Photovoltaic data division method and device based on permutation entropy, equipment and medium

Abstract

The invention discloses a photovoltaic data division method and device based on permutation entropy, equipment and a medium, and relates to the field of power generation data processing, and the method comprises the steps: obtaining to-be-divided photovoltaic historical power generation power data and meteorological data; according to the photovoltaic historical power generation power data and the meteorological data, calculating the feature relevancy of each category of the meteorological data, and screening the categories of the meteorological data according to the feature relevancy to obtain a feature category sequence; according to the feature category sequence, calculating the permutation entropy of the target sequence of each day in the photovoltaic historical generation power data; wherein the target sequence is correspondingly obtained by the feature category sequence; and obtaining a clustering result of the permutation entropy based on a clustering algorithm according to the permutation entropy of the target sequence of each day, and dividing the photovoltaic historical generation power data according to the clustering result of the permutation entropy. According to the embodiment of the invention, the technical problem of insufficient division accuracy of the photovoltaic power generation data in the prior art can be solved.

Description

Technical Field

[0001] The present application relates to the field of power generation data processing, and in particular to a photovoltaic data partitioning method, device, equipment and medium based on permutation entropy. Background Art

[0002] This section is only used to introduce the technical background mastered by the applicant and does not represent that the following is prior art.

[0003] With the large-scale construction of distributed photovoltaics, the unstable characteristics of photovoltaic power generation have become more prominent. Therefore, it is necessary to predict the photovoltaic power generation to improve the utilization rate of photovoltaic power generation. Generally, meteorological changes have the most significant impact on the output power of photovoltaic power generation, and there are obvious differences in photovoltaic output power under different meteorological types. Therefore, when predicting photovoltaic power, it is necessary to distinguish different meteorological types to obtain more accurate results.

[0004] In the prior art, there are some methods for classifying or clustering photovoltaic power generation data to distinguish different meteorological types. However, among these methods, one is that there may be subjective judgment factors, such as the classification granularity is artificially defined too fine or too coarse; the second is that the classification or clustering accuracy is insufficient. For example, directly clustering photovoltaic data without pre-processing and post-processing leads to insufficient accuracy in meteorological type classification. These will affect the accuracy of photovoltaic power generation prediction based on the divided data to a certain extent. Therefore, how to improve the accuracy of photovoltaic power generation data classification is still a difficult problem that needs to be solved urgently in the prior art. Summary of the invention

[0005] The present application provides a photovoltaic data division method, device, equipment and medium based on permutation entropy to solve the technical problem of insufficient accuracy in the division of photovoltaic power generation data in the prior art.

[0006] According to a first aspect of the implementation of the present application, a photovoltaic data partitioning method based on permutation entropy is provided, comprising:

[0007] Obtain historical photovoltaic power generation data and meteorological data to be divided;

[0008] Calculating the feature correlation of each category of the meteorological data according to the historical photovoltaic power generation data and the meteorological data, and screening the category of the meteorological data according to the feature correlation to obtain a feature category sequence;

[0009] According to the feature category sequence, calculating the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data; wherein the target sequence is obtained by corresponding to the feature category sequence;

[0010] According to the permutation entropy of the target sequence of each day, based on a clustering algorithm, a clustering result of the permutation entropy is obtained, and the photovoltaic historical power generation data is divided according to the clustering result of the permutation entropy.

[0011] The present application first obtains the photovoltaic historical power generation data and meteorological data to be divided, and calculates the feature correlation of each category of the meteorological data, so that the feature correlation of each category of the meteorological data can be obtained, and then when screening according to the feature correlation, a feature category sequence with a higher degree of correlation with photovoltaic power generation is obtained; then the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data is calculated according to the feature category sequence, and the permutation entropy of the target sequence of each day can be calculated, and the objective change trend of the target sequence can be reflected thereby, avoiding subjective influence, and clustering is performed according to the permutation entropy of the target sequence of each day to obtain a clustering result, and a more objective clustering result can be obtained, thereby improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering result.

[0012] In certain embodiments of the present application, the feature correlation of each category of the meteorological data is calculated based on the photovoltaic historical power generation data and the meteorological data, and the category of the meteorological data is screened according to the feature correlation to obtain a feature category sequence, specifically including:

[0013] Aligning the photovoltaic historical power generation data and the meteorological data in time series, and calculating the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data;

[0014] According to a preset correlation threshold, based on the feature correlation, the categories of the meteorological data are screened to obtain a feature category sequence corresponding to the meteorological data.

[0015] The present application first aligns the photovoltaic historical power generation data and meteorological data in chronological order, and then calculates the feature correlation of each category of the meteorological data, so as to obtain the feature correlation of each category of the meteorological data, and then, when screening according to the feature correlation, obtains a feature category sequence with a higher degree of correlation with photovoltaic power generation, thereby improving the accuracy of the clustering results, and indirectly improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering results.

[0016] In certain embodiments of the present application, the calculating, based on the photovoltaic historical power generation data and the meteorological data, the characteristic correlation of each category of the meteorological data specifically includes:

[0017] The feature relevance is specifically:

[0018]

[0019] Among them, xi is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

[0020] In certain embodiments of the present application, the calculation of the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence specifically includes:

[0021] Selecting the target sequence of each day in the photovoltaic historical power generation data as the current target sequence in sequence, and calculating the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method, until the target sequence in the photovoltaic historical power generation data is completely selected;

[0022] The first calculation method is specifically:

[0023] Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components;

[0024] The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components;

[0025] A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

[0026] The present application first selects the current target sequence, reconstructs the phase space of the current target sequence and divides it into multiple first target components according to the row space, then sorts the elements in the multiple first target components to obtain multiple second target components, and obtains multiple sorting mappings according to the subscripts of the elements of the multiple second target components, thereby obtaining the permutation entropy of the current target sequence, and can calculate the permutation entropy of the target sequence for each day, and thereby reflect the objective change trend of the target sequence, avoiding subjective influence, thereby improving the clustering accuracy when clustering according to the permutation entropy, and indirectly improving the accuracy when dividing the photovoltaic historical power generation data according to the clustering results.

[0027] In certain embodiments of the present application, calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes:

[0028] The permutation entropy of the current target sequence is specifically:

[0029]

[0030] Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

[0031] In certain embodiments of the present application, the clustering result of the permutation entropy is obtained based on the permutation entropy of the target sequence of each day and based on a clustering algorithm, specifically including:

[0032] According to the permutation entropy of the target sequence of each day and the preset number of clustering clusters, based on the clustering algorithm, the clustering result of the permutation entropy is obtained.

[0033] The present application obtains a clustering result of the permutation entropy based on a clustering algorithm according to the permutation entropy of the target sequence of each day and the preset number of clustering clusters. By limiting the number of clustering clusters, the division result obtained according to the clustering result can be more suitable for different scenarios of photovoltaic power generation prediction.

[0034] According to a second aspect of the implementation of the present application, there is provided a photovoltaic data partitioning device based on permutation entropy, comprising a to-be-partitioned data acquisition module, a feature sequence acquisition module, a permutation entropy calculation module and a permutation entropy clustering module;

[0035] The to-be-divided data acquisition module is used to acquire the photovoltaic historical power generation data and meteorological data to be divided;

[0036] The feature sequence acquisition module is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, and to filter the category of the meteorological data according to the feature correlation to obtain a feature category sequence;

[0037] The permutation entropy calculation module is used to calculate the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence; wherein the target sequence is obtained by corresponding to the feature category sequence;

[0038] The permutation entropy clustering module is used to obtain the clustering result of the permutation entropy based on the permutation entropy of the target sequence of each day based on a clustering algorithm, and to divide the photovoltaic historical power generation data according to the clustering result of the permutation entropy.

[0039] In certain embodiments of the present application, the feature sequence acquisition module includes a feature correlation calculation unit and a feature sequence acquisition unit;

[0040] The feature correlation calculation unit is used to align the photovoltaic historical power generation data and the meteorological data in time series, and calculate the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data;

[0041] The feature sequence acquisition unit is used to screen the categories of the meteorological data based on the feature relevance according to a preset relevance threshold, so as to obtain a feature category sequence corresponding to the meteorological data.

[0042] In certain embodiments of the present application, the feature correlation calculation unit is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, specifically including:

[0043] The feature relevance is specifically:

[0044]

[0045] Among them, x i is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

[0046] In certain embodiments of the present application, the permutation entropy calculation module includes a permutation entropy calculation unit;

[0047] The permutation entropy calculation unit is used to sequentially select the target sequence of each day in the photovoltaic historical power generation data as the current target sequence, and calculate the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method until the target sequence in the photovoltaic historical power generation data is completely selected;

[0048] The first calculation method is specifically:

[0049] Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components;

[0050] The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components;

[0051] A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

[0052] In certain embodiments of the present application, calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes:

[0053] The permutation entropy of the current target sequence is specifically:

[0054]

[0055] Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

[0056] In certain embodiments of the present application, the permutation entropy clustering module includes a permutation entropy clustering unit;

[0057] The permutation entropy clustering unit is used to obtain the clustering result of the permutation entropy based on the clustering algorithm according to the permutation entropy of the target sequence of each day and the preset number of clustering clusters.

[0058] The present application first obtains the photovoltaic historical power generation data and meteorological data to be divided, and calculates the feature correlation of each category of the meteorological data, so that the feature correlation of each category of the meteorological data can be obtained, and then when screening according to the feature correlation, a feature category sequence with a higher degree of correlation with photovoltaic power generation is obtained; then the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data is calculated according to the feature category sequence, and the permutation entropy of the target sequence of each day can be calculated, and the objective change trend of the target sequence can be reflected thereby, avoiding subjective influence, and clustering is performed according to the permutation entropy of the target sequence of each day to obtain a clustering result, and a more objective clustering result can be obtained, thereby improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering result. BRIEF DESCRIPTION OF THE DRAWINGS

[0059] Figure 1 : A schematic flow chart of a photovoltaic data partitioning method based on permutation entropy shown in certain embodiments of the present application;

[0060] Figure 2 : A module structure diagram of a photovoltaic data partitioning device based on permutation entropy shown in certain embodiments of the present application. DETAILED DESCRIPTION

[0061] The embodiments of the present application are described in detail below, and examples of the embodiments are shown in the accompanying drawings, wherein the same or similar reference numerals represent the same or similar elements or elements having the same or similar functions from beginning to end. The embodiments described below in conjunction with the accompanying drawings are exemplary and are only used to explain some embodiments of the present application, and should not be construed as limitations on the embodiments of the present application. Based on the embodiments shown in the present application, all other embodiments obtained by ordinary technicians in the field without making creative work are within the scope of protection of the present application.

[0062] In the description of this application, it should be understood that the terms "first" and "second" are used for descriptive purposes only and should not be understood as indicating or implying relative importance or implicitly indicating the number of technical features indicated. Therefore, a feature defined as "first" or "second" may explicitly or implicitly include at least one of the features. In the description of this application, unless otherwise clearly and specifically defined, "multiple" and "several" mean two or more.

[0063] When predicting photovoltaic power, the impact of meteorological changes is the most significant, and there are obvious differences in photovoltaic output power under different meteorological types. Therefore, when predicting photovoltaic power, it is necessary to distinguish different meteorological types to obtain more accurate results. However, the methods of distinguishing different meteorological types in the prior art, including classification or clustering, have some shortcomings: first, there may be subjective judgment factors, such as the classification granularity is artificially defined too fine or too coarse, and second, the classification or clustering division accuracy is insufficient. For example, photovoltaic data is directly clustered, lacking pre-processing and post-processing, resulting in insufficient accuracy in meteorological type division. These will affect the accuracy of photovoltaic power generation prediction based on the divided data to a certain extent. Therefore, how to improve the accuracy of the division of photovoltaic power generation data is still a difficult problem that needs to be solved urgently in the prior art.

[0064] Based on the above technical background, please refer to Figure 1 The embodiment of the present application provides a photovoltaic data partitioning method based on permutation entropy, including steps S101 to S104, each of which is specifically as follows:

[0065] Step S101: Acquire historical photovoltaic power generation data and meteorological data to be divided.

[0066] In certain embodiments of the present application, the obtaining of the photovoltaic historical power generation data and meteorological data to be divided specifically includes:

[0067] Obtain initial photovoltaic historical power generation data and initial meteorological data;

[0068] The initial photovoltaic historical power generation data is filtered according to a preset sunshine period to obtain first photovoltaic historical power generation data; wherein the first photovoltaic historical power generation data is data of the initial photovoltaic historical power generation data that only includes sunshine periods;

[0069] Based on the error status of each element in the first photovoltaic historical power generation data and the initial meteorological data, the first photovoltaic historical power generation data and the initial meteorological data are corrected to obtain second photovoltaic historical power generation data and meteorological data;

[0070] The second photovoltaic historical power generation data is normalized to obtain photovoltaic historical power generation data.

[0071] Consider filtering the historical photovoltaic power generation data according to the preset sunshine period because the photovoltaic power generation is zero at night. By filtering and deleting the historical photovoltaic power generation data at night, the data range can be narrowed and the data information entropy can be improved, thereby improving the subsequent calculation feature correlation and clustering accuracy.

[0072] In certain embodiments of the present application, the categories of meteorological data include, but are not limited to, solar irradiance, ambient temperature, relative humidity, air pressure, wind speed, wind direction and photovoltaic module surface temperature.

[0073] Step S102: Calculate the feature correlation of each category of the meteorological data based on the historical photovoltaic power generation data and the meteorological data, and filter the categories of the meteorological data based on the feature correlation to obtain a feature category sequence.

[0074] In certain embodiments of the present application, the feature correlation of each category of the meteorological data is calculated based on the photovoltaic historical power generation data and the meteorological data, and the category of the meteorological data is screened according to the feature correlation to obtain a feature category sequence, which specifically includes:

[0075] Aligning the photovoltaic historical power generation data and the meteorological data in time series, and calculating the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data;

[0076] According to a preset correlation threshold, based on the feature correlation, the categories of the meteorological data are screened to obtain a feature category sequence corresponding to the meteorological data.

[0077] In certain embodiments of the present application, the calculating, based on the photovoltaic historical power generation data and the meteorological data, the characteristic correlation of each category of the meteorological data specifically includes:

[0078] The feature relevance is specifically:

[0079]

[0080] Among them, x i is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

[0081] The present application first aligns the photovoltaic historical power generation data and meteorological data in chronological order, and then calculates the feature correlation of each category of the meteorological data, so as to obtain the feature correlation of each category of the meteorological data, and then, when screening according to the feature correlation, obtains a feature category sequence with a higher degree of correlation with photovoltaic power generation, thereby improving the accuracy of the clustering results, and indirectly improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering results.

[0082] Step S103: Calculate the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence; wherein the target sequence is obtained by corresponding to the feature category sequence.

[0083] In certain embodiments of the present application, the calculation of the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence specifically includes:

[0084] Selecting the target sequence of each day in the photovoltaic historical power generation data as the current target sequence in sequence, and calculating the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method, until the target sequence in the photovoltaic historical power generation data is completely selected;

[0085] The first calculation method is specifically:

[0086] Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components;

[0087] The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components;

[0088] A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

[0089] Specifically, the current target sequence is recorded as y, and its length is l. Then the first target phase space obtained after the phase space reconstruction of the current target sequence y is expressed as follows:

[0090]

[0091] Where m is the embedding dimension, is the delay time, is the sequence length.

[0092] Specifically, the first target phase space is divided into a plurality of first target components Y(i) according to the row space and is expressed as:

[0093] Wherein, k is the number of the second target components.

[0094] Specifically, the elements in the plurality of first target components are sorted, and the plurality of second target components are obtained by sorting in a non-decreasing order according to the value of each element during the sorting, and the sorting of each second target component is expressed as:

[0095] Among them, when there are two equal values, according to the j corresponding to the two values i The subscript i of is sorted in non-decreasing order.

[0096] Through the above sorting operation, each second target component (first target component) Y(i) can be mapped to the same symbol sequence S(i)=(j1, j2, ..., j m ).

[0097] Specifically, after obtaining a plurality of second target components, the plurality of second target components (the plurality of first target components) correspond to the symbol sequence S(i)=(j1, j2, ..., j m ) is also determined. Therefore, when calculating the permutation entropy of the current target sequence according to the multiple sorting maps, it is only necessary to calculate the symbol sequence S(i)=(j1, j2, ..., j in the k groups of second target components. m ) is the probability of occurrence of a sequence composed of different orderings of the elements in the time series, and the permutation entropy of the current target sequence can be obtained. Generally speaking, the more regular the time series, the smaller the corresponding permutation entropy, and the more chaotic the data changes, the larger the corresponding permutation entropy.

[0098] In certain embodiments of the present application, calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes:

[0099] The permutation entropy of the current target sequence is specifically:

[0100]

[0101] Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

[0102] The present application first selects the current target sequence, reconstructs the phase space of the current target sequence and divides it into multiple first target components according to the row space, then sorts the elements in the multiple first target components to obtain multiple second target components, and obtains multiple sorting mappings according to the subscripts of the elements of the multiple second target components, thereby obtaining the permutation entropy of the current target sequence, and can calculate the permutation entropy of the target sequence for each day, and thereby reflect the objective change trend of the target sequence, avoiding subjective influence, thereby improving the clustering accuracy when clustering according to the permutation entropy, and indirectly improving the accuracy when dividing the photovoltaic historical power generation data according to the clustering results.

[0103] Step S104: obtaining a clustering result of the permutation entropy according to the permutation entropy of the target sequence of each day based on a clustering algorithm, and dividing the photovoltaic historical power generation data according to the clustering result of the permutation entropy.

[0104] In certain embodiments of the present application, the clustering result of the permutation entropy is obtained based on the permutation entropy of the target sequence of each day and based on a clustering algorithm, specifically including:

[0105] According to the permutation entropy of the target sequence of each day and the preset number of clustering clusters, based on the clustering algorithm, the clustering result of the permutation entropy is obtained.

[0106] In certain embodiments of the present application, the clustering algorithm includes but is not limited to a K-means clustering algorithm, a mean shift clustering algorithm, a DBSCAN clustering algorithm, and a hierarchical clustering algorithm, and a preferred implementation is the K-means clustering algorithm.

[0107] The present application obtains a clustering result of the permutation entropy based on a clustering algorithm according to the permutation entropy of the target sequence of each day and the preset number of clustering clusters. By limiting the number of clustering clusters, the division result obtained according to the clustering result can be more suitable for different scenarios of photovoltaic power generation prediction.

[0108] Compared with the prior art, the present application first obtains the photovoltaic historical power generation data and meteorological data to be divided, and calculates the feature correlation of each category of the meteorological data, so as to obtain the feature correlation of each category of the meteorological data, and then obtains a feature category sequence with a higher degree of correlation with photovoltaic power generation when screening according to the feature correlation; then calculates the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence, and can calculate the permutation entropy of the target sequence of each day, and thereby reflect the objective change trend of the target sequence, avoiding subjective influence, and clustering is performed according to the permutation entropy of the target sequence of each day to obtain a clustering result, and can obtain a more objective clustering result, thereby improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering result.

[0109] Corresponding to the previous method, see Figure 2 , the embodiment of the present application provides a photovoltaic data partitioning device based on permutation entropy, including a to-be-partitioned data acquisition module 210, a feature sequence acquisition module 220, a permutation entropy calculation module 230 and a permutation entropy clustering module 240;

[0110] The to-be-divided data acquisition module 210 is used to acquire the photovoltaic historical power generation data and meteorological data to be divided;

[0111] The feature sequence acquisition module 220 is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, and filter the category of the meteorological data according to the feature correlation to obtain a feature category sequence;

[0112] The permutation entropy calculation module 230 is used to calculate the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence; wherein the target sequence is obtained by corresponding to the feature category sequence;

[0113] The permutation entropy clustering module 240 is used to obtain the clustering result of the permutation entropy based on the permutation entropy of the target sequence of each day based on a clustering algorithm, and to divide the photovoltaic historical power generation data according to the clustering result of the permutation entropy.

[0114] In certain embodiments of the present application, the feature sequence acquisition module 220 includes a feature correlation calculation unit and a feature sequence acquisition unit;

[0115] The feature correlation calculation unit is used to align the photovoltaic historical power generation data and the meteorological data in time series, and calculate the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data;

[0116] The feature sequence acquisition unit is used to screen the categories of the meteorological data based on the feature relevance according to a preset relevance threshold, so as to obtain a feature category sequence corresponding to the meteorological data.

[0117] In certain embodiments of the present application, the feature correlation calculation unit is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, specifically including:

[0118] The feature relevance is specifically:

[0119]

[0120] Among them, x i is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

[0121] In certain embodiments of the present application, the permutation entropy calculation module 230 includes a permutation entropy calculation unit;

[0122] The permutation entropy calculation unit is used to sequentially select the target sequence of each day in the photovoltaic historical power generation data as the current target sequence, and calculate the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method until the target sequence in the photovoltaic historical power generation data is completely selected;

[0123] The first calculation method is specifically:

[0124] Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components;

[0125] The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components;

[0126] A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

[0127] In certain embodiments of the present application, calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes:

[0128] The permutation entropy of the current target sequence is specifically:

[0129]

[0130] Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

[0131] In certain embodiments of the present application, the permutation entropy clustering module 240 includes a permutation entropy clustering unit;

[0132] The permutation entropy clustering unit is used to obtain the clustering result of the permutation entropy based on the clustering algorithm according to the permutation entropy of the target sequence of each day and the preset number of clustering clusters.

[0133] The present application first obtains the photovoltaic historical power generation data and meteorological data to be divided, and calculates the feature correlation of each category of the meteorological data, so that the feature correlation of each category of the meteorological data can be obtained, and then when screening according to the feature correlation, a feature category sequence with a higher degree of correlation with photovoltaic power generation is obtained; then the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data is calculated according to the feature category sequence, and the permutation entropy of the target sequence of each day can be calculated, and the objective change trend of the target sequence can be reflected thereby, avoiding subjective influence, and clustering is performed according to the permutation entropy of the target sequence of each day to obtain a clustering result, and a more objective clustering result can be obtained, thereby improving the accuracy of dividing the photovoltaic historical power generation data according to the clustering result.

[0134] It should be understood that the device provided in the embodiment of the present application corresponds to the aforementioned method, and a photovoltaic data division device based on permutation entropy provided in the embodiment of the present application can implement a photovoltaic data division method based on permutation entropy provided in any embodiment of the present application.

[0135] Adaptively, the embodiments of the present application also provide a computer device and a computer-readable storage medium.

[0136] The computer device comprises: a processor, a memory, and a computer program stored in the memory and configured to be executed by the processor;

[0137] Wherein, when the processor executes the computer program, a photovoltaic data partitioning method based on permutation entropy of the present application is implemented.

[0138] The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute a photovoltaic data partitioning method based on permutation entropy of the present application.

[0139] The above are some implementation methods of the present application, and the purpose, technical scheme and beneficial effects of the present application are further described in detail. It should be clear that the above-mentioned partial implementation methods of the present application cannot be understood as limiting the present application. It is particularly pointed out that for those skilled in the art, any changes, modifications, equivalent substitutions and variations made within the spirit and principles of the present application should be included in the protection scope of the present application.

Claims

1. A photovoltaic data partitioning method based on permutation entropy, characterized in that: include: Obtain historical photovoltaic power generation data and meteorological data to be divided; Calculating the feature correlation of each category of the meteorological data according to the historical photovoltaic power generation data and the meteorological data, and screening the category of the meteorological data according to the feature correlation to obtain a feature category sequence; According to the feature category sequence, calculating the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data; wherein the target sequence is obtained by corresponding to the feature category sequence; According to the permutation entropy of the target sequence of each day, based on a clustering algorithm, a clustering result of the permutation entropy is obtained, and the photovoltaic historical power generation data is divided according to the clustering result of the permutation entropy.

2. The photovoltaic data partitioning method based on permutation entropy according to claim 1, characterized in that: The step of calculating the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, and screening the category of the meteorological data according to the feature correlation to obtain a feature category sequence specifically includes: Aligning the photovoltaic historical power generation data and the meteorological data in time series, and calculating the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data; According to a preset correlation threshold, based on the feature correlation, the categories of the meteorological data are screened to obtain a feature category sequence corresponding to the meteorological data.

3. The photovoltaic data partitioning method based on permutation entropy according to claim 2, characterized in that: The calculating, based on the photovoltaic historical power generation data and the meteorological data, the feature correlation of each category of the meteorological data specifically includes: The feature relevance is specifically: Among them, x i is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

4. The photovoltaic data partitioning method based on permutation entropy according to claim 1, characterized in that: The step of calculating the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence specifically includes: Selecting the target sequence of each day in the photovoltaic historical power generation data as the current target sequence in sequence, and calculating the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method, until the target sequence in the photovoltaic historical power generation data is completely selected; The first calculation method is specifically: Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components; The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components; A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

5. The photovoltaic data partitioning method based on permutation entropy according to claim 4, characterized in that: The calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes: The permutation entropy of the current target sequence is specifically: Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

6. The photovoltaic data partitioning method based on permutation entropy according to claim 1, characterized in that: The clustering result of the permutation entropy is obtained based on the permutation entropy of the target sequence of each day and based on the clustering algorithm, specifically including: According to the permutation entropy of the target sequence of each day and the preset number of clustering clusters, based on the clustering algorithm, the clustering result of the permutation entropy is obtained.

7. A photovoltaic data partitioning device based on permutation entropy, characterized in that: It includes a module for obtaining data to be divided, a module for obtaining a feature sequence, a module for calculating permutation entropy and a module for clustering permutation entropy; The to-be-divided data acquisition module is used to acquire the photovoltaic historical power generation data and meteorological data to be divided; The feature sequence acquisition module is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, and to filter the category of the meteorological data according to the feature correlation to obtain a feature category sequence; The permutation entropy calculation module is used to calculate the permutation entropy of the target sequence of each day in the photovoltaic historical power generation data according to the feature category sequence; wherein the target sequence is obtained by corresponding to the feature category sequence; The permutation entropy clustering module is used to obtain the clustering result of the permutation entropy based on the permutation entropy of the target sequence of each day based on a clustering algorithm, and to divide the photovoltaic historical power generation data according to the clustering result of the permutation entropy.

8. The photovoltaic data partitioning device based on permutation entropy according to claim 7, characterized in that: The feature sequence acquisition module includes a feature correlation calculation unit and a feature sequence acquisition unit; The feature correlation calculation unit is used to align the photovoltaic historical power generation data and the meteorological data in time series, and calculate the feature correlation of each category of the meteorological data based on the photovoltaic historical power generation data and the meteorological data; The feature sequence acquisition unit is used to screen the categories of the meteorological data based on the feature relevance according to a preset relevance threshold, so as to obtain a feature category sequence corresponding to the meteorological data.

9. The photovoltaic data partitioning device based on permutation entropy according to claim 8, characterized in that: The feature correlation calculation unit is used to calculate the feature correlation of each category of the meteorological data according to the photovoltaic historical power generation data and the meteorological data, specifically including: The feature relevance is specifically: Among them, x i is the photovoltaic power generation power on the i-th day, is the average photovoltaic power generation power, is the value of the ω-th type of meteorological data on the i-th day, is the mean of the ωth type of meteorological data, and n is the number of statistical days.

10. The photovoltaic data partitioning device based on permutation entropy according to claim 7, characterized in that: The permutation entropy calculation module includes a permutation entropy calculation unit; The permutation entropy calculation unit is used to sequentially select the target sequence of each day in the photovoltaic historical power generation data as the current target sequence, and calculate the permutation entropy of the current target sequence based on the feature category sequence according to the preset first calculation method until the target sequence in the photovoltaic historical power generation data is completely selected; The first calculation method is specifically: Reconstructing the current target sequence in phase space to obtain a first target phase space, and dividing the first target phase space according to row space to obtain a plurality of first target components; The elements in the plurality of first target components are sorted respectively, and the elements are sorted in a non-decreasing order according to the value of each element to obtain a plurality of second target components; A plurality of sorting mappings are obtained according to the subscripts of the elements of the plurality of second target components, and the permutation entropy of the current target sequence is calculated according to the plurality of sorting mappings.

11. The photovoltaic data partitioning device based on permutation entropy according to claim 10, characterized in that: The calculating the permutation entropy of the current target sequence according to the multiple sorting mappings specifically includes: The permutation entropy of the current target sequence is specifically: Among them, H pe is the permutation entropy of the current target sequence, P i is the probability that the sorting mapping corresponding to the i-th second target component appears in the multiple sorting mappings, and k is the number of second target components.

12. The photovoltaic data partitioning device based on permutation entropy according to claim 7, characterized in that: The permutation entropy clustering module includes a permutation entropy clustering unit; The permutation entropy clustering unit is used to obtain the clustering result of the permutation entropy based on the clustering algorithm according to the permutation entropy of the target sequence of each day and the preset number of clustering clusters.

13. A computer device, characterized in that: include: processor; Memory; a computer program stored in the memory and configured to be executed by the processor; Wherein, when the processor executes the computer program, a photovoltaic data partitioning method based on permutation entropy as described in any one of claims 1 to 6 is implemented.

14. A computer-readable storage medium, characterized in that: The computer-readable storage medium stores a plurality of instructions, and the instructions are suitable for being loaded by a processor to execute a photovoltaic data partitioning method based on permutation entropy as described in any one of claims 1 to 6.

Images