A method and system for predicting new energy power generation based on meteorological elements

Abstract

The present invention provides a new energy power generation prediction method and system based on meteorological elements, belonging to the field of new energy prediction; the problem of low accuracy of new energy power generation prediction results is solved; specifically as follows: historical parameter data is acquired; principal component analysis is performed on the historical parameter data to obtain screened historical parameters; historical wind power generation data, historical photovoltaic power generation data and screened historical parameter data are acquired; ultra-short-term wind prediction model, ultra-short-term photovoltaic prediction model, short-term wind prediction model, short-term photovoltaic prediction model, medium- and long-term wind prediction model and medium- and long-term photovoltaic prediction model are acquired; prediction parameter data is acquired; prediction data is acquired; prediction data is corrected to obtain corrected wind power generation prediction data and corrected photovoltaic power generation prediction data; the present invention improves the accuracy of the new energy power generation prediction method by extracting main characteristic parameters and verifying model data.

Description

Technical Field

[0001] The present invention discloses a method and system for predicting renewable energy power generation based on meteorological elements, and relates to the field of renewable energy power generation prediction. Background Art

[0002] The existing methods or systems for forecasting renewable energy power generation based on meteorological factors have the following deficiencies:

[0003] 1. Inaccurate feature selection: There are different prediction methods based on different classification standards. There are multi-feature high-dimensional data according to the spatial scale differences of the prediction, which cannot be clearly expressed and it is difficult to extract effective features, resulting in unsatisfactory prediction accuracy;

[0004] 2. Single data prediction method: A single prediction method is dependent on data: Some methods are highly dependent on specific types of data, the data quality is degraded or incomplete, and a single method cannot take advantage of the integration advantages of multiple methods. Summary of the invention

[0005] In view of the shortcomings of the prior art, the purpose of the present invention is to provide a new energy power generation power prediction method based on meteorological elements, aiming to solve the problem of accuracy of new energy power generation power prediction results.

[0006] In order to achieve the above object, the present invention is implemented by the following technical solution: a new energy power generation power prediction method based on meteorological elements, the prediction method comprises:

[0007] Step S1: Acquire historical wind speed data, historical temperature data, historical humidity data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, and historical photovoltaic module data to obtain historical parameter data;

[0008] Step S2: Perform principal component analysis on the historical parameter data to obtain the screened historical parameters;

[0009] Step S3: Acquire historical wind power generation data and historical photovoltaic power generation data;

[0010] Find the data corresponding to the filtered historical parameters in the historical parameter data, extract the corresponding data, and obtain the filtered historical parameter data;

[0011] The historical parameter data is divided into time and data proportions, and the historical wind power generation data and the historical photovoltaic power generation data are combined for calculation to obtain an ultra-short-term wind prediction model, an ultra-short-term photovoltaic prediction model, a short-term wind prediction model, a short-term photovoltaic prediction model, a medium- and long-term wind prediction model, and a medium- and long-term photovoltaic prediction model;

[0012] Step S4: acquiring wind speed, wind direction, humidity, temperature and solar radiation data, obtaining wind speed data, wind direction data, humidity data, temperature data and solar radiation data, and performing calculations to obtain prediction parameter data;

[0013] Step S5: acquiring ultra-short-term wind power forecast data, ultra-short-term photovoltaic forecast data, short-term wind power forecast data, short-term photovoltaic forecast data, medium- and long-term wind power forecast data, and medium- and long-term photovoltaic forecast data according to the forecast parameter data and the forecast model;

[0014] The corrected wind power generation power prediction data and the corrected photovoltaic power generation power prediction data are acquired to obtain corrected prediction data.

[0015] Furthermore, the workflow of step S1 is as follows:

[0016] Read the historical wind speed, temperature, humidity and wind direction records to obtain historical wind direction data, historical temperature data, historical humidity data and historical wind speed data HU;

[0017] Read the historical wind speed data to obtain the historical wind speed data volume Hun and the historical wind speed data variable HUi;

[0018] Calculate the historical wind speed data volume Hun and the historical wind speed data variable HUi to obtain the historical wind speed mean

[0019]

[0020] Historical wind speed average The historical wind speed data volume Hun and the historical wind speed data variable HUi are calculated to obtain the historical wind speed standard deviation HUσ;

[0021]

[0022] The historical wind speed standard deviation and historical wind speed mean are calculated to obtain the historical turbulence data HA;

[0023]

[0024] Read the historical wind speed and get the historical wind speed HUih corresponding to the height h;

[0025] The historical wind speed HUih1 corresponding to the height h1 and the historical wind speed HUih2 corresponding to the height h2 are calculated to obtain the historical wind shear index data HB;

[0026]

[0027] Acquire solar radiation and obtain historical solar radiation data;

[0028] Acquire the historical azimuth angle data and historical component efficiency data of the new energy equipment to obtain the historical photovoltaic component data;

[0029] The historical wind speed data, historical temperature data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, historical humidity data, historical photovoltaic module azimuth data and historical module efficiency data are recorded as historical parameter data HK, and the process goes to step S2.

[0030] Furthermore, the workflow of step S2 is as follows:

[0031] Read the historical parameter data HK to obtain the historical parameter data variable HKmi and the historical parameter data quantity HKn;

[0032] The historical parameter data variables are expressed in the form of: {HK1i, HK2i, HK3i, HK4i, HK5i, HK6i, HK7i, HK8i, HK9i};

[0033] Calculate the historical parameter data variable HKmi and the historical parameter data volume HKn to obtain the historical parameter mean

[0034]

[0035] Historical parameter mean The manifestation is:

[0036] The historical parameter mean The historical parameter data variable HKmi and the historical parameter data quantity HKn are calculated to obtain the historical parameter standard deviation HKmσ;

[0037]

[0038] The historical parameter standard deviation HKmσ is expressed in the form of: {HK1σ, HK2σ, HK3σ, HK4σ, HK5σ, HK6σ, HK7σ, HK8σ, HK9σ};

[0039] The standard deviation HKmσ of historical parameters and the mean value of historical parameters Calculate with historical parameter data variable HKmi to obtain standardized historical parameter HZmi;

[0040]

[0041] The standardized historical parameter HZmi is expressed as: {HZ1i, HZ2i, HZ3i, HZ4i, HZ5i, HZ6i, HZ7i, HZ8i, HZ9i};

[0042] The standardized historical parameter HZmi, the historical parameter standard deviation HKmσ and the historical parameter data volume HKmn are calculated to obtain the covariance matrix C;

[0043]

[0044]

[0045] The covariance matrix C is expressed as:

[0046]

[0047] Calculate the covariance matrix C to obtain the eigenvalue Hλm and eigenvector Hαm;

[0048] C*Hαm=Hλm*Hαm;

[0049] The eigenvalue Hλm is expressed as: {Hλ1, Hλ2, Hλ3, Hλ4, Hλ5, Hλ6, Hλ7, Hλ8, Hλ9};

[0050] Screen the characteristic value Hλm to obtain the screening history parameter;

[0051] The screening steps are as follows:

[0052] Sort the eigenvalues ​​Hλm in order of size to obtain a set of sequential eigenvalues;

[0053] The historical parameters corresponding to the first five characteristic values ​​of the sequential characteristic value set are read to obtain the screening historical parameters, and then the process goes to step S3.

[0054] Furthermore, the workflow of step S3 is as follows:

[0055] Obtain historical wind power generation data and historical photovoltaic power generation data;

[0056] The screening historical parameters include: the first historical parameter, the second historical parameter, the third historical parameter, the fourth historical parameter, and the fifth historical parameter;

[0057] Find the data corresponding to the filtered historical parameters in the historical parameter data, extract the corresponding data, and obtain the filtered historical parameter data;

[0058] According to the time period, the screened historical parameter data is divided into ultra-short-term historical parameter data, short-term historical parameter data and medium- and long-term historical parameter data;

[0059] The time period division rules are as follows:

[0060] The screening historical parameter data within four hours is recorded as ultra-short-term historical parameter data;

[0061] The screening historical parameter data within 72 hours shall be recorded as short-term historical parameter data;

[0062] The historical parameter data of screening within one month is recorded as medium- and long-term parameter data;

[0063] The ultra-short-term historical parameter data, the short-term historical parameter data and the medium- and long-term parameter data are divided according to the data ratio to obtain an ultra-short-term historical data parameter training set, an ultra-short-term historical parameter data test set, a short-term historical data parameter training set, a short-term historical parameter data test set, a medium- and long-term historical data parameter training set and a medium- and long-term historical parameter data test set;

[0064] Using TensorFlow's tf.keras, we calculated the ultra-short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and ultra-short-term historical parameter data test set to obtain the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic prediction model.

[0065] Using TensorFlow's tf.keras, we calculated the short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and short-term historical parameter data test set to obtain the short-term wind power prediction model and the short-term photovoltaic prediction model.

[0066] Using TensorFlow's tf.keras, the medium- and long-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and medium- and long-term historical parameter data test set are calculated to obtain a medium- and long-term wind prediction model and a medium- and long-term photovoltaic prediction model. The ultra-short-term wind prediction model, ultra-short-term photovoltaic prediction model, short-term wind prediction model, short-term photovoltaic prediction model, medium- and long-term wind prediction model, and medium- and long-term photovoltaic prediction model are defined as prediction models, and step S4 is entered.

[0067] Furthermore, the workflow of step S4 is as follows:

[0068] Acquire the first parameter to obtain first parameter data HP1;

[0069] Acquire the second parameter to obtain second parameter data HP2;

[0070] Acquire the third parameter to obtain third parameter data HP3;

[0071] Acquire the fourth parameter to obtain fourth parameter data HP4;

[0072] Acquire the fifth parameter to obtain fifth parameter data HP5;

[0073] Data processing is performed on the first parameter data HP1, the second parameter data HP2, the third parameter data HP3, the fourth parameter data HP4 and the fifth parameter data HP5.

[0074] Furthermore, the workflow of data processing is as follows:

[0075] Read the first parameter data HP1 to obtain the first parameter data quantity HPn and the first parameter data variable HP1j;

[0076] Read the second parameter data HP2 to obtain the second parameter data quantity HPn and the second parameter data variable HP2j;

[0077] Read the third parameter data HP3 to obtain the third parameter data quantity HPn and the third parameter data variable HP3j;

[0078] The fourth parameter data HP4 is read to obtain the fourth parameter data quantity HPn and the third parameter data variable HP4j;

[0079] Read the fifth parameter data HP5 to obtain the fifth parameter data amount HPn and the fifth parameter data variable HP5j;

[0080] Calculate the first parameter data volume HPn and the first parameter data variable HP1j to obtain the first parameter mean

[0081]

[0082] Calculate the second parameter data volume HPn and the second parameter data variable HP2j to obtain the second parameter mean

[0083]

[0084] Calculate the third parameter data volume HPn and the third parameter data variable HP3j to obtain the third parameter mean

[0085]

[0086] Calculate the fourth parameter data volume HPn and the fourth parameter data variable HP4j to obtain the fourth parameter mean

[0087]

[0088] Calculate the fifth parameter data volume HPn and the fifth parameter data variable HP5j to obtain the fifth parameter mean

[0089]

[0090] The mean of the first parameter The first parameter data quantity HPn and the first parameter data variable HP1j are calculated to obtain the first parameter standard deviation HP1σ;

[0091]

[0092] The mean of the second parameter The second parameter data quantity HPn and the second parameter data variable HP2j are calculated to obtain the second parameter standard deviation HP2σ;

[0093]

[0094] The mean of the third parameter The third parameter data quantity HPn and the third parameter data variable HP3j are calculated to obtain the third parameter standard deviation HP3σ;

[0095]

[0096] The mean of the fourth parameter The fourth parameter data quantity HPn and the fourth parameter data variable HP4j are calculated to obtain the fourth parameter standard deviation HP4σ;

[0097]

[0098] The mean of the fifth parameter The fifth parameter data quantity HPn and the fifth parameter data variable HP5j are calculated to obtain the fifth parameter standard deviation HP5σ;

[0099]

[0100] For the first parameter standard deviation HP1σ, the first parameter mean Calculate with the first parameter data variable HP1j to obtain the first parameter standard value HZ1j;

[0101]

[0102] The second parameter standard deviation HP2σ, the second parameter mean Calculate with the second parameter data variable HP2j to obtain the second parameter standard value HZ2j;

[0103]

[0104] The third parameter standard deviation HP3σ, the third parameter mean Calculate with the third parameter data variable HP3j to obtain the third parameter standard value HZ3j;

[0105]

[0106] The fourth parameter standard deviation HP4σ, the fourth parameter mean Calculate with the fourth parameter data variable HP4j to obtain the fourth parameter standard value HZ4j;

[0107]

[0108] The fifth parameter standard deviation HP5σ, the fifth parameter mean Calculate with the fifth parameter data variable HP5j to obtain the fifth parameter standard value HZ5j;

[0109]

[0110] The first parameter standard value HZ1j, the second parameter standard value HZ2j, the third parameter standard value HZ3j, the fourth parameter standard value HZ4j and the fifth parameter standard value HZ5j are recorded as prediction parameter data.

[0111] Furthermore, the workflow of step S5 is as follows:

[0112] The ultra-short-term wind power prediction data F1 and the ultra-short-term photovoltaic power prediction data V1 are obtained by combining the prediction parameter data with the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic power prediction model;

[0113] By using the prediction parameter data combined with the short-term wind prediction model and the short-term photovoltaic prediction model, the short-term wind prediction data F2 and the short-term photovoltaic prediction data V2 are obtained;

[0114] The medium- and long-term wind power forecast data F3 and the medium- and long-term photovoltaic forecast data V3 are obtained by combining the forecast parameter data with the medium- and long-term wind power forecast model and the medium- and long-term photovoltaic forecast model;

[0115] The weights of the ultra-short-term wind prediction model, the short-term wind prediction model and the medium- and long-term wind prediction model are set to obtain the ultra-short-term wind prediction model weight A1, the short-term wind prediction model weight A2 and the medium- and long-term wind prediction model weight A3;

[0116] The weights of the ultra-short-term photovoltaic prediction model, the short-term photovoltaic prediction model and the medium- and long-term photovoltaic prediction model are set to obtain the weight B1 of the ultra-short-term photovoltaic prediction model, the weight B2 of the short-term photovoltaic prediction model and the weight B3 of the medium- and long-term photovoltaic prediction model;

[0117] Calculate the ultra-short-term wind forecast data F1, the short-term wind forecast data F2, the medium- and long-term wind forecast data F3, the ultra-short-term wind forecast model weight A1, the short-term wind forecast model weight A2 and the medium- and long-term wind forecast model weight A3 to obtain the modified wind power generation forecast data F;

[0118]

[0119] The ultra-short-term photovoltaic prediction data V1, the short-term photovoltaic prediction data V2, the medium- and long-term photovoltaic prediction data V3, the ultra-short-term photovoltaic prediction model weight B1, the short-term photovoltaic prediction model weight B2 and the medium- and long-term photovoltaic prediction model weight B3 are calculated to obtain the corrected photovoltaic power generation prediction data V;

[0120]

[0121] The obtained corrected wind power generation power prediction data and corrected photovoltaic power generation power prediction data are corrected prediction data.

[0122] A new energy power generation prediction system based on meteorological elements, the prediction system comprising:

[0123] Collection module: acquire historical wind speed data, historical temperature data, historical humidity data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, and historical photovoltaic module data to obtain historical parameter data;

[0124] Acquire historical wind power generation data and historical photovoltaic power generation data;

[0125] Acquire the first parameter, the second parameter, the fourth parameter, the third parameter and the fifth parameter to obtain first parameter data, second parameter data, fourth parameter data, third parameter data and fifth parameter data;

[0126] Data processing module: perform principal component analysis on historical parameter data to obtain filtered historical parameters;

[0127] According to the screening historical parameters and the historical parameter data, the screening historical parameter data is obtained;

[0128] Calculating the first parameter data, the second parameter data, the fourth parameter data, the third parameter data and the fifth parameter data to obtain predicted parameter data;

[0129] Model training module: divide the selected historical parameter data by time and data ratio, and collect historical wind power data and historical photovoltaic power data for calculation to obtain ultra-short-term wind prediction model, ultra-short-term photovoltaic prediction model, short-term wind prediction model, short-term photovoltaic prediction model, medium- and long-term wind prediction model, and medium- and long-term photovoltaic prediction model;

[0130] Prediction data verification module: obtains ultra-short-term wind power prediction data, ultra-short-term photovoltaic prediction data, short-term wind power prediction data, short-term photovoltaic prediction data, medium- and long-term wind power prediction data, and medium- and long-term photovoltaic prediction data according to prediction parameter data and prediction models;

[0131] The revised wind power generation power forecast data and the revised photovoltaic power generation power forecast data are obtained.

[0132] In the existing field of renewable energy power generation, accurate prediction of power generation is crucial for the stable operation of the power grid and the effective management of energy.

[0133] Compared with the prior art, the present invention has the following beneficial effects:

[0134] 1. Extract the main characteristic parameters. The present invention performs principal component analysis on the selected parameters, sorts the obtained analysis results according to the contribution ratio, obtains several characteristic parameters with relatively large contributions, reduces the dimension of model training, and makes the results of model training more accurate.

[0135] 2. Model prediction data verification. The present invention establishes three prediction models, establishes the models according to different time ranges, verifies the prediction results according to time, obtains corrected prediction data, and obtains more accurate data. BRIEF DESCRIPTION OF THE DRAWINGS

[0136] Other features, objects and advantages of the present invention will become more apparent from the detailed description of non-limiting embodiments made with reference to the following drawings:

[0137] Figure 1 It is a schematic diagram of the process of the present invention;

[0138] Figure 2 This is a schematic diagram of the parameters of this hair;

[0139] Figure 3 It is a schematic diagram of the system of the present invention;

[0140] Figure 4 Schematic diagram of the principal component analysis method of the present invention; DETAILED DESCRIPTION

[0141] In order to make the above-mentioned objects, features and advantages of the present invention more obvious and understandable, the present invention is further described in detail below with reference to the accompanying drawings and specific embodiments.

[0142] See also Figure 1 and Figure 3 , a new energy power generation prediction method based on meteorological elements, the specific steps for obtaining the power generation are as follows:

[0143] Step S1: Acquire historical wind speed data, historical temperature data, historical humidity data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, and historical photovoltaic module data to obtain historical parameter data;

[0144] The specific steps for obtaining historical parameter data are as follows:

[0145] Through the database, the wind speed, temperature, humidity and wind direction records are read to obtain historical wind direction data, historical temperature data, historical humidity data and historical wind speed data HU;

[0146] Read the historical wind speed data to obtain the historical wind speed data volume Hun and the historical wind speed data variable HUi;

[0147] Among them, the historical wind speed data variable HUi is the composition of the historical wind speed, and the historical wind speed data volume HUn is the number of historical wind speed data variables; for example, if the historical wind speed data variable is [2, 3, 3.2, 4], then the historical wind speed data variable is 4;

[0148] It should be noted that HUi represents the wind speed at the i-th time point; Hun represents the number of historical wind speed data; i represents the corresponding subscript of the historical wind speed data, and the value range of i is: 1~HUn.

[0149] Calculate the historical wind speed data volume Hun and the historical wind speed data variable HUi to obtain the historical wind speed mean

[0150]

[0151] Historical wind speed average The historical wind speed data volume Hun and the historical wind speed data variable HUi are calculated to obtain the historical wind speed standard deviation HUσ;

[0152]

[0153] The historical wind speed standard deviation and historical wind speed mean are calculated to obtain the historical turbulence data HA;

[0154]

[0155] Read the historical wind speed and get the historical wind speed HUih corresponding to the height h;

[0156] It should be noted that HUih represents the wind speed at height h at the i-th time point;

[0157] The historical wind speed HUih1 corresponding to the height h1 and the historical wind speed HUih2 corresponding to the height h2 are calculated to obtain the historical wind shear index data HB;

[0158]

[0159] The solar radiation is acquired through the database to obtain historical solar radiation data;

[0160] Obtain the historical azimuth angle data and historical component efficiency data of the new energy equipment through the database;

[0161] Azimuth angle data and historical module efficiency data constitute historical PV module data;

[0162] It should be noted that the data stored in the database include historical data recorded by the China Meteorological Data Network, new energy equipment parameters and new energy equipment power generation historical data;

[0163] The historical wind speed data, the historical temperature data, the historical wind direction data, the historical turbulence data, the historical wind shear index data, the historical solar radiation data, the historical humidity data, the historical photovoltaic module azimuth data and the historical module efficiency data are recorded as the historical parameter data HK;

[0164] See also Figure 2 and Figure 4 , Step S2: Perform principal component analysis on the historical parameter data to obtain the screened historical parameters;

[0165] The steps to obtain the screening history parameters are as follows:

[0166] Read the historical parameter data HK to obtain the historical parameter data variable HKmi and the historical parameter data quantity HKn;

[0167] The historical parameter data variables are expressed in the form of: {HK1i, HK2i, HK3i, HK4i, HK5i, HK6i, HK7i, HK8i, HK9i};

[0168] It should be noted that HKmi represents the i-th variable of the m-th parameter; HKn represents the number of variables of the parameter; the value range of m is: 1 to 9, and the value range of i is: 1 to HKn;

[0169] Calculate the historical parameter data variable HKmi and the historical parameter data volume HKn to obtain the historical parameter mean

[0170]

[0171] Historical parameter mean The manifestation is:

[0172] The historical parameter mean The historical parameter data variable HKmi and the historical parameter data quantity HKn are calculated to obtain the historical parameter standard deviation HKmσ;

[0173]

[0174] The historical parameter standard deviation HKmσ is expressed in the form of: {HK1σ, HK2σ, HK3σ, HK4σ, HK5σ, HK6σ, HK7σ, HK8σ, HK9σ};

[0175] The standard deviation HKmσ of historical parameters and the mean value of historical parameters Calculate with historical parameter data variable HKmi to obtain standardized historical parameter HZmi;

[0176]

[0177] The standardized historical parameter HZmi is expressed as: {HZ1i, HZ2i, HZ3i, HZ4i, HZ5i, HZ6i, HZ7i, HZ8i, HZ9i};

[0178] The standardized historical parameter HZmi, the historical parameter standard deviation HKmσ and the historical parameter data volume HKmn are calculated to obtain the covariance matrix C;

[0179]

[0180] The covariance matrix C is expressed as:

[0181]

[0182] Calculate the covariance matrix C to obtain the eigenvalue Hλm and eigenvector Hαm;

[0183] C*Hαm=Hλm*Hαm;

[0184] The eigenvalue Hλn is expressed as: {Hλ1, Hλ2, Hλ3, Hλ4, Hλ5, Hλ6, Hλ7, Hλ8, Hλ9};

[0185] Screen the characteristic value Hλm to obtain the screening history parameter;

[0186] The screening steps are as follows:

[0187] Sort the eigenvalues ​​Hλm in order of size to obtain a set of sequential eigenvalues;

[0188] Read the historical parameters corresponding to the first five eigenvalues ​​of the sequential eigenvalue set to obtain the screening historical parameters;

[0189] It should be noted that the present invention selects the parameters corresponding to the first five eigenvalues ​​for screening, and the specific screening eigenvalue results vary depending on the historical data parameters selected by the model training personnel;

[0190] Step S3: Acquire historical wind power generation data and historical photovoltaic power generation data;

[0191] According to the screening historical parameters and the historical parameter data, the screening historical parameter data is obtained;

[0192] The historical parameter data is divided into time and data proportions, and the historical wind power generation data and the historical photovoltaic power generation data are combined for calculation to obtain an ultra-short-term wind prediction model, an ultra-short-term photovoltaic prediction model, a short-term wind prediction model, a short-term photovoltaic prediction model, a medium- and long-term wind prediction model, and a medium- and long-term photovoltaic prediction model;

[0193] Obtain historical wind power generation data and historical photovoltaic power generation data through the database;

[0194] In this embodiment, if the five parameters selected in the historical parameters are: historical wind speed, historical wind direction, historical temperature, historical humidity, and historical solar radiation;

[0195] According to the screening historical parameters, the historical parameter data is read to obtain the screening historical parameter data;

[0196] Steps to obtain filtered historical parameter data:

[0197] Find the data corresponding to the filtered historical parameters in the historical parameter data, extract the corresponding data, and obtain the filtered historical parameter data;

[0198] According to the time period, the screened historical parameter data is divided into ultra-short-term historical parameter data, short-term historical parameter data and medium- and long-term historical parameter data;

[0199] The time period division rules are as follows:

[0200] The screening historical parameter data within four hours is recorded as ultra-short-term historical parameter data;

[0201] The screening historical parameter data within 72 hours shall be recorded as short-term historical parameter data;

[0202] The historical parameter data of screening within one month is recorded as medium- and long-term parameter data;

[0203] It should be noted that the present invention sets historical data within four hours as ultra-short term, historical data within seventy-two hours as short term, and historical data within one month as medium- to long-term. The specific time setting can be determined by the model training personnel based on the actual situation.

[0204] It should be noted that, from the perspective of time span, the short-term time range may include the ultra-short-term time range, but in actual forecasting practice, these two terms are used to distinguish different forecasting tasks and needs. This division helps to clarify the forecasting objectives, select appropriate forecasting methods, and optimize the operation and management of the power grid.

[0205] The ultra-short-term historical parameter data, the short-term historical parameter data and the medium- and long-term parameter data are divided according to the data ratio to obtain an ultra-short-term historical data parameter training set, an ultra-short-term historical parameter data test set, a short-term historical data parameter training set, a short-term historical parameter data test set, a medium- and long-term historical data parameter training set and a medium- and long-term historical parameter data test set;

[0206] The proportion division rules are as follows:

[0207] The data set is divided into 7:3 parts according to the data ratio, with the 7-part as the training set and the 3-part as the test set;

[0208] It should be noted that the ratio of 7:3 is the ratio setting of the present invention. The specific ratio setting can be set by the model training personnel according to actual needs.

[0209] Using TensorFlow's tf.keras, we calculated the ultra-short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and ultra-short-term historical parameter data test set to obtain the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic prediction model.

[0210] Using TensorFlow's tf.keras, we calculated the short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and short-term historical parameter data test set to obtain the short-term wind power prediction model and the short-term photovoltaic prediction model.

[0211] Using TensorFlow's tf.keras, we calculated the medium- and long-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and medium- and long-term historical parameter data test set to obtain the medium- and long-term wind power prediction model and the medium- and long-term photovoltaic prediction model.

[0212] Step S4: acquiring wind speed, wind direction, humidity, temperature and solar radiation data to obtain wind speed data, wind direction data, humidity data, temperature data and solar radiation data;

[0213] Calculate wind speed data, wind direction data, humidity data, temperature data and solar radiation data to obtain prediction parameter data;

[0214] The steps for obtaining prediction parameter data are as follows:

[0215] Use an anemometer to obtain the wind speed and obtain wind speed data HP1;

[0216] Use the wind vane to obtain the wind direction and obtain the wind direction data HP2;

[0217] Using a thermometer, the temperature is acquired to obtain temperature data HP3;

[0218] Use a hygrometer to obtain humidity and obtain humidity data HP4;

[0219] Use a solar radiometer to obtain solar radiation and obtain solar radiation data HP5;

[0220] Read the wind speed data HP1 to obtain the wind speed data quantity HPn and the wind speed data variable HP1j;

[0221] Read the wind direction data HP2 to obtain the wind direction data quantity HPn and the wind direction data variable HP2j;

[0222] Read the temperature data HP3 to obtain the temperature data quantity HPn and the temperature data variable HP3j;

[0223] Read the humidity data HP4 to obtain the humidity data quantity HPn and the temperature data variable HP4j;

[0224] Read the solar radiation data HP5 to obtain the solar radiation data quantity HPn and the solar radiation data variable HP5j;

[0225] Calculate the wind speed data volume HPn and wind speed data variable HP1j to get the wind speed mean

[0226]

[0227] Calculate the wind direction data volume HPn and wind direction data variable HP2j to get the wind direction mean

[0228]

[0229] Calculate the temperature data volume HPn and temperature data variable HP3j to get the temperature mean

[0230]

[0231] Calculate the humidity data volume HPn and the temperature data variable HP4j to get the humidity mean

[0232]

[0233] Calculate the solar radiation data volume HPn and solar radiation data variable HP5j to get the mean solar radiation value

[0234]

[0235] Average wind speed The wind speed data volume HPn and wind speed data variable HP1j are calculated to obtain the wind speed standard deviation HP1σ;

[0236]

[0237] Mean wind direction The wind direction data volume HPn and the wind direction data variable HP2j are calculated to obtain the wind direction standard deviation HP2σ;

[0238]

[0239] For the mean temperature The temperature data quantity HPn and the temperature data variable HP3j are calculated to obtain the temperature standard deviation HP3σ;

[0240]

[0241] Average humidity The humidity data volume HPn and the temperature data variable HP4j are calculated to obtain the humidity standard deviation HP4σ;

[0242]

[0243] The mean solar radiation The solar radiation data quantity HPn and the solar radiation data variable HP5j are calculated to obtain the solar radiation standard deviation HP5σ;

[0244]

[0245] For wind speed standard deviation HP1σ, wind speed mean Calculate with wind speed data variable HP1j to obtain wind speed standard value HZ1j;

[0246]

[0247] The standard deviation of wind direction HP2σ and the mean of wind direction Calculate with wind direction data variable HP2j to obtain wind direction standard value HZ2j;

[0248]

[0249] For temperature standard deviation HP3σ, temperature mean Calculate with temperature data variable HP3j to obtain temperature standard value HZ3j;

[0250]

[0251] Humidity standard deviation HP4σ, humidity mean Calculate with temperature data variable HP4j to obtain humidity standard value HZ4j;

[0252]

[0253] The standard deviation of solar radiation HP5σ and the mean solar radiation Calculate with solar radiation data variable HP5j to obtain solar radiation standard value HZ5j;

[0254]

[0255] The standard value of wind speed HZ1j, the standard value of wind direction HZ2j, the standard value of temperature HZ3j, the standard value of humidity HZ4j and the standard value of solar radiation HZ5j are recorded as prediction parameter data;

[0256] Step S5: acquiring ultra-short-term wind power forecast data, ultra-short-term photovoltaic forecast data, short-term wind power forecast data, short-term photovoltaic forecast data, medium- and long-term wind power forecast data, and medium- and long-term photovoltaic forecast data;

[0257] Acquiring revised wind power generation power forecast data and revised photovoltaic power generation power forecast data;

[0258] The ultra-short-term wind power prediction data F1 and the ultra-short-term photovoltaic power prediction data V1 are obtained by combining the prediction parameter data with the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic power prediction model;

[0259] By using the prediction parameter data combined with the short-term wind prediction model and the short-term photovoltaic prediction model, the short-term wind prediction data F2 and the short-term photovoltaic prediction data V2 are obtained;

[0260] The medium- and long-term wind power forecast data F3 and the medium- and long-term photovoltaic forecast data V3 are obtained by combining the forecast parameter data with the medium- and long-term wind power forecast model and the medium- and long-term photovoltaic forecast model;

[0261] It should be noted that the medium- and long-term prediction model predicts data at 60-minute time intervals; the short-term prediction model predicts data at 30-minute time intervals; and the ultra-short-term prediction model predicts data at 15-minute time intervals.

[0262] The weights of the ultra-short-term wind prediction model, the short-term wind prediction model and the medium- and long-term wind prediction model are set to obtain the ultra-short-term wind prediction model weight A1, the short-term wind prediction model weight A2 and the medium- and long-term wind prediction model weight A3;

[0263] It should be noted that in this application, A1<A2<A3; B1<B2<B3; in the specific data prediction process, the data prediction personnel shall set it according to the actual situation;

[0264] The weights of the ultra-short-term photovoltaic prediction model, the short-term photovoltaic prediction model and the medium- and long-term photovoltaic prediction model are set to obtain the weight B1 of the ultra-short-term photovoltaic prediction model, the weight B2 of the short-term photovoltaic prediction model and the weight B3 of the medium- and long-term photovoltaic prediction model;

[0265] Calculate the ultra-short-term wind forecast data F1, the short-term wind forecast data F2, the medium- and long-term wind forecast data F3, the ultra-short-term wind forecast model weight A1, the short-term wind forecast model weight A2 and the medium- and long-term wind forecast model weight A3 to obtain the modified wind power generation forecast data F;

[0266]

[0267] The ultra-short-term photovoltaic prediction data V1, the short-term photovoltaic prediction data V2, the medium- and long-term photovoltaic prediction data V3, the ultra-short-term photovoltaic prediction model weight B1, the short-term photovoltaic prediction model weight B2 and the medium- and long-term photovoltaic prediction model weight B3 are calculated to obtain the corrected photovoltaic power generation prediction data V;

[0268]

[0269] The obtained revised wind power generation power prediction data and revised photovoltaic power generation power prediction data are revised prediction data, which realize the prediction of new energy power generation power.

[0270] The above formulas are all dimensionless and numerical calculations. The formula is a formula for the most recent real situation obtained by collecting a large amount of data and performing software simulation. The preset parameters in the formula are set by technical personnel in this field according to actual conditions. For example, if there are weight coefficients and proportional coefficients, their set sizes are to quantify each parameter to obtain a specific value, which is convenient for subsequent comparison. Regarding the size of the weight coefficient and the proportional coefficient, it is sufficient as long as it does not affect the proportional relationship between the parameter and the quantized value.

[0271] Finally, it should be noted that the above-described embodiments are only specific implementations of the present invention, which are used to illustrate the technical solutions of the present invention, rather than to limit them. The protection scope of the present invention is not limited thereto. Although the present invention is described in detail with reference to the above-described embodiments, ordinary technicians in the field should understand that any technician familiar with the technical field can still modify the technical solutions recorded in the above-described embodiments within the technical scope disclosed by the present invention, or can easily think of changes, or make equivalent replacements for some of the technical features therein; and these modifications, changes or replacements do not make the essence of the corresponding technical solutions deviate from the spirit and scope of the technical solutions of the embodiments of the present invention, and should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be based on the protection scope of the claims.

Claims

1. A method for predicting renewable energy power generation based on meteorological factors, characterized in that: Prediction methods include: Step S1: Acquire historical wind speed data, historical temperature data, historical humidity data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, and historical photovoltaic module data to obtain historical parameter data; Step S2: Perform principal component analysis on the historical parameter data to obtain the screened historical parameters; Step S3: Acquire historical wind power generation data and historical photovoltaic power generation data; Find the data corresponding to the filtered historical parameters in the historical parameter data, extract the corresponding data, and obtain the filtered historical parameter data; The historical parameter data is divided into time and data proportions, and the historical wind power generation data and the historical photovoltaic power generation data are combined for calculation to obtain an ultra-short-term wind prediction model, an ultra-short-term photovoltaic prediction model, a short-term wind prediction model, a short-term photovoltaic prediction model, a medium- and long-term wind prediction model, and a medium- and long-term photovoltaic prediction model; Step S4: acquiring wind speed, wind direction, humidity, temperature and solar radiation data, obtaining wind speed data, wind direction data, humidity data, temperature data and solar radiation data, and performing calculations to obtain prediction parameter data; Step S5: acquiring ultra-short-term wind power forecast data, ultra-short-term photovoltaic forecast data, short-term wind power forecast data, short-term photovoltaic forecast data, medium- and long-term wind power forecast data, and medium- and long-term photovoltaic forecast data according to the forecast parameter data and the forecast model; Acquiring the revised wind power generation power prediction data and the revised photovoltaic power generation power prediction data to obtain revised prediction data; The workflow of step S1 is as follows: Read the historical wind speed, temperature, humidity and wind direction records to obtain historical wind direction data, historical temperature data, historical humidity data and historical wind speed data HU; Read the historical wind speed data to obtain the historical wind speed data volume Hun and the historical wind speed data variable HUi; Calculate the historical wind speed data volume Hun and the historical wind speed data variable HUi to obtain the historical wind speed mean Historical wind speed average The historical wind speed data volume Hun and the historical wind speed data variable HUi are calculated to obtain the historical wind speed standard deviation HUσ; The historical wind speed standard deviation and historical wind speed mean are calculated to obtain the historical turbulence data HA; Read the historical wind speed and get the historical wind speed HUih corresponding to the height h; The historical wind speed HUih1 corresponding to the height h1 and the historical wind speed HUih2 corresponding to the height h2 are calculated to obtain the historical wind shear index data HB; Acquire solar radiation and obtain historical solar radiation data; Acquire the historical azimuth angle data and historical component efficiency data of the new energy equipment to obtain the historical photovoltaic component data; The historical wind speed data, the historical temperature data, the historical wind direction data, the historical turbulence data, the historical wind shear index data, the historical solar radiation data, the historical humidity data, the historical photovoltaic module azimuth data and the historical module efficiency data are recorded as the historical parameter data HK; Read the historical parameter data HK to obtain the historical parameter data variable HKmi and the historical parameter data quantity HKn; The historical parameter data variables are expressed in the form of: {HK1i, HK2i, HK3i, HK4i, HK5i, HK6i, HK7i, HK8i, HK9i}; Calculate the historical parameter data variable HKmi and the historical parameter data volume HKn to obtain the historical parameter mean Historical parameter mean The manifestation is: The historical parameter mean The historical parameter data variable HKmi and the historical parameter data quantity HKn are calculated to obtain the historical parameter standard deviation HKmσ; The historical parameter standard deviation HKmσ is expressed in the form of: {HK1σ, HK2σ, HK3σ, HK4σ, HK5σ, HK6σ, HK7σ, HK8σ, HK9σ}; The standard deviation HKmσ of historical parameters and the mean value of historical parameters Calculate with historical parameter data variable HKmi to obtain standardized historical parameter HZmi; The standardized historical parameter HZmi is expressed as: {HZ1i, HZ2i, HZ3i, HZ4i, HZ5i, HZ6i, HZ7i, HZ8i, HZ9i}; The standardized historical parameter HZmi, the historical parameter standard deviation HKmσ and the historical parameter data volume HKmn are calculated to obtain the covariance matrix C; The covariance matrix C is expressed as: Calculate the covariance matrix C to obtain the eigenvalue Hλm and eigenvector Hαm; C*Hαm=Hλm*Hαm; The eigenvalue Hλm is expressed as: {Hλ1, Hλ2, Hλ3, Hλ4, Hλ5, Hλ6, Hλ7, Hλ8, Hλ9}; Screen the characteristic value Hλm to obtain the screening history parameter; The screening steps are as follows: Sort the eigenvalues ​​Hλm in order of size to obtain a set of sequential eigenvalues; Read the historical parameters corresponding to the first five characteristic values ​​of the sequential characteristic value set to obtain the screening historical parameters, and enter step S3; The workflow of step S5 is as follows: The ultra-short-term wind power prediction data F1 and the ultra-short-term photovoltaic power prediction data V1 are obtained by combining the prediction parameter data with the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic power prediction model; By using the prediction parameter data combined with the short-term wind prediction model and the short-term photovoltaic prediction model, the short-term wind prediction data F2 and the short-term photovoltaic prediction data V2 are obtained; The medium- and long-term wind power forecast data F3 and the medium- and long-term photovoltaic forecast data V3 are obtained by combining the forecast parameter data with the medium- and long-term wind power forecast model and the medium- and long-term photovoltaic forecast model; The weights of the ultra-short-term wind prediction model, the short-term wind prediction model and the medium- and long-term wind prediction model are set to obtain the ultra-short-term wind prediction model weight A1, the short-term wind prediction model weight A2 and the medium- and long-term wind prediction model weight A3; The weights of the ultra-short-term photovoltaic prediction model, the short-term photovoltaic prediction model and the medium- and long-term photovoltaic prediction model are set to obtain the weight B1 of the ultra-short-term photovoltaic prediction model, the weight B2 of the short-term photovoltaic prediction model and the weight B3 of the medium- and long-term photovoltaic prediction model; Calculate the ultra-short-term wind forecast data F1, the short-term wind forecast data F2, the medium- and long-term wind forecast data F3, the ultra-short-term wind forecast model weight A1, the short-term wind forecast model weight A2 and the medium- and long-term wind forecast model weight A3 to obtain the modified wind power generation forecast data F; The ultra-short-term photovoltaic prediction data V1, the short-term photovoltaic prediction data V2, the medium- and long-term photovoltaic prediction data V3, the ultra-short-term photovoltaic prediction model weight B1, the short-term photovoltaic prediction model weight B2 and the medium- and long-term photovoltaic prediction model weight B3 are calculated to obtain the corrected photovoltaic power generation prediction data V; The obtained corrected wind power generation power prediction data and corrected photovoltaic power generation power prediction data are corrected prediction data.

2. The method for predicting power generation of renewable energy based on meteorological elements according to claim 1, characterized in that: The workflow of step S3 is as follows: Obtain historical wind power generation data and historical photovoltaic power generation data; The screening historical parameters include: the first historical parameter, the second historical parameter, the third historical parameter, the fourth historical parameter, and the fifth historical parameter; Find the data corresponding to the filtered historical parameters in the historical parameter data, extract the corresponding data, and obtain the filtered historical parameter data; According to the time period, the screened historical parameter data is divided into ultra-short-term historical parameter data, short-term historical parameter data and medium- and long-term historical parameter data; The time period division rules are as follows: The screening historical parameter data within four hours is recorded as ultra-short-term historical parameter data; The screening historical parameter data within 72 hours shall be recorded as short-term historical parameter data; The historical parameter data of screening within one month is recorded as medium- and long-term parameter data; The ultra-short-term historical parameter data, the short-term historical parameter data and the medium- and long-term parameter data are divided according to the data ratio to obtain an ultra-short-term historical data parameter training set, an ultra-short-term historical parameter data test set, a short-term historical data parameter training set, a short-term historical parameter data test set, a medium- and long-term historical data parameter training set and a medium- and long-term historical parameter data test set; Using TensorFlow's tf.keras, we calculated the ultra-short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and ultra-short-term historical parameter data test set to obtain the ultra-short-term wind power prediction model and the ultra-short-term photovoltaic prediction model. Using TensorFlow's tf.keras, we calculated the short-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and short-term historical parameter data test set to obtain the short-term wind power prediction model and the short-term photovoltaic prediction model. Using TensorFlow's tf.keras, the medium- and long-term historical data parameter training set, historical wind power data, historical photovoltaic power data, and medium- and long-term historical parameter data test set are calculated to obtain a medium- and long-term wind prediction model and a medium- and long-term photovoltaic prediction model. The ultra-short-term wind prediction model, ultra-short-term photovoltaic prediction model, short-term wind prediction model, short-term photovoltaic prediction model, medium- and long-term wind prediction model, and medium- and long-term photovoltaic prediction model are defined as prediction models, and step S4 is entered.

3. The method for predicting power generation of renewable energy based on meteorological elements according to claim 2 is characterized in that: The workflow of step S4 is as follows: Acquire the first parameter to obtain first parameter data HP1; Acquire the second parameter to obtain second parameter data HP2; Acquire the third parameter to obtain third parameter data HP3; Acquire the fourth parameter to obtain fourth parameter data HP4; Acquire the fifth parameter to obtain fifth parameter data HP5; Data processing is performed on the first parameter data HP1, the second parameter data HP2, the third parameter data HP3, the fourth parameter data HP4 and the fifth parameter data HP5.

4. The method for predicting renewable energy power generation based on meteorological elements according to claim 3 is characterized in that: The workflow of data processing is as follows: Read the first parameter data HP1 to obtain the first parameter data quantity HPn and the first parameter data variable HP1j; Read the second parameter data HP2 to obtain the second parameter data quantity HPn and the second parameter data variable HP2j; Read the third parameter data HP3 to obtain the third parameter data quantity HPn and the third parameter data variable HP3j; The fourth parameter data HP4 is read to obtain the fourth parameter data quantity HPn and the third parameter data variable HP4j; Read the fifth parameter data HP5 to obtain the fifth parameter data amount HPn and the fifth parameter data variable HP5j; Calculate the first parameter data volume HPn and the first parameter data variable HP1j to obtain the first parameter mean Calculate the second parameter data volume HPn and the second parameter data variable HP2j to obtain the second parameter mean Calculate the third parameter data volume HPn and the third parameter data variable HP3j to obtain the third parameter mean Calculate the fourth parameter data volume HPn and the fourth parameter data variable HP4j to obtain the fourth parameter mean Calculate the fifth parameter data volume HPn and the fifth parameter data variable HP5j to obtain the fifth parameter mean The mean of the first parameter The first parameter data quantity HPn and the first parameter data variable HP1j are calculated to obtain the first parameter standard deviation HP1σ; The mean of the second parameter The second parameter data quantity HPn and the second parameter data variable HP2j are calculated to obtain the second parameter standard deviation HP2σ; The mean of the third parameter The third parameter data quantity HPn and the third parameter data variable HP3j are calculated to obtain the third parameter standard deviation HP3σ; The mean of the fourth parameter The fourth parameter data quantity HPn and the fourth parameter data variable HP4j are calculated to obtain the fourth parameter standard deviation HP4σ; The mean of the fifth parameter The fifth parameter data quantity HPn and the fifth parameter data variable HP5j are calculated to obtain the fifth parameter standard deviation HP5σ; For the first parameter standard deviation HP1σ, the first parameter mean Calculate with the first parameter data variable HP1j to obtain the first parameter standard value HZ1j; The second parameter standard deviation HP2σ, the second parameter mean Calculate with the second parameter data variable HP2j to obtain the second parameter standard value HZ2j; The third parameter standard deviation HP3σ, the third parameter mean Calculate with the third parameter data variable HP3j to obtain the third parameter standard value HZ3j; The fourth parameter standard deviation HP4σ, the fourth parameter mean Calculate with the fourth parameter data variable HP4j to obtain the fourth parameter standard value HZ4j; The fifth parameter standard deviation HP5σ, the fifth parameter mean Calculate with the fifth parameter data variable HP5j to obtain the fifth parameter standard value HZ5j; The first parameter standard value HZ1j, the second parameter standard value HZ2j, the third parameter standard value HZ3j, the fourth parameter standard value HZ4j and the fifth parameter standard value HZ5j are recorded as prediction parameter data.

5. A new energy power generation prediction system based on meteorological elements, applicable to a new energy power generation prediction method based on meteorological elements as claimed in any one of claims 1 to 4, characterized in that: The forecasting system includes: Collection module: acquire historical wind speed data, historical temperature data, historical humidity data, historical wind direction data, historical turbulence data, historical wind shear index data, historical solar radiation data, and historical photovoltaic module data to obtain historical parameter data; Acquire historical wind power generation data and historical photovoltaic power generation data; Acquire the first parameter, the second parameter, the fourth parameter, the third parameter and the fifth parameter to obtain first parameter data, second parameter data, fourth parameter data, third parameter data and fifth parameter data; Data processing module: perform principal component analysis on historical parameter data to obtain filtered historical parameters; According to the screening historical parameters and the historical parameter data, the screening historical parameter data is obtained; Calculating the first parameter data, the second parameter data, the fourth parameter data, the third parameter data and the fifth parameter data to obtain predicted parameter data; Model training module: divide the selected historical parameter data by time and data ratio, and collect historical wind power data and historical photovoltaic power data for calculation to obtain ultra-short-term wind prediction model, ultra-short-term photovoltaic prediction model, short-term wind prediction model, short-term photovoltaic prediction model, medium- and long-term wind prediction model, and medium- and long-term photovoltaic prediction model; Prediction data verification module: obtains ultra-short-term wind power prediction data, ultra-short-term photovoltaic prediction data, short-term wind power prediction data, short-term photovoltaic prediction data, medium- and long-term wind power prediction data, and medium- and long-term photovoltaic prediction data according to prediction parameter data and prediction models; The revised wind power generation power forecast data and the revised photovoltaic power generation power forecast data are obtained.

Images