Reconstruction method of hourly output data of wind and light power station based on reservoir high-low periods

Abstract

The invention relates to a method for reconstructing hourly output data of wind and light power stations based on reservoir full-dry periods, and the method comprises the following steps: screening the hourly output data of all wind and light power stations in the sections in the full year based on the time of the full-dry periods and the full-dry periods of reservoirs with different sections; carrying out anomaly detection and cleaning processing on the screened output data, and carrying out missing value interpolation; extracting a typical scene from the interpolated output data to obtain hourly output data of the wind and light power station in different output strength scenes corresponding to the time of the wet season and the dry season of the reservoir; based on the capacity and efficiency of each wind power plant and each photovoltaic power station in a section reservoir, the output weights of each wind power plant and each photovoltaic power station in the section are set, and the sum of hourly output data of the wind and light power stations in different output strength scenes corresponding to the time of a wet season and a dry season of the reservoir is calculated through the output weights of each wind power plant and each photovoltaic power station in the section. And obtaining total hourly output data of the wind and light power station in different force output scenes in the section.

Description

Technical Field

[0001] The present invention relates to the technical field of big data technology, and in particular to a method for reconstructing hourly output data of wind and solar power stations based on reservoir flood and dry periods. Background Art

[0002] Promoting the consumption and utilization of renewable energy at a high rate; adhering to the requirements of innovation and high-quality development of renewable energy, comprehensively promoting the large-scale development and high-quality development of wind power and solar power generation, accelerating the construction of a modern energy system, and building a number of multi-energy complementary clean energy bases are important paths to promote the realization of high-quality economic and social development. As the speed of integrated development and construction of water, wind and light is further accelerated, the water, wind and light clean energy base will gradually become an important form of energy operation.

[0003] With the rapid growth of wind and solar power in various river basins, the randomness, volatility and intermittency of wind and solar power output have increasingly serious impacts on the power grid. At the same time, the scale of wind and solar power output data is also increasing rapidly, and the quality of data will directly affect the accuracy of wind and solar power output. The inability to accurately grasp the wind and solar power output law will affect the accuracy of complementary regulation of the integrated system of water, wind and solar energy bases, and at the same time affect the quality of power grid power data and the stable operation of the power grid.

[0004] Most of the existing data quality detection and cleaning reconstruction are aimed at the output scenario of a single typical wind and solar power station, without considering the problem of inconsistency of wind and solar power output under the influence of the section transmission capacity of multiple stations in the entire section and the output of cascade power stations in the basin at different times. Moreover, after the existing data is cleaned and reconstructed, most of them output a single result, without distinguishing the different output scenarios of wind and solar power output characteristics, lacking the extraction of typical wind and solar power output scenarios, which is not conducive to the subsequent analysis and mining of wind and solar power output rules. Summary of the invention

[0005] The present invention provides a reconstruction method for the hourly output data of wind and solar power stations based on the wet and dry periods of the reservoir. By performing abnormal detection and cleaning and reconstruction on the output data of wind power stations and photovoltaic power stations in the water, wind and solar clean energy base, the quality of wind and solar output data is ensured. By extracting typical scenarios, the mining of wind and solar output rules is strengthened, the complementary capacity of the integrated system of the water, wind and solar clean energy base is improved, and the quality of power grid power data and the safe and stable operation of the power grid are guaranteed.

[0006] The present invention solves the above technical problems as follows: A method for reconstructing hourly output data of wind and solar power stations based on the flood and dry periods of a reservoir, comprising the following steps:

[0007] Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year are screened out, and the screened output data are subjected to abnormal detection and cleaning;

[0008] Interpolate missing values ​​for output data that has undergone anomaly detection and cleaning;

[0009] Extract k typical output intensity scenarios from the interpolated output data, and obtain the hourly output data of the wind and solar power stations under different output intensity scenarios corresponding to the reservoir's flood season and dry season;

[0010] Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of each wind farm and photovoltaic power station in the cross-section are established. The hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the flood season and dry season of the reservoir are summed up through the output weights of each wind farm and photovoltaic power station in the cross-section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the cross-section.

[0011] Preferably, based on the flood season and dry season of reservoirs in different sections, the hourly output data of the flood season and dry season corresponding to all wind and solar power stations in the section are screened out throughout the year, and the screened output data are subjected to abnormal detection and cleaning processing, including:

[0012] Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year are screened out;

[0013] Perform dimensionality reduction processing on the screened output data;

[0014] The output data that has completed dimensionality reduction processing is denoised and cleaned based on the DBSCAN clustering algorithm.

[0015] Preferably, the dimensionality reduction processing of the screened output data includes:

[0016] The flood season and dry season of reservoirs in different sections are selected to select the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year to establish the standardized output sample X (x1, x2, x3...x 23 , x 24 );

[0017] Based on the daily power station output characteristics, the output curve skewness (S) and the daily output peak mean (E) of the standardized output sample X are extracted. peak ) is processed by dimensionality reduction to obtain the output feature matrix O i (E peak , S), the formula is as follows:

[0018]

[0019] Where: S represents the degree of left or right deviation of the power station daily output curve about the point l in the day. If it is greater than 0, it means right deviation about point l; p is the set of possible values ​​of point l. Considering the influence of different diameters on the direct radiation time, different sections may have the same value, and the same section has a certain value; l is selected as the l within p that makes the S value of each power station in a certain section the smallest; x l+i is the power station output at time l+i; E peak Represents the average of the four points with the largest daily output of the power station; m i It is the ith largest daily output value of the power station.

[0020] Preferably, the denoising and cleaning of the output data after dimensionality reduction processing based on the DBSCAN clustering algorithm includes:

[0021] Set the DBSCAN parameter scanning radius (Eps) and the minimum number of included points (k), and increase the model parameters to calculate the output characteristic matrix O of the power station. i (E peak , S) for coordinate centrifugation, the formula is as follows:

[0022] T=C×n×|S| q

[0023] Where: T is the value of S after centrifugal deviation. The larger the absolute value of this value, the greater the skewness of the daily output curve; C takes 1 when S is greater than 0 and takes -1 when it is less than 0, indicating left or right deviation; n is the uniform scaling coefficient. The larger the value, the greater the output characteristic matrix O in clustering. i The S direction is compared to E peak The more sensitive the direction is; q is the eccentric coefficient, the value is greater than 1, the larger the eccentric coefficient is, the more sensitive the output feature matrix O is when clustering. i The higher the centrifugal effect in the S direction.

[0024] Centrifugal denoising is performed according to the proportion of noise data in the noise sequence to the total data (generally 2% to 31%), and the denoised feature matrix data sets D corresponding to the flood and dry periods of reservoirs in different sections under different noise rates are obtained. 0，x , x is the proportion of noise data to the total data.

[0025] Preferably, the interpolation of missing values ​​of the output data after abnormality detection and cleaning includes:

[0026] For the non-continuous missing data in the output data that has been processed by anomaly detection and cleaning, the mean of the hourly output data of the two days before and after is taken to interpolate all data sets;

[0027] For the continuous missing data in the output data after anomaly detection and cleaning, multiple historical similar days are found according to the characteristic values ​​of the meteorological data on the missing day, and the average of the hourly output data of multiple historical similar days is taken to interpolate all data sets.

[0028] Preferably, the interpolated output data is extracted into k typical output intensity scenes to obtain hourly output data of wind and solar power stations in different output intensity scenes corresponding to the reservoir's flood season and dry season, including:

[0029] Select Euclidean distance as the sample point distance, determine the k value according to the different force types set, and extract k typical scenes of force output;

[0030] For data sets D with different noise rates 1，x Perform Kmeans++ clustering respectively to generate corresponding different clustering result sets D 2，x ; For the clustering result set D 2，x Use the ratio F as an indicator and use the remapped original output feature matrix O when calculating the ratio F. i (E peak ,slop), the formula is as follows:

[0031]

[0032] Where: the ratio F is the ratio of the sum of squares of the between-class deviations to the sum of squares of the within-class deviations. The larger the ratio, the more reasonable the clustering. p is the number of stations in the basin. N k and are the sample size and sample mean of the kth subclass, respectively. Represents X j The value of the i-th sample observation in the k-th subclass; the term (NK) / (K-1) is introduced to eliminate the influence of the sample size N and the number of clusters K on the results;

[0033] Take the inflection point on the curve of noise rate and F value to infer the noise rate X suitable for the target section; after determining the noise rate, take the corresponding data set D 1，X is the final cleaned data, and the corresponding clustering result set D 2，X is the final clustering result;

[0034] Typical scenarios are extracted from the clustering results with a set confidence level (95%) to obtain the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of wind and solar power stations.

[0035] Preferably, based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of each wind farm and photovoltaic power station in the cross-section are established, and the hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the flood season and dry season of the reservoir are summed through the output weights of each wind farm and photovoltaic power station in the cross-section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the cross-section, including:

[0036] Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of the wind farm and photovoltaic power station in the cross-section are set. The formula is as follows:

[0037] w lw =N lw *h lw

[0038] Where: w lw is the weight of the lth wind power station; N lw is the installed capacity of the lth wind power station; h lw is the annual utilization hours of the lth wind power station;

[0039]

[0040] Where: w w is the sum of the weights of all wind power stations in the section;

[0041] w lp =N lp *h lp

[0042] Where: w lp is the weight of the lth photovoltaic power station; N lp is the installed capacity of the lth photovoltaic power station; h lp is the annual utilization hours of the lth photovoltaic power station;

[0043]

[0044] Where: w p is the sum of the weights of all PV power stations in the section;

[0045] The hourly output of all wind power stations in different output intensity scenarios corresponding to the flood season and dry season of the reservoir is summed up through the output weights of each wind farm and photovoltaic power station in the section, that is, the hourly output of all wind farms in the section is equivalent to the hourly output of one wind farm; the hourly output of all photovoltaic power stations is equivalent to the hourly output of one photovoltaic power station, and the total hourly output data of wind and photovoltaic power stations in different output intensity scenarios in the section are obtained. The formula is as follows:

[0046]

[0047] Where: P wt is the equivalent output of the equivalent wind power station in the section at time t; P lw，t is the output of the lth wind power station at time t; is the output weight ratio of the lth wind power station;

[0048]

[0049] Where: P pt is the equivalent output of the equivalent photovoltaic power station in the section at time t; P lp，t is the output of the lth PV power station at time t; is the output weight ratio of the lth photovoltaic power station.

[0050] The present invention also provides a dispatching device for a water-wind-solar complementary power system based on cascade reservoirs, comprising:

[0051] The data cleaning module is used to filter out the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year based on the flood season and dry season of reservoirs in different sections, and perform abnormal detection and cleaning on the filtered output data;

[0052] Data interpolation module, used to interpolate missing values ​​of output data that has undergone anomaly detection and cleaning;

[0053] The typical scenario extraction module is used to extract k typical scenarios of output intensity based on the interpolated output data, and obtain the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of the wind and solar power stations;

[0054] The equivalent output calculation module is used to establish the output weights of each wind farm and photovoltaic power station in the section based on the capacity and efficiency of each wind farm and photovoltaic power station in the section reservoir. The hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed up through the output weights of each wind farm and photovoltaic power station in the section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the section.

[0055] The present invention also provides a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for reconstructing the hourly output data of wind and solar power stations based on the reservoir's wet and dry periods are implemented as described above.

[0056] The present invention also provides an electronic device, comprising a memory and a processor: the memory is used to store computer-executable instructions, the processor is used to execute the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the method for reconstructing the hourly output data of wind and solar power stations based on the reservoir's wet and dry periods are implemented as described above.

[0057] The beneficial effects of the present invention are as follows: the present invention cleans and denoises the hourly output data of wind and solar power stations in the section based on the wet and dry periods of the reservoir through DBSCAN clustering, and obtains complete and effective output data through interpolation of the previous and next daily average values ​​and multiple historical similar daily average values. Through Kmeans++ clustering, typical high, medium and low output scenarios are extracted to obtain the hourly output data of wind and solar power stations in each scenario, thereby improving the quality of the hourly output data of wind and solar power, facilitating the analysis and grasping of the changing laws of different wind and solar power output scenarios during the wet and dry periods of the reservoir, improving the research efficiency of the water-wind-solar complementary output data in the section, and facilitating the improvement of the stability of the water-wind-solar output in the section.

[0058] The above description is only an overview of the technical solution of the present invention. In order to more clearly understand the technical means of the present invention and implement it according to the contents of the specification, the following is a detailed description of the preferred embodiments of the present invention in conjunction with the accompanying drawings. The specific implementation of the present invention is given in detail by the following embodiments and their accompanying drawings. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] The drawings described herein are used to provide a further understanding of the present invention and constitute a part of this application. The exemplary embodiments of the present invention and their descriptions are used to explain the present invention and do not constitute an improper limitation of the present invention. In the drawings:

[0060] Figure 1 A flow chart of a reconstruction method based on hourly output data of wind and solar power stations during the flood and dry periods of a reservoir provided in Example 1;

[0061] Figure 2 This is a module diagram of a dispatching device for a water-wind-solar complementary power system based on cascade reservoirs according to Example 2. DETAILED DESCRIPTION

[0062] The principles and features of the present invention are described below in conjunction with the accompanying drawings. The examples given are only used to explain the present invention and are not used to limit the scope of the present invention.

[0063] Example 1

[0064] This embodiment provides a method for reconstructing hourly output data of wind and solar power stations based on the flood and dry periods of the reservoir, including the following steps:

[0065] S1. Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year are screened out, and the screened output data is subjected to abnormal detection and cleaning, including:

[0066] 1.1 Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year are screened out;

[0067] 1.2 Perform dimensionality reduction processing on the selected output data;

[0068] The flood season and dry season of reservoirs in different sections are selected to select the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year to establish the standardized output sample X (x1, x2, x3...x 23 , x 24 );

[0069] Based on the daily power station output characteristics, the output curve skewness (S) and the daily output peak mean (E) of the standardized output sample X are extracted. peak ) is processed by dimensionality reduction to obtain the output feature matrix O i (E peak , S), the formula is as follows:

[0070]

[0071] Where: S represents the degree of left or right deviation of the power station daily output curve about the point l in the day. If it is greater than 0, it means right deviation about point l; p is the set of possible values ​​of point l. Considering the influence of different diameters on the direct radiation time, different sections may have the same value, and the same section has a certain value; l is selected as the l within p that makes the S value of each power station in a certain section the smallest; x l+i is the power station output at time l+i; E peak Represents the average of the four points with the largest daily output of the power station; m i It is the i-th largest output value of the power station per day.

[0072] 1.3 Based on the DBSCAN clustering algorithm, the output data after dimensionality reduction is denoised and cleaned, including:

[0073] Set the DBSCAN parameter scanning radius (Eps) and the minimum number of included points (k), and increase the model parameters to calculate the output characteristic matrix O of the power station. i (E peak , S) for coordinate centrifugation, the formula is as follows:

[0074] T=C×n×|S| q

[0075] Where: T is the value of S after centrifugal deviation. The larger the absolute value of this value, the greater the skewness of the daily output curve; C takes 1 when S is greater than 0 and takes -1 when it is less than 0, indicating left or right deviation; n is the uniform scaling coefficient. The larger the value, the greater the output characteristic matrix O in clustering. i The S direction is compared to E peak The more sensitive the direction is; q is the eccentric coefficient, the value is greater than 1, the larger the eccentric coefficient is, the more sensitive the output feature matrix O is when clustering. i The higher the centrifugal effect in the S direction.

[0076] Centrifugal denoising is performed on 2% to 31% of the noise sequence to obtain the denoised feature matrix data sets D corresponding to the flood and dry periods of reservoirs in different sections under different noise rates. 0，x , x is the proportion of noise data to the total data.

[0077] S2. Interpolate missing values ​​of output data that have undergone abnormality detection and cleaning;

[0078] 2.1 For the non-continuous missing data in the output data that has been processed by anomaly detection and cleaning, the mean of the hourly output data of the two days before and after is taken to interpolate all data sets;

[0079] 2.2 For the continuous missing data in the output data after anomaly detection and cleaning, find multiple historical similar days based on the characteristic values ​​of the meteorological data on the missing day, and take the mean of the hourly output data of multiple historical similar days to interpolate all data sets.

[0080] S3, extracting k typical output intensity scenarios from the interpolated output data, and obtaining the hourly output data of the wind and solar power stations under different output intensity scenarios corresponding to the reservoir's flood season and dry season;

[0081] Select Euclidean distance as the sample point distance, determine the k value according to the different force types set, and extract k typical scenes of force output;

[0082] For data sets D with different noise rates 1，x Perform Kmeans++ clustering respectively to generate corresponding different clustering result sets D 2，x ; For the clustering result set D 2，x Use the ratio F as an indicator and use the remapped original output feature matrix O when calculating the ratio F. i (E peak ,slop), the formula is as follows:

[0083]

[0084] Where: the ratio F is the ratio of the sum of squares of the between-class deviations to the sum of squares of the within-class deviations. The larger the ratio, the more reasonable the clustering. p is the number of stations in the basin. N k and are the sample size and sample mean of the kth subclass, respectively. Represents X j The value of the i-th sample observation in the k-th subclass; the term (NK) / (K-1) is introduced to eliminate the influence of the sample size N and the number of clusters K on the results;

[0085] Take the inflection point on the curve of noise rate and F value to infer the noise rate X suitable for the target section; after determining the noise rate, take the corresponding data set D 1，X is the final cleaned data, and the corresponding clustering result set D 2，X is the final clustering result;

[0086] Typical scenarios were extracted from the clustering results with a confidence level of 95%, and the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of wind and solar power stations were obtained.

[0087] S4. Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of each wind farm and photovoltaic power station in the cross-section are established. The hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed through the output weights of each wind farm and photovoltaic power station in the cross-section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the cross-section, including:

[0088] 4.1 Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of the wind farm and photovoltaic power station in the cross-section are set. The formula is as follows:

[0089] w lw =N lw *h lw

[0090] Where: w lw is the weight of the lth wind power station; N lw is the installed capacity of the lth wind power station; h lw is the annual utilization hours of the lth wind power station;

[0091]

[0092] Where: w w is the sum of the weights of all wind power stations in the section;

[0093] w lp =N lp *h lp

[0094] Where: w lp is the weight of the lth photovoltaic power station; N lp is the installed capacity of the lth photovoltaic power station; h lp is the annual utilization hours of the lth photovoltaic power station;

[0095]

[0096] Where: w p is the sum of the weights of all PV power stations in the section;

[0097] 4.2 The hourly output of all wind power stations in different output intensity scenarios corresponding to the flood season and dry season of the reservoir is summed up through the output weights of each wind farm and photovoltaic power station in the section, that is, the hourly output of all wind farms in the section is equivalent to the hourly output of one wind farm; the hourly output of all photovoltaic power stations is equivalent to the hourly output of one photovoltaic power station, and the total hourly output data of wind and photovoltaic power stations in different output intensity scenarios in the section are obtained. The formula is as follows:

[0098]

[0099] Where: P wt is the equivalent output of the equivalent wind power station in the section at time t; P lw，t is the output of the lth wind power station at time t; is the output weight ratio of the lth wind power station;

[0100]

[0101] Where: P pt is the equivalent output of the equivalent photovoltaic power station in the section at time t; P lp，t is the output of the lth PV power station at time t; is the output weight ratio of the lth photovoltaic power station.

[0102] Example 2

[0103] This embodiment provides a dispatching device for a water-wind-solar hybrid power system based on cascade reservoirs, including:

[0104] The data cleaning module is used to filter out the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year based on the flood season and dry season of reservoirs in different sections, and perform abnormal detection and cleaning on the filtered output data;

[0105] Data interpolation module, used to interpolate missing values ​​of output data that has undergone anomaly detection and cleaning;

[0106] The typical scenario extraction module is used to extract k typical scenarios of output intensity based on the interpolated output data, and obtain the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of the wind and solar power stations;

[0107] The equivalent output calculation module is used to establish the output weights of each wind farm and photovoltaic power station in the section based on the capacity and efficiency of each wind farm and photovoltaic power station in the section reservoir. The hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed up through the output weights of each wind farm and photovoltaic power station in the section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the section.

[0108] Example 3

[0109] The present invention also provides a computer storage medium, wherein the computer-readable storage medium stores a computer program, and when the computer program is executed by a processor, the steps of the method for reconstructing the hourly output data of wind and solar power stations based on the reservoir's wet and dry periods as described in Example 1 are implemented.

[0110] Example 4

[0111] The present invention also provides an electronic device, comprising a memory and a processor: the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions, and when the computer-executable instructions are executed by the processor, the steps of the method for reconstructing the hourly output data of wind and solar power stations based on the reservoir's wet and dry periods are implemented as described in Example 1.

[0112] The above description is only a preferred embodiment of the present invention and does not limit the present invention in any form. Any ordinary technician in the industry can smoothly implement the present invention as shown in the drawings and described above. However, any equivalent changes, modifications and evolutions made by technicians familiar with the profession without departing from the scope of the technical solution of the present invention using the technical content disclosed above are all equivalent embodiments of the present invention. At the same time, any equivalent changes, modifications and evolutions made to the above embodiments based on the essential technology of the present invention are still within the protection scope of the technical solution of the present invention.

Claims

1. A reconstruction method based on hourly output data of wind and solar power stations during the flood and dry periods of a reservoir, characterized in that: The following steps are involved: Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year are screened out, and the screened output data are subjected to abnormal detection and cleaning; Interpolate missing values ​​for output data that has undergone anomaly detection and cleaning; Extract k typical output intensity scenarios from the interpolated output data, and obtain the hourly output data of the wind and solar power stations under different output intensity scenarios corresponding to the reservoir's flood season and dry season; Based on the capacity and efficiency of each wind farm and photovoltaic power station in the reservoir section, the output weights of each wind farm and photovoltaic power station in the section are established; the hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed up through the output weights of each wind farm and photovoltaic power station in the section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the section.

2. According to claim 1, a reconstruction method based on hourly output data of wind and solar power stations during the flood and dry periods of a reservoir, characterized in that: Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year are screened out, and the screened output data are subjected to abnormal detection and cleaning processing, including: Based on the flood season and dry season of reservoirs in different sections, the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year are screened out; Perform dimensionality reduction processing on the screened output data; The output data that has completed dimensionality reduction processing is denoised and cleaned based on the DBSCAN clustering algorithm.

3. According to claim 1, a reconstruction method based on hourly output data of wind and solar power stations during the flood and dry periods of a reservoir, characterized in that: The dimensionality reduction processing of the filtered output data includes: The flood season and dry season of reservoirs in different sections are selected to select the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season in different months throughout the year to establish the standardized output sample X (x1, x2, x3...x 23 , x 24 ); Based on the daily power station output characteristics, the output curve skewness (S) and the daily output peak mean (E) of the standardized output sample X are extracted. peak ) is processed by dimensionality reduction to obtain the output feature matrix O i (E peak , S), the formula is as follows: Where: S represents the degree of left or right deviation of the power station daily output curve about the time point l within the day. If it is greater than 0, it means right deviation about point l; p is the set of possible values ​​of point l, and the value of l is the l within p that makes the S value of each power station in a certain section the smallest; x l+i is the power station output at time l+i; E peak Represents the average of the four points with the largest daily output of the power station; m i It is the ith largest daily output value of the power station.

4. A method for reconstructing hourly output data of wind and solar power stations based on the reservoir's wet and dry periods according to claim 3, characterized in that: The denoising and cleaning of the output data after dimensionality reduction processing based on the DBSCAN clustering algorithm includes: Set the DBSCAN parameter scanning radius (Eps) and the minimum number of included points (k), and increase the model parameters to calculate the output characteristic matrix O of the power station. i (E peak , S) for coordinate centrifugation, the formula is as follows: T=C×n×|S| q Where: T is the value of S after centrifugal deviation; C is 1 when S is greater than 0 and -1 when S is less than 0; n is the uniform scaling coefficient; q is the centrifugal coefficient, which is greater than 1; Centrifugal denoising is performed according to the proportion of noise data in the noise sequence to the total data, and the denoised feature matrix data sets D corresponding to the flood and dry periods of reservoirs in different sections under different noise rates are obtained. 0，x , x is the proportion of noise data to the total data.

5. A method for reconstructing hourly output data of wind and solar power stations based on the reservoir's wet and dry periods according to claim 4, characterized in that: The missing value interpolation of the output data after abnormal detection and cleaning includes: For the non-continuous missing data in the output data that has been processed by anomaly detection and cleaning, the mean of the hourly output data of the two days before and after is taken to interpolate all data sets; For the continuous missing data in the output data after anomaly detection and cleaning, multiple historical similar days are found according to the characteristic values ​​of the meteorological data on the missing day, and the average of the hourly output data of multiple historical similar days is taken to interpolate all data sets.

6. A method for reconstructing hourly output data of wind and solar power stations based on the reservoir's wet and dry periods according to claim 1, characterized in that: The interpolated output data is used to extract k typical output scenarios, and the hourly output data of the wind and solar power stations under different output scenarios corresponding to the reservoir's flood season and dry season are obtained, including: Select Euclidean distance as the sample point distance, determine the k value according to the different force types set, and extract k typical scenes of force output; For data sets D with different noise rates 1，x Perform Kmeans++ clustering respectively to generate corresponding different clustering result sets D 2，x ; For the clustering result set D 2，x Use the ratio F as an indicator and use the remapped original output feature matrix O when calculating the ratio F. i (E peak ,slop), the formula is as follows: Where: the ratio F is the ratio of the sum of squares of the between-class deviations to the sum of squares of the within-class deviations; p is the number of stations in the basin; N k and are the sample size and sample mean of the kth subclass, respectively. Represents X j The value of the i-th sample observation in the k-th subclass; the term (NK) / (K-1) is introduced to eliminate the influence of the sample size N and the number of clusters K on the results; Typical scenarios are extracted from the clustering results with a set confidence level to obtain the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of wind and solar power stations.

7. A method for reconstructing hourly output data of wind and solar power stations based on the reservoir's wet and dry periods according to claim 6, characterized in that: Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of each wind farm and photovoltaic power station in the cross-section are established; the hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed through the output weights of each wind farm and photovoltaic power station in the cross-section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the cross-section, including: Based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir, the output weights of the wind farm and photovoltaic power station in the cross-section are set. The formula is as follows: w lw =N lw *h lw Where: w lw is the weight of the lth wind power station; N lw is the installed capacity of the lth wind power station; h lw is the annual utilization hours of the lth wind power station; Where: w w is the sum of the weights of all wind power stations in the section; w lp =N lp *h lp Where: w lp is the weight of the lth photovoltaic power station; N lp is the installed capacity of the lth photovoltaic power station; h lp is the annual utilization hours of the lth photovoltaic power station; Where: w p is the sum of the weights of all PV power stations in the section; The hourly output of all wind power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season is summed up through the output weights of each wind farm and photovoltaic power station in the section to obtain the total hourly output data of wind and photovoltaic power stations in different output intensity scenarios in the section. The formula is as follows: Where: P wt is the equivalent output of the equivalent wind power station in the section at time t; P lw，t is the output of the lth wind power station at time t; is the output weight ratio of the lth wind power station; Where: P pt is the equivalent output of the equivalent photovoltaic power station in the section at time t; P lp，t is the output of the lth PV power station at time t; is the output weight ratio of the lth photovoltaic power station.

8. A dispatching device for a water-wind-solar hybrid power system based on cascade reservoirs, comprising: The data cleaning module is used to filter out the hourly output data of all wind and solar power stations in the section corresponding to the flood season and dry season throughout the year based on the flood season and dry season of reservoirs in different sections, and perform abnormal detection and cleaning on the filtered output data; Data interpolation module, used to interpolate missing values ​​of output data that has undergone anomaly detection and cleaning; The typical scenario extraction module is used to extract k typical scenarios of output intensity based on the interpolated output data, and obtain the hourly output data of the reservoir's flood season and dry season corresponding to different output intensity scenarios of the wind and solar power stations; The equivalent output calculation module is used to establish the output weights of each wind farm and photovoltaic power station in the cross-section based on the capacity and efficiency of each wind farm and photovoltaic power station in the cross-section reservoir; the hourly output data of the wind and photovoltaic power stations in different output intensity scenarios corresponding to the reservoir's flood season and dry season are summed up through the output weights of each wind farm and photovoltaic power station in the cross-section to obtain the total hourly output data of the wind and photovoltaic power stations in different output intensity scenarios in the cross-section.

9. A computer storage medium, wherein the computer readable storage medium stores a computer program, characterized in that: When the computer program is executed by the processor, the steps of the method for reconstructing the hourly output data of the wind and solar power station based on the wet and dry periods of the reservoir as described in any one of claims 1 to 7 are implemented.

10. An electronic device, characterized in that: It includes a memory and a processor: the memory is used to store computer-executable instructions, and the processor is used to execute the computer-executable instructions. When the computer-executable instructions are executed by the processor, the steps of the method for reconstructing the hourly output data of a wind and solar power station based on the flood and dry periods of a reservoir as described in any one of claims 1 to 7 are implemented.

Images