatility of the A-th wind speed data on the wind speed time series and the wind speed on the left side of the time sequence; |Z A -Z A″ | represents the volatility of the A-th wind speed data on the wind speed time series and the wind speed on the right side of the time sequence;

[0094] represents the volatility of the A-th wind speed data on the wind speed time series; |Z p -Z p+1 | represents the volatility between the i-th and (i + 1)-th wind speed data on the wind speed time series; represents the sum of the differences between the volatility of the A-th wind speed data and the volatility of the entire wind speed time series.

[0095] In this embodiment, for the wind speed change caused by the wind blowing, it is a continuous change and will not mutate. Therefore, the volatility of the wind speed data and the volatility generated at all adjacent moments as a whole, the smaller the error, the more desirable. Therefore, the above error desirability evaluation function formula can better measure the error desirability of each wind speed data.

[0096] S2142. Construct a sum of squared wind speed error function based on the error desirability evaluation function and the wind speed function to be solved.

[0097] Specifically, in this embodiment, the sum of squared wind speed error function is as follows:

[0098]

[0099] Where S w represents the sum of squared errors of all wind speed data points; n represents the total number of wind speed data; P p represents the error desirability evaluation value of the p-th wind speed data in time series, Z(p) represents the wind speed value calculated by substituting the time corresponding to the p-th wind speed data into the wind speed function, and Z p represents the true wind speed value collected at the time corresponding to the p-th wind speed data; || represents the absolute value function.

[0100] Since in this embodiment, the method of minimizing the sum of squared data point errors is used to determine the m wind speed parameters of the wind speed function, but if there are unacceptable points with large errors, it will cause extremely large errors in the selection of wind speed parameters. This step makes the acquisition of the wind speed function more accurate by reducing the contribution of data points with lower error desirability evaluation values.

[0101] S2143. Use the sum of squared wind speed error function as the fitting function of the least squares method to fit the wind speed time series and the power time series to obtain the solution of each wind speed parameter.

[0102] Among them, the process of using the sum of squared wind speed error function for fitting by the least squares method is prior art and will not be elaborated in this embodiment.

[0103] S215. Calculate the effective wind speed time series based on the wind speed function. That is, in this step, each wind speed data in the effective wind speed time series is calculated by the wind speed function.

[0104] S216. Calculate the authenticity evaluation value of each wind speed based on the effective wind speed time series and the wind speed time series. Specifically, in this embodiment, the calculation formula for the authenticity evaluation value of the wind speed is:

[0105] V A = exp(-|Z(A)-Z A |)

[0106] Where V A represents the authenticity evaluation value of the A-th wind speed data in the wind speed time series; exp represents the exponential function with the natural constant e as the base; Z(A) represents the wind speed obtained by substituting the A-th wind speed data into the wind speed function at the corresponding time; Z Arepresents the true wind speed value collected at the moment corresponding to the wind speed data of the A-th; || represents the absolute value function.

[0107] S22. Construct a wind speed-power generation power function model based on the authenticity evaluation value of each said wind speed data and the power time series.

[0108] In this embodiment, considering that the power generation power of the wind turbine is linearly proportional to the cube of the wind speed, so in this embodiment, a wind speed-power generation power function model is constructed accordingly, and then the authenticity of the power generation power is obtained through the wind speed-power generation power function model.

[0109] At the same time, in this embodiment, it is also considered that when constructing this wind speed-power generation power function model, because the wind speed-power generation power function model requires that the independent variable can only correspond to one dependent variable, and the same historical wind speed in time series can correspond to multiple power generation power data at different times. Here, it is necessary to make the power generation power value corresponding to the more real wind speed data contribute more to it, obtain all the data point pairs required by the wind speed-power generation power function model, construct the wind speed-power generation power function model accordingly, and then through the desirability of these data point pairs, obtain the optimal parameters of the model, and the authenticity of the power generation power data can be obtained using the error of the model. Specifically, refer to steps S221 - S224.

[0110] S221. Statistically obtain a wind speed-power sequence based on the authenticity evaluation value of each said wind speed data, the wind speed time series, and the power time series.

[0111] In the embodiment, since the same wind speed numerical data can correspond to multiple power generation power data at different times. This is not conducive to confirming the wind speed-power generation power function model in the subsequent steps. Therefore, in this step, the power generation power corresponding to each wind speed is confirmed through the following calculation formula:

[0112]

[0113] where G M represents the power generation power that the wind speed M should have; l represents the total number of moments corresponding to the same wind speed M. Assuming that the wind speed M appears 30 times in a day, then l = 30; represents the authenticity evaluation value corresponding to the wind speed data when the wind speed M appears for the j-th time, and G j represents the power generation power data at the moment corresponding to the j-th wind speed data.

[0114] Through the above calculation formula, a unique generated power can be confirmed for each wind speed, thereby overcoming the problem that in the time series, the same historical wind speed value data can correspond to multiple generated power data at different times. At the same time, by using the above method, the contribution degree of each wind speed data can be changed according to the authenticity evaluation value, so that the confirmation of a unique generated power for the wind speed is more in line with the actual situation. And a wind speed-power sequence is formed thereby.

[0115] S222. Analyze the wind speed corresponding to the maximum generated power based on the wind speed-power sequence.

[0116] Since in reality, the generated power of a wind turbine does not increase infinitely with the increase of the wind speed, but is limited by the upper limit of the energy conversion efficiency of the wind turbine. Therefore, in this embodiment, steps S2221 - S2223 are further included to analyze the wind speed corresponding to the maximum generated power.

[0117] S2221. Calculate the standard deviations on both sides of each wind speed in the wind speed-power sequence respectively, to obtain the first standard deviation and the second standard deviation corresponding to each wind speed. The first standard deviation is calculated from the generated powers corresponding to a number of first wind speeds that increase sequentially, and the second standard deviation is calculated from the generated powers corresponding to a number of second wind speeds that increase sequentially. The first wind speed is less than the second wind speed.

[0118] S2222. Calculate the maximum effective wind speed probability corresponding to each wind speed based on the first standard deviation and the second standard deviation corresponding to each wind speed. The maximum effective wind speed probability is used to represent the probability that the wind speed just reaches the maximum generated power.

[0119] Among them, the calculation formula for the maximum effective wind speed probability is as follows:

[0120]

[0121] Among them, H A represents the probability that the wind speed A is just the wind speed that reaches the maximum generated power; represents the standard deviation of the generated power of the wind speed A and the generated powers of the m wind speeds on the left side of the wind speed in the wind speed-power sequence; S1 ′ represents the generated power of the first wind speed located on the left side of the wind speed A; S ′ m represents the generated power of the mth wind speed located on the left side of the wind speed A; represents the standard deviation of the generated power of the wind speed A and the generated powers of the m wind speeds on the right side of the wind speed in the wind speed-power sequence; S1 ″ represents the generated power of the first wind speed located on the right side of the wind speed A; S ′ m ′It represents the power generation power of the m-th wind speed located on the right side of wind speed A; exp represents the exponential function with the natural constant e as the base.

[0122] In the above calculation formula, [S, S1 ′ ...S ′ m represents the first standard deviation calculated from the power generation power corresponding to the first wind speed; [S, S1 ″ ...S ′ m ′ represents the second standard deviation calculated from the power generation power corresponding to the second wind speed. In this embodiment, considering that in reality, the power generation power of the wind turbine will not increase infinitely with the increase of the wind speed, but will be restricted by the upper limit of the energy conversion efficiency of the wind turbine. Therefore, for the wind speed just reaching the maximum power, the left side is stable, so the smaller the better. On this basis, the right side changes greatly, so the larger the better.

[0123] S2223. Screen the wind speed corresponding to the maximum power generation based on the maximum effective wind speed probability.

[0124] S223. Construct a wind speed-power generation power function model to be solved based on the wind speed corresponding to the maximum power generation. The wind speed-power generation power function model to be solved includes at least one power parameter;

[0125] Among them, in this embodiment, the wind speed-power generation power function model to be solved is:

[0126]

[0127] Among them, G(Z) represents the wind speed-power generation power function model to be solved; k represents the power parameter; Z represents the wind speed, that is, the collected sequential wind speed; represents the limit power generation; S represents the wind speed just reaching the maximum power generation.

[0128] Through the above wind speed-power generation power function model to be solved, it can be well represented that the maximum power generation of the wind turbine has a limit. Before this limit, because the cube of the wind speed is linearly proportional to the power generation power, after the limit, the power generation is certain.

[0129] S224. Perform power parameter optimization processing on all the power parameters based on the least squares method to obtain the wind speed-power generation power function model.

[0130] To clarify the process of the wind speed parameter optimization processing, please refer to steps S2241 - S2243 specifically.

[0131] S2241. Construct an error contribution function based on the authenticity evaluation value of each of the wind speed data.

[0132] Among them, the error contribution function in this step is:

[0133]

[0134] Among them, D M represents the error contribution of the wind speed M corresponding to the power generation power; l represents the total number of corresponding moments of the same wind speed M; represents the authenticity evaluation value corresponding to the wind speed data when the wind speed M appears for the jth time.

[0135] S2242. Construct a sum of squared power errors function based on the to-be-solved wind speed-power generation power function model and the error contribution function.

[0136] Specifically, in this embodiment, the sum of squared power errors function is:

[0137]

[0138] Among them, S P represents the sum of squared errors of all power generation power data points; n represents the total number of wind speed data; D p represents the error contribution of the power generation power corresponding to the pth wind speed data in time series; G(p) represents the power generation power calculated by substituting the wind speed collected at the moment corresponding to the pth wind speed data in time series into the to-be-solved wind speed-power generation power function model, and G p represents the power generation power data collected at the moment corresponding to the pth wind speed data in time series.

[0139] Since in this embodiment, the method of minimizing the sum of squares of data point errors is to be used to determine the power parameter k of the to-be-solved wind speed-power generation power function model, but if there are unacceptable points with large errors, it will cause a great error in the selection of the power parameter k. This step balances the data through error contribution, making the confirmation of the power parameter more in line with the actual situation.

[0140] S2243. Fit the wind speed-power sequence with the sum of squared power errors function as the fitting function of the least squares method to obtain the solution of the power parameter.

[0141] S3. Confirm the authenticity evaluation value of each power data according to the wind speed-power generation power function model, the wind speed time series, and the power time series.

[0142] Specifically, in this embodiment, the calculation formula for the authenticity evaluation value of the power data is:

[0143] F p = exp(-|G(Zp ) - G p |)

[0144] Formula Explanation: F p represents the power generation power G corresponding to the moment of the i-th wind speed data in time series. p The authenticity evaluation value of, exp represents the exponential function with the natural constant e as the base; Z(p) represents the wind speed value calculated by substituting the p-th wind speed data corresponding moment into the wind speed function; G(Z p ) represents the wind speed value calculated by substituting the p-th wind speed data corresponding moment in time series into the wind speed function, and then substituting it into the wind speed-power generation power function model to obtain the power generation power value, G p represents the power generation power data collected at the moment corresponding to the p-th wind speed data in time series; || represents the absolute value function.

[0145] S4. Divide the wind speed time series into a training set and a prediction set.

[0146] It should be noted that in this embodiment, the above wind speed time series is further divided. The prediction set consists of ten wind speed data, and the corresponding training set consists of several groups of samples composed of ten wind speed data. And each sample in the training set corresponds to a power generation power data for training the model.

[0147] S5. Assign different weights to each power data in the training set according to the vector weighting algorithm and the authenticity evaluation value of each power data, and train the preset neural network model to obtain a prediction model.

[0148] S6. Assign different weights to each power data in the prediction set according to the vector weighting algorithm and the authenticity evaluation value of each power data and input it into the prediction model for prediction to obtain predicted power data.

[0149] It should be noted that the vector weighting algorithm mentioned in steps S5 and S6 in this embodiment is a prior art and will not be elaborated in this embodiment.

[0150] S7. Compare the predicted power data with the power data corresponding to the prediction moment, and perform tracking detection on the virtual power plant based on the comparison result. The prediction moment is the moment corresponding to the predicted power data.

[0151] In this embodiment, it is judged whether the predicted power data matches the actual power data corresponding to the prediction moment. If the predicted power data matches the actual power data, the power generation power is directly output; if the prediction result does not match the actual power generation power result, the data is continuously tracked to judge whether there is an abnormal situation in the wind turbine, such as whether there is a situation of continuous mismatched data.

[0152] It should be noted that the matching mentioned in the above steps refers to whether the absolute value of the difference between the predicted power data and the actual power data results is within an error range. For example, an error range can be 0 to 50 kW.

[0153] Embodiment 2:

[0154] As Figure 4 shown, this embodiment provides a virtual power plant data tracking and detection system, and the system includes:

[0155] A data acquisition module, configured to collect wind speed and power generation in real time to obtain a wind speed time series and a power time series;

[0156] A model confirmation module, configured to analyze and confirm the wind speed time series and the power time series to obtain a wind speed - power generation power function model;

[0157] A power evaluation module, configured to confirm the authenticity evaluation value of each power data according to the wind speed - power generation power function model, the wind speed time series, and the power time series;

[0158] A division module, configured to divide the wind speed time series into a training set and a prediction set;

[0159] A training module, configured to assign different weights to each power data in the training set according to the vector weighting algorithm and the authenticity evaluation value of each power data, and train a preset neural network model to obtain a prediction model;

[0160] A prediction module, configured to assign different weights to each power data in the prediction set according to the vector weighting algorithm and the authenticity evaluation value of each power data and input it into the prediction model for prediction to obtain predicted power data;

[0161] A comparison module, configured to compare the predicted power data with the power data corresponding to the prediction moment, and perform tracking and detection on the virtual power plant based on the comparison result, where the prediction moment is the moment corresponding to the predicted power data.

[0162] In some specific embodiments, as Figure 5 shown, the model confirmation module includes:

[0163] A wind speed module, configured to analyze and confirm the authenticity evaluation value of each wind speed data based on the Weibull distribution for the wind speed time series;

[0164] A model construction module, configured to construct a wind speed - power generation power function model based on the authenticity evaluation value of each wind speed data and the power time series.

[0165] It should be noted that regarding the system in the above embodiments, the specific manners in which each module performs operations have been described in detail in the embodiments related to the method, and will not be elaborated here.

[0166] It should be noted that the above sequence of the embodiments of the present invention is only for description and does not represent the superiority or inferiority of the embodiments. The processes depicted in the accompanying drawings do not necessarily require the specific order or continuous order shown to achieve the desired results. In some embodiments, multitasking and parallel processing are also possible or may be advantageous.

[0167] Each embodiment in this specification is described in a progressive manner. For the same or similar parts among the embodiments, reference can be made to each other, and the key point of each embodiment is to illustrate the differences from other embodiments.

Claims

1. A virtual power plant data tracking and detection method, characterized in that: The method comprises: Real-time collection of wind speed and power generation to obtain wind speed time series and power time series; Analyze and confirm the wind speed time series and the power time series to obtain a wind speed-power generation function model; Confirm and obtain the authenticity evaluation value of each power data according to the wind speed-power generation function model, the wind speed time series and the power time series; Dividing the wind speed time series into a training set and a prediction set; Assigning different weights to each power data in the training set according to a vector weighting algorithm and an authenticity evaluation value of each power data, and training a preset neural network model to obtain a prediction model; Assigning different weights to each power data in the prediction set according to a vector weighting algorithm and a authenticity evaluation value of each power data and inputting the weights into the prediction model for prediction to obtain predicted power data; The predicted power data is compared with the power data corresponding to the predicted time, and the virtual power plant is tracked and detected based on the comparison result, wherein the predicted time is the time corresponding to the predicted power data.

2. The virtual power plant data tracking and detection method according to claim 1 is characterized in that: The wind speed time series and the power time series are analyzed and confirmed to obtain a wind speed-power generation function model, including: Analyze the wind speed time series based on Weibull distribution to confirm the authenticity assessment value of each wind speed data; A wind speed-generation power function model is constructed based on the authenticity evaluation value of each wind speed data and the power time series.

3. The virtual power plant data tracking and detection method according to claim 2 is characterized in that: The wind speed time series is analyzed based on Weibull distribution to confirm the authenticity assessment value of each wind speed data, including: Calculate the probability that the moment corresponding to each wind speed data is the Weber center based on the wind speed time series; The total number of Weber centers is obtained by judging the probability corresponding to each moment according to a preset probability threshold, and the probability corresponding to the Weber center is greater than the preset probability threshold; A wind speed function to be solved is constructed based on the total number of Weber centers and a preset Weber function, wherein the wind speed function to be solved includes at least one wind speed parameter; Performing wind speed parameter optimization processing on all the wind speed parameters based on the least square method to obtain a wind speed function; Calculate an effective wind speed time series based on the wind speed function; The authenticity evaluation value of each wind speed is calculated based on the effective wind speed time series and the wind speed time series.

4. The virtual power plant data tracking and detection method according to claim 3 is characterized in that: The wind speed parameter optimization process includes: The error desirability evaluation function is constructed according to each wind speed data in the wind speed time series and the data located on both sides of each wind speed data; A wind speed error square sum function is constructed based on the error desirability evaluation function and the wind speed function to be solved; The wind speed error sum of square function is used as the fitting function of the least square method to fit the wind speed time series and the power time series to obtain the solution of each wind speed parameter.

5. The virtual power plant data tracking and detection method according to claim 3 is characterized in that: The preset probability threshold is 0.

7.

6. The virtual power plant data tracking and detection method according to claim 2 is characterized in that: A wind speed-generation power function model is constructed based on the authenticity evaluation value of each wind speed data and the power time series, including: Obtaining a wind speed-power sequence based on the authenticity evaluation value of each wind speed data, the wind speed time series and the power time series statistics; Obtaining the wind speed corresponding to the maximum power generation based on the wind speed-power sequence analysis; A wind speed-generation power function model to be solved is constructed based on the wind speed corresponding to the maximum power generation, wherein the wind speed-generation power function model to be solved includes at least one power parameter; All the power parameters are optimized based on the least square method to obtain a wind speed-power generation function model.

7. The virtual power plant data tracking and detection method according to claim 6 is characterized in that: The power parameter optimization process includes: Constructing an error contribution function based on the authenticity evaluation value of each wind speed data; A power error square sum function is constructed based on the wind speed-generation power function model to be solved and the error contribution function; The wind speed-power sequence is fitted using the power error square sum function as the fitting function of the least square method to obtain a solution for the power parameter.

8. The virtual power plant data tracking and detection method according to claim 6, characterized in that: The wind speed corresponding to the maximum power generation is obtained based on the wind speed-power sequence analysis, including: Calculate the standard deviations on both sides of each wind speed in the wind speed-power sequence respectively to obtain a first standard deviation and a second standard deviation corresponding to each wind speed, wherein the first standard deviation is obtained by calculating the standard deviation of a plurality of sequentially increasing power generation corresponding to the first wind speed, and the second standard deviation is obtained by calculating the standard deviation of a plurality of sequentially increasing power generation corresponding to the second wind speed, and the first wind speed is less than the second wind speed; The maximum effective wind speed probability corresponding to each wind speed is calculated based on the first standard deviation and the second standard deviation corresponding to each wind speed, and the maximum effective wind speed probability is used to characterize the probability that the wind speed has just reached the maximum power generation power; The wind speed corresponding to the maximum power generation is selected based on the probability of the maximum effective wind speed.

9. A virtual power plant data tracking and detection system, characterized in that: include: A data acquisition module is used to collect wind speed and power generation in real time to obtain wind speed time series and power time series; A model confirmation module, used to analyze and confirm the wind speed time series and the power time series to obtain a wind speed-power generation function model; A power evaluation module, used to confirm the authenticity evaluation value of each power data according to the wind speed-generation power function model, the wind speed time series and the power time series; A division module, used for dividing the wind speed time series into a training set and a prediction set; A training module, used for assigning different weights to each power data in the training set according to a vector weighting algorithm and an authenticity evaluation value of each power data, and training a preset neural network model to obtain a prediction model; A prediction module, used for assigning different weights to each power data in the prediction set according to a vector weighting algorithm and an authenticity evaluation value of each power data, and inputting the weights into the prediction model for prediction to obtain predicted power data; A comparison module is used to compare the predicted power data with the power data corresponding to the predicted time, and to track and detect the virtual power plant based on the comparison result, wherein the predicted time is the time corresponding to the predicted power data.

10. The virtual power plant data tracking and detection system according to claim 9, characterized in that: The model confirmation module includes: Wind speed module, used to analyze the wind speed time series based on Weibull distribution to confirm the authenticity evaluation value of each wind speed data; The model building module is used to build a wind speed-generation power function model based on the authenticity evaluation value of each wind speed data and the power time series.

Images