New energy plant station meteorological data quality control method and system based on EOF

Abstract

The invention relates to the technical field of data quality control, in particular to a new energy plant station meteorological data control method and system based on EOF, and the method comprises the steps: collecting meteorological observation information of a new energy station, extracting meteorological factors, carrying out the preprocessing of the meteorological factors, obtaining a meteorological observation sequence, carrying out the abnormal data recognition of the meteorological observation sequence, and removing the abnormal data. The method comprises the following steps: acquiring a meteorological correction subsequence, correcting the meteorological correction subsequence, replacing abnormal data to generate a meteorological correction sequence, carrying out collaborative analysis on the meteorological correction sequence, and identifying and optimizing potential anomalies by adopting a collaborative analysis result, thereby effectively solving the problems that the traditional meteorological data anomaly detection is inaccurate and the correction is not timely; the meteorological data can be quickly corrected and optimized under complex working conditions, so that the data reliability is improved, and the precision and efficiency of operation management of a new energy plant station are enhanced.

Description

Technical Field

[0001] The present invention relates to the technical field of data quality control, and in particular to a method and system for quality control of meteorological data of a new energy plant based on EOF. Background Art

[0002] In the field of modern renewable energy and electricity, high-precision quality control of meteorological data is one of the key factors in improving the operating efficiency of renewable energy plants. With the development of large-scale renewable energy grid connection and smart grid technology, the requirements for the real-time and accuracy of meteorological data are increasing. Traditional data processing methods often face problems such as inaccurate anomaly detection and untimely data correction. This leads to the inability to effectively guarantee the reliability and practicality of data in the face of complex meteorological conditions.

[0003] Although existing technologies have made progress in meteorological data collection and preliminary correction, they are still insufficient in comprehensive quality control. Many methods are unable to accurately identify and quickly correct abnormal data when faced with multi-factor coordinated changes and dynamic anomalies, resulting in accumulated errors in meteorological data, which in turn affects the scheduling decisions and operational safety of new energy plants. Under highly dynamic meteorological conditions, meteorological data needs to be detected, corrected and optimized in a very short time, and traditional static analysis methods are difficult to meet this requirement.

[0004] The information disclosed in this background technology section is only intended to deepen the understanding of the overall background technology of the present disclosure, and should not be regarded as acknowledging or suggesting in any form that the information constitutes the prior art known to those skilled in the art. Summary of the invention

[0005] The present invention provides a method and system for quality control of meteorological data of new energy plants and stations based on EOF, which can effectively solve the problems in the background technology. In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A method for quality control of meteorological data of new energy plants and stations based on EOF, the method comprising:

[0007] Collect meteorological observation information of new energy plants and stations, extract meteorological factor data, and pre-process some of the meteorological factor data to obtain a meteorological observation sequence;

[0008] Performing data anomaly detection on the meteorological observation sequence, determining and eliminating comprehensive abnormal records, and obtaining a meteorological correction subsequence;

[0009] Performing data correction on the comprehensive anomaly record to generate a meteorological replacement subsequence, integrating the meteorological replacement subsequence with the meteorological correction subsequence to obtain a meteorological correction sequence;

[0010] The meteorological correction sequence is collaboratively analyzed to obtain collaborative analysis results, potential anomalies are identified based on the collaborative analysis results, and the potential anomalies are optimized.

[0011] Furthermore, preprocessing is performed on the meteorological factor data to obtain a meteorological observation sequence, including:

[0012] Performing time homogenization processing on the data of several meteorological factors, unifying the time bases of several meteorological factors, and obtaining a meteorological pre-observation sequence;

[0013] The missing values ​​of the meteorological observation sequence are extracted according to the meteorological observation sequence, and the missing values ​​are filled and marked to obtain the meteorological observation sequence.

[0014] Furthermore, performing data anomaly detection on the meteorological observation sequence includes:

[0015] Collecting historical meteorological observation information, and setting a temperature range threshold and a temperature change threshold based on the historical meteorological observation information;

[0016] Based on the temperature range threshold, a point-by-point comparison is performed with each observation record of the meteorological observation sequence, and if the observation record is not within the temperature range threshold, it is marked as a range abnormality record;

[0017] Calculating the difference between each observation record and the adjacent observation record point by point according to the meteorological observation sequence, comparing the calculated difference with the temperature change threshold, and marking the observation record as an abnormal change record if the observation record is not within the temperature change threshold;

[0018] The anomaly detection is also applicable to other meteorological factors.

[0019] Furthermore, performing data anomaly detection on the meteorological observation sequence includes:

[0020] Calculating the observation median of the meteorological observation sequence and obtaining the observation time range of the meteorological observation sequence;

[0021] Calculate the standard deviation of any of the observation records based on the historical meteorological observation information as a weight value of the observation record;

[0022] A double-weighted standard deviation test is performed based on the observation median, the observation time range, and the observation record weight value, and the calculation formula is as follows:

[0023]

[0024] Among them, AVE bw is the double weighted average, STD bw is the double-weighted standard deviation, is the observed temperature of the nth observation record, M is the observation median, N is the observation time range, MAD is the median of the difference between the observation record and the observation median, is the weight value of the nth observation record, if |w n |>1.0, then w n =1.0, when the weighted anomaly detection condition is met , it is marked as a weight abnormal record;

[0025] Integrate the range anomaly record, the change anomaly record and the weight anomaly record to obtain a comprehensive anomaly record;

[0026] The anomaly detection is also applicable to other meteorological factors.

[0027] Furthermore, data correction is performed on the comprehensive abnormal record to generate a meteorological replacement subsequence, including:

[0028] Constructing a cubic spline interpolation function based on the meteorological correction subsequence, wherein the cubic spline interpolation function is continuous on the meteorological correction subsequence;

[0029] Substituting the time point of the comprehensive anomaly record into the cubic spline interpolation function, calculating the anomaly interpolation, and obtaining the meteorological replacement subsequence;

[0030] The meteorological replacement subsequence replaces the comprehensive anomaly record to generate a meteorological correction sequence.

[0031] Further, collaborative analysis is performed on the meteorological correction sequence to obtain collaborative analysis results, potential anomalies are identified based on the collaborative analysis results, and the potential anomalies are optimized, including:

[0032] Constructing a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series, and the matrix columns represent meteorological factors; respectively calculating eigenvalues ​​of the covariance matrix and eigenvectors corresponding to the eigenvalues, wherein the eigenvalues ​​are used to represent weight ratios corresponding to the meteorological factors, and the eigenvectors are used to indicate the mutual relationships and change trends of the meteorological factors;

[0033] A cumulative contribution threshold is set according to historical meteorological observation information, the eigenvalue and the corresponding eigenvector are screened based on the cumulative contribution threshold, a screening result is obtained, and the screened eigenvectors are combined into an eigenvector matrix;

[0034] The principal component analysis is performed on the characteristic vector matrix to obtain the time series change information of the meteorological factors, and the statistical anomaly detection method is combined to identify potential abnormal records and optimize the potential abnormal records.

[0035] Further, constructing a covariance matrix based on the meteorological correction sequence includes:

[0036] constructing a meteorological matrix according to the meteorological correction sequence, and performing standardization processing on the meteorological matrix to obtain a meteorological standard matrix;

[0037] Each of the meteorological factors is combined with all the meteorological factors in a single cycle to generate a number of meteorological variable pairs, and covariance calculation is performed on the meteorological variable pairs to obtain a number of covariances. The covariance calculation formula is as follows:

[0038]

[0039] Among them, z ij With z ik is the observation record of the jth and kth meteorological factors at the i-th time node, and is the observed average value of the jth and kth meteorological factors;

[0040] A covariance matrix is ​​constructed based on the plurality of covariances.

[0041] Furthermore, principal component analysis is performed on the eigenvector matrix to obtain information on coordinated changes in meteorological factors, including:

[0042] Performing a matrix multiplication operation on the meteorological standard matrix and the eigenvector matrix to obtain a principal component analysis matrix, wherein each column in the principal component analysis matrix represents a principal component direction, and each row under the column represents a principal component score of a time node;

[0043] According to the principal component scores, the coordinated change patterns of meteorological factors at the same time node are identified, and the time series change information of meteorological factors is obtained.

[0044] A meteorological data control system for a new energy plant based on EOF, the system comprising:

[0045] A meteorological data preprocessing module collects meteorological observation information of new energy plants and stations, extracts meteorological factor data, and preprocesses some of the meteorological factor data to obtain a meteorological observation sequence;

[0046] An abnormal data elimination module performs data anomaly detection on the meteorological observation sequence, determines and eliminates comprehensive abnormal records, and obtains a meteorological correction subsequence; a data correction integration module performs data correction on the comprehensive abnormal records, generates a meteorological replacement subsequence, integrates the meteorological replacement subsequence with the meteorological correction subsequence, and obtains a meteorological correction sequence;

[0047] The collaborative anomaly optimization module collaboratively analyzes the meteorological correction sequence, obtains collaborative analysis results, identifies potential anomalies based on the collaborative analysis results, and optimizes the potential anomalies.

[0048] Furthermore, the collaborative abnormality optimization module includes:

[0049] A covariance matrix construction unit, which constructs a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series and the matrix columns represent meteorological factors;

[0050] an eigenvalue vector analysis unit, for respectively calculating the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues, wherein the eigenvalues ​​are used to represent the weight ratios corresponding to the meteorological factors, and the eigenvectors are used to indicate the mutual relationships and change trends of the meteorological factors;

[0051] A feature vector screening unit, which sets a cumulative contribution threshold according to historical meteorological observation information, screens the feature values ​​and corresponding feature vectors based on the cumulative contribution threshold, obtains screening results, and forms a feature vector matrix with the screened feature vectors;

[0052] The principal component anomaly optimization unit performs principal component analysis on the eigenvector matrix to obtain the time series change information of the meteorological factors, and combines the statistical anomaly detection method to identify potential anomaly records and optimize the potential anomaly records.

[0053] The technical solution of the present invention can achieve the following technical effects:

[0054] It effectively solves the problems of inaccurate anomaly detection and untimely correction of traditional meteorological data, enabling meteorological data to be quickly corrected and optimized under complex working conditions, thereby improving data reliability and enhancing the accuracy and efficiency of new energy plant operation management.

[0055] The above description is only an overview of the technical solution of the present application. In order to more clearly understand the technical means of the present application, it can be implemented in accordance with the contents of the specification. In order to make the above and other purposes, features and advantages of the present application more obvious and easy to understand, the specific implementation methods of the present application are listed below. BRIEF DESCRIPTION OF THE DRAWINGS

[0056] In order to more clearly illustrate the embodiments of the present invention or the technical solutions in the prior art, the drawings required for use in the embodiments or the description of the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments recorded in the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without paying creative work.

[0057] Figure 1 It is a flow chart of the meteorological data control method of new energy plants based on EOF;

[0058] Figure 2 Schematic diagram of the process flow for collaborative exception optimization;

[0059] Figure 3 This is a schematic diagram of the collaborative exception optimization module structure. DETAILED DESCRIPTION

[0060] The technical solutions in the embodiments of the present invention will be described clearly and completely below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, rather than all the embodiments.

[0061] Unless otherwise defined, all technical and scientific terms used herein have the same meaning as those commonly understood by those skilled in the art of the present invention. The terms used in the specification of the present invention are only for the purpose of describing specific embodiments and are not intended to limit the present invention. The term "and / or" used herein includes any and all combinations of one or more related listed items.

[0062] Embodiment 1;

[0063] like Figure 1 As shown, the present application provides a method for controlling meteorological data of a new energy plant based on EOF, the method comprising:

[0064] S10: Collect meteorological observation information of new energy plants and stations, extract meteorological factor data, and pre-process several meteorological factor data to obtain a meteorological observation sequence;

[0065] S20: Perform data anomaly detection on the meteorological observation sequence, determine and remove comprehensive abnormal records, and obtain a meteorological correction subsequence;

[0066] S30: performing data correction on the comprehensive abnormal record, generating a meteorological replacement subsequence, integrating the meteorological replacement subsequence and the meteorological correction subsequence, and obtaining a meteorological correction sequence;

[0067] S40: Perform collaborative analysis on the meteorological correction sequence, obtain collaborative analysis results, identify potential anomalies based on the collaborative analysis results, and optimize the potential anomalies.

[0068] Specifically, high-performance meteorological sensors installed in new energy plants and stations are used to collect key meteorological parameters in real time, including temperature, humidity, wind speed, solar radiation, etc. The collected raw data are automatically cleaned, including the identification and filtering of erroneous data, such as temperature readings that do not conform to physical possibilities, including filtering out obvious noise and data points that do not conform to physical limitations. Subsequently, the data is normalized so that information from different sensors and data sources can be uniformly processed and compared. Statistical methods can be used to detect anomalies in meteorological observation sequences, and data records identified as abnormal are removed from the observation sequence. Then, numerical analysis techniques such as numerical interpolation methods are used to intelligently fill in the removed abnormal points to generate meteorological replacement subsequences. The calculated meteorological replacement subsequences are integrated with the original meteorological correction subsequences to form a complete meteorological correction sequence. The integrated meteorological correction sequence is collaboratively analyzed to reveal the relationship and collaborative change trend between different meteorological factors, and to identify subtle potential anomalies in the collaborative analysis results. After in-depth analysis of these potential anomalies, corresponding optimization measures are taken to adjust them, ultimately ensuring the quality of the data and the efficiency of the plant. The technical solution of the present invention effectively solves the problems of inaccurate anomaly detection and untimely correction of traditional meteorological data, so that meteorological data can be quickly corrected and optimized under complex working conditions, thereby improving data reliability and enhancing the accuracy and efficiency of new energy plant operation management.

[0069] Furthermore, several meteorological factor data are preprocessed to obtain meteorological observation sequences, including:

[0070] Perform time homogenization processing on several meteorological factor data, unify the time base of several meteorological factors, and obtain meteorological pre-observation sequence;

[0071] According to the meteorological observation sequence, the missing values ​​of the meteorological observation sequence are extracted, and the missing values ​​are filled and marked to obtain the meteorological observation sequence.

[0072] As a preferred embodiment of the above embodiment, the collected data of several meteorological factors are time-homogenized, and a standard time base is determined by checking the timestamp of each data point, such as recording every hour, every day or every 15 minutes. If the time recording intervals of the original data are inconsistent, for example, some stations record every 10 minutes, while other stations record every 30 minutes, then the linear interpolation method can be used to convert the data to the selected time base. At the same time, linear interpolation can also be used to estimate the data values ​​of these time points and fill in the missing time points, and set an identifier for the filled missing values ​​in the data set, such as by adding a new column to mark whether these values ​​are original records or interpolated, so as to facilitate identification in the subsequent data analysis and quality control process.

[0073] Furthermore, data anomaly detection is performed on the meteorological observation sequence, including:

[0074] Collect historical meteorological observation information, and set temperature range thresholds and temperature change thresholds based on the historical meteorological observation information;

[0075] Based on the temperature range threshold, each observation record in the meteorological observation sequence is compared point by point. If the observation record is not within the temperature range threshold, it is marked as a range anomaly record;

[0076] According to the meteorological observation sequence, the difference between each observation record and the adjacent observation record is calculated point by point, and the calculated difference is compared with the temperature change threshold. If the observation record is not within the temperature change threshold, it is marked as an abnormal change record;

[0077] Among them, anomaly detection is also applicable to other meteorological factors.

[0078] In this embodiment, first, past meteorological observation data are collected from relevant meteorological databases. These data should include meteorological records for many years, including temperature records, humidity records, etc., and then the historical data are used to determine the normal range of temperature and the rate of temperature change. The maximum and minimum temperature values ​​are calculated through historical data, or climate statistical data are used to set a safe temperature range, such as ±5% of the minimum and maximum temperatures. Then, the rate of change of temperature over time in the historical data is analyzed to determine the threshold of abnormal changes, such as the temperature change per hour should not exceed 2°C. Subsequently, each observation record is traversed, and while comparing its temperature value, the temperature difference of the previous observation point is calculated to determine whether the observed temperature and temperature change are within the established temperature range or whether they exceed the set temperature change threshold. If the observed value exceeds this range, the data point is marked as a range abnormality. Similar methods are also applicable to other meteorological factors. It is only necessary to adjust the corresponding range threshold and change threshold for each factor.

[0079] Furthermore, data anomaly detection is performed on meteorological observation sequences, including:

[0080] Calculate the observation median of the meteorological observation sequence and obtain the observation time range of the meteorological observation sequence;

[0081] Calculate the standard deviation of any observation record based on historical meteorological observation information as the weight value of the observation record;

[0082] A double-weighted standard deviation test is performed based on the observation median, observation time range, and observation record weight value. The calculation formula is as follows:

[0083]

[0084] Among them, AVE bw is the double weighted average, STDbw is the double-weighted standard deviation, is the observed temperature of the nth observation record, M is the observation median, N is the observation time range, MAD is the median of the difference between the observation record and the observation median, is the weight value of the nth observation record. If |w n |>1.0, then w n =1.0,

[0085] When the weighted anomaly detection condition is met , it is marked as a weight abnormal record;

[0086] Integrate range anomaly records, change anomaly records and weight anomaly records to obtain comprehensive anomaly records;

[0087] Among them, anomaly detection is also applicable to other meteorological factors.

[0088] Specifically, first, sort the temperature or other meteorological factors of the entire meteorological observation sequence, calculate the median to represent the typical value of the sequence, and then determine the start and end points of the meteorological data collection to understand the time length. Then, for each data point of the historical meteorological observation information, calculate its standard deviation with the historical data. Based on the calculated standard deviation, assign a weight value to each observation record. The weight can be set based on the inverse of the standard deviation or other correlation measures. Then, by calculating the double-weighted average and double-weighted deviation values, the observation value is judged as abnormal. When there is an observation value that meets the above weighted detection conditions, it is marked as a weighted abnormal record. Similar methods are also applicable to other meteorological factors. It is only necessary to adjust the corresponding weights and standard deviations for each factor. Finally, the comprehensive range abnormality, change abnormality and weight abnormality records are integrated to form a comprehensive abnormality record. Further, the comprehensive abnormality record is data corrected to generate a meteorological replacement subsequence, including:

[0089] A cubic spline interpolation function is constructed based on the meteorological correction subsequence, and the cubic spline interpolation function is continuous on the meteorological correction subsequence;

[0090] Substitute the time point of the comprehensive anomaly record into the cubic spline interpolation function, calculate the anomaly interpolation, and obtain the meteorological replacement subsequence;

[0091] The meteorological replacement subsequence replaces the comprehensive anomaly record to generate a meteorological correction sequence.

[0092] As a preferred embodiment of the above embodiment, first, those meteorological data points that are not marked as abnormal are selected as reference points for cubic spline interpolation, and a cubic spline interpolation function is constructed. This function can provide a continuous and smooth data curve on the entire meteorological correction subsequence. Then, each time point in the identified comprehensive abnormal record is brought into the previously constructed cubic spline interpolation function, and an interpolation calculation is performed on each abnormal time point to obtain an estimated value. This estimated value is predicted based on the data trend of the surrounding non-abnormal data points. This can ensure the continuity and consistency of the data and reduce the impact of abnormal data on the final result. Subsequently, the interpolation results obtained from the cubic spline interpolation function are used to replace the comprehensive abnormal records in the original meteorological observation sequence. By replacing these abnormal values, a final meteorological correction sequence is generated. This correction sequence will integrate the non-abnormal original records and the newly calculated interpolation to form a complete meteorological data sequence with improved data quality and availability.

[0093] Furthermore, if Figure 2 As shown, the meteorological correction sequence is collaboratively analyzed, collaborative analysis results are obtained, potential anomalies are identified based on the collaborative analysis results, and potential anomalies are optimized, including:

[0094] S41: constructing a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series and the matrix columns represent meteorological factors;

[0095] S42: respectively calculating the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues, wherein the eigenvalues ​​are used to represent the weight ratios of the corresponding meteorological factors, and the eigenvectors are used to indicate the mutual relationships and change trends of the meteorological factors;

[0096] S43: setting a cumulative contribution threshold according to historical meteorological observation information, screening eigenvalues ​​and corresponding eigenvectors based on the cumulative contribution threshold, obtaining screening results, and forming a eigenvector matrix with the screened eigenvectors;

[0097] S44: Perform principal component analysis on the eigenvector matrix to obtain the temporal variation information of meteorological factors, and combine statistical anomaly detection methods to identify potential abnormal records and optimize the potential abnormal records.

[0098] In this embodiment, first, a covariance matrix is ​​constructed based on the meteorological correction sequence. In this matrix, the rows represent different time points of the time series, and the columns represent different meteorological factors, such as temperature, humidity, wind speed, etc. Then, the covariance matrix is ​​calculated to reveal the mutual relationship and dependence between different meteorological factors, and the covariance matrix is ​​subjected to eigenvalue decomposition to obtain eigenvalues ​​and corresponding eigenvectors. The eigenvalues ​​represent the importance of each eigenvector, and the eigenvectors describe the mutual relationship and change trend between the meteorological factors. Then, according to the historical meteorological observation information, a cumulative contribution threshold is set for screening important eigenvectors, which is generally set to 90%, meaning that the selected eigenvectors should be able to explain 90% of the original data. The eigenvectors whose cumulative contributions reach the set threshold are selected and combined into a new eigenvector matrix. The filtered eigenvector matrix is ​​used for principal component analysis. The principal component analysis can reveal the main change patterns of meteorological factors, which may not appear in the analysis of a single meteorological factor due to hidden environmental factors or other influences. According to the data obtained from the principal component analysis, statistical anomaly detection methods such as threshold determination and distance measurement are applied to identify potential abnormal records that deviate from the main change pattern, and these potential anomalies are further analyzed and optimized, such as data correction.

[0099] Furthermore, a covariance matrix is ​​constructed based on the meteorological correction sequence, including:

[0100] A meteorological matrix is ​​constructed according to the meteorological correction sequence, and the meteorological matrix is ​​standardized to obtain a meteorological standard matrix;

[0101] Combine each meteorological factor with all meteorological factors in a single cycle to generate several meteorological variable pairs. Perform covariance calculation on several meteorological variable pairs to obtain several covariances. The covariance calculation formula is as follows:

[0102]

[0103] Among them, z ij With z ik is the observation record of the jth and kth meteorological factors at the i-th time node, and is the observed average value of the jth and kth meteorological factors;

[0104] Construct a covariance matrix from several covariances.

[0105] Specifically, the data in the calibrated meteorological correction sequence are organized into a matrix, in which each column represents a different meteorological factor, such as temperature, humidity, wind speed, etc., and each row represents an observation record at a time point. Each pair of meteorological factors in the meteorological standard matrix is ​​combined in a single cycle, which means that each factor will be combined with all other factors including itself once to calculate the covariance between them. The covariance calculation formula is used to calculate the covariance between two variables to represent their covariance. All covariance values ​​are organized into a covariance matrix. In this matrix, each element represents the covariance between each two factors. The covariance matrix is ​​symmetrical, and the elements on the diagonal represent the variance of the meteorological factor itself. It should be noted that before constructing the covariance matrix, ensure that all variables are properly standardized to avoid the impact of dimensionality problems on the results. Further, the principal component analysis is performed on the eigenvector matrix to obtain the information on the coordinated changes of meteorological factors, including:

[0106] Perform matrix multiplication operation on the meteorological standard matrix and the eigenvector matrix to obtain the principal component analysis matrix, wherein each column in the principal component analysis matrix represents a principal component direction, and each row under the column represents the principal component score of the time node;

[0107] According to the principal component scores, the coordinated change patterns of meteorological factors at the same time node are identified, and the temporal change information of meteorological factors is obtained.

[0108] As a preferred embodiment of the above embodiment, matrix multiplication operation is performed on the meteorological standard matrix and the eigenvector matrix constructed in the covariance matrix of the data. In this step, the standardized values ​​of all meteorological factors at a time point represented by each row in the meteorological standard matrix will be multiplied by the eigenvector matrix to obtain the principal component analysis matrix. Each column in this matrix represents a principal component, and each row shows the score of a specific time node on these principal components. Then, the variables represented by each principal component are analyzed to understand which meteorological factors changes they each express. The rows of the principal component score matrix are observed and the principal component scores at different time points are analyzed. These scores show which change patterns dominate at different time points. For example, in the hot summer months, the principal component analysis may show a significant increase in the scores of high temperature and low humidity, indicating that these two meteorological factors show a clear synergistic upward trend during this period. Subsequently, by comparing the principal component scores at different time points, the changing trends and periodic patterns of the meteorological factors can be identified, and it is determined which meteorological factors tend to change together during specific climate events or seasonal changes. For example, high temperature may occur simultaneously with low humidity, and such a pattern can be identified through the principal component scores.

[0109] Embodiment 2:

[0110] Based on the same inventive concept as the EOF-based new energy plant meteorological data quality control method in the aforementioned embodiment, the present invention also provides an EOF-based new energy plant meteorological data quality control system, the system comprising:

[0111] The meteorological data preprocessing module collects meteorological observation information of new energy plants and stations, extracts meteorological factor data, and preprocesses some meteorological factor data to obtain meteorological observation sequences;

[0112] The abnormal data elimination module performs data anomaly detection on the meteorological observation sequence, determines and eliminates the comprehensive abnormal records, and obtains the meteorological correction subsequence; the data correction integration module performs data correction on the comprehensive abnormal records, generates a meteorological replacement subsequence, integrates the meteorological replacement subsequence and the meteorological correction subsequence, and obtains the meteorological correction sequence;

[0113] The collaborative anomaly optimization module performs collaborative analysis on the meteorological correction sequence, obtains collaborative analysis results, identifies potential anomalies based on the collaborative analysis results, and optimizes potential anomalies.

[0114] The above-mentioned adjustment system in the present invention can effectively realize a new energy plant meteorological data quality control method based on EOF, and the technical effects that can be achieved are as described in the above-mentioned embodiments and will not be repeated here.

[0115] Further, if Figure 3 As shown, the collaborative exception optimization module includes:

[0116] A covariance matrix construction unit, which constructs a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series and the matrix columns represent meteorological factors;

[0117] An eigenvalue vector analysis unit calculates the eigenvalue of the covariance matrix and the eigenvector corresponding to the eigenvalue, respectively. The eigenvalue is used to represent the weight ratio of the corresponding meteorological factor, and the eigenvector is used to indicate the relationship and change trend of the meteorological factors.

[0118] The eigenvector screening unit sets a cumulative contribution threshold according to historical meteorological observation information, screens eigenvalues ​​and corresponding eigenvectors based on the cumulative contribution threshold, obtains screening results, and forms a eigenvector matrix with the screened eigenvectors;

[0119] The principal component anomaly optimization unit performs principal component analysis on the eigenvector matrix to obtain the temporal variation information of meteorological factors, and combines statistical anomaly detection methods to identify potential anomaly records and optimize potential anomaly records.

[0120] Similarly, the above-mentioned optimization schemes for the system can also respectively achieve the optimization effects corresponding to the method in Example 1, which will not be repeated here.

[0121] Although the present application has been described in conjunction with specific features and embodiments thereof, it is obvious that various modifications and combinations may be made thereto without departing from the spirit and scope of the present application. Accordingly, this specification and the accompanying drawings are merely exemplary illustrations of the present application as defined therein, and are deemed to have covered any and all modifications, variations, combinations or equivalents within the scope of the present application. Obviously, a person skilled in the art may make various modifications and variations to the present application without departing from the scope of the present application. Thus, if these modifications and variations of the present application fall within the scope of the present application and its equivalents, the present application is intended to include these modifications and variations.

Claims

1. A meteorological data control method for new energy plants based on EOF, characterized in that: The method comprises: collecting meteorological observation information of a new energy plant, extracting meteorological factor data, and preprocessing a number of the meteorological factor data to obtain a meteorological observation sequence; Performing data anomaly detection on the meteorological observation sequence, determining and eliminating comprehensive abnormal records, and obtaining a meteorological correction subsequence; Performing data correction on the comprehensive anomaly record to generate a meteorological replacement subsequence, integrating the meteorological replacement subsequence with the meteorological correction subsequence to obtain a meteorological correction sequence; The meteorological correction sequence is collaboratively analyzed to obtain collaborative analysis results, potential anomalies are identified based on the collaborative analysis results, and the potential anomalies are optimized.

2. The EOF-based meteorological data control method for new energy plants according to claim 1 is characterized in that: Preprocessing the plurality of meteorological factor data to obtain a meteorological observation sequence further comprises: Performing time homogenization processing on the data of several meteorological factors, unifying the time bases of several meteorological factors, and obtaining a meteorological pre-observation sequence; The missing values ​​of the meteorological observation sequence are extracted according to the meteorological observation sequence, and the missing values ​​are filled and marked to obtain the meteorological observation sequence.

3. The EOF-based meteorological data control method for new energy plants according to claim 2 is characterized in that: Performing data anomaly detection on the meteorological observation sequence further includes: Collect historical meteorological observation information, and set a temperature range threshold and a temperature change threshold based on the historical meteorological observation information; compare each observation record of the meteorological observation sequence point by point based on the temperature range threshold, and if the observation record is not within the temperature range threshold, it is marked as a range abnormality record; Calculating the difference between each observation record and the adjacent observation record point by point according to the meteorological observation sequence, comparing the calculated difference with the temperature change threshold, and marking the observation record as an abnormal change record if the observation record is not within the temperature change threshold; The anomaly detection is also applicable to other meteorological factors.

4. The EOF-based meteorological data control method for new energy plants according to claim 3 is characterized in that: Performing data anomaly detection on the meteorological observation sequence further includes: Calculating the observation median of the meteorological observation sequence and obtaining the observation time range of the meteorological observation sequence; The standard deviation of any of the observation records is calculated based on the historical meteorological observation information as the observation record weight value; a double-weighted standard deviation test is performed based on the observation median, the observation time range, and the observation record weight value, and the calculation formula is as follows: Among them, AVE bw is the double weighted average, STD bw is the double-weighted standard deviation, is the observed temperature of the nth observation record, M is the observation median, N is the observation time range, MAD is the median of the difference between the observation record and the observation median, is the weight value of the nth observation record, if |w n |>1.0, then w n =1.0, When the weighted anomaly detection condition is met When , it is marked as a weight abnormal record; the range abnormal record, the change abnormal record and the weight abnormal record are integrated to obtain a comprehensive abnormal record; The anomaly detection is also applicable to other meteorological factors.

5. The EOF-based meteorological data control method for new energy plants according to claim 4 is characterized in that: Performing data correction on the comprehensive abnormal record to generate a meteorological replacement subsequence further includes: Constructing a cubic spline interpolation function based on the meteorological correction subsequence, wherein the cubic spline interpolation function is continuous on the meteorological correction subsequence; Substituting the time point of the comprehensive anomaly record into the cubic spline interpolation function, calculating the anomaly interpolation, and obtaining the meteorological replacement subsequence; The meteorological replacement subsequence replaces the comprehensive anomaly record to generate a meteorological correction sequence.

6. The EOF-based meteorological data control method for new energy plants according to claim 1 is characterized in that: Performing collaborative analysis on the meteorological correction sequence, obtaining collaborative analysis results, identifying potential anomalies based on the collaborative analysis results, and optimizing the potential anomalies, further comprising: Constructing a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series and the matrix columns represent meteorological factors; Calculating the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues ​​respectively, the eigenvalues ​​are used to represent the weight ratios corresponding to the meteorological factors, and the eigenvectors are used to indicate the mutual relationships and change trends of the meteorological factors; A cumulative contribution threshold is set according to historical meteorological observation information, the eigenvalue and the corresponding eigenvector are screened based on the cumulative contribution threshold, a screening result is obtained, and the screened eigenvectors are combined into an eigenvector matrix; The principal component analysis is performed on the characteristic vector matrix to obtain the time series change information of the meteorological factors, and the statistical anomaly detection method is combined to identify potential abnormal records and optimize the potential abnormal records.

7. The EOF-based meteorological data control method for new energy plants according to claim 6 is characterized in that: Constructing a covariance matrix based on the meteorological correction sequence further includes: constructing a meteorological matrix according to the meteorological correction sequence, and performing standardization processing on the meteorological matrix to obtain a meteorological standard matrix; Each of the meteorological factors is combined with all the meteorological factors in a single cycle to generate a number of meteorological variable pairs, and covariance calculation is performed on the meteorological variable pairs to obtain a number of covariances. The covariance calculation formula is as follows: Among them, z ij With z ik is the observation record of the jth and kth meteorological factors at the i-th time node, and is the observed average value of the jth and kth meteorological factors; A covariance matrix is ​​constructed based on the plurality of covariances.

8. The EOF-based meteorological data control method for new energy plants according to claim 7 is characterized in that: Performing principal component analysis on the eigenvector matrix to obtain information on coordinated changes in meteorological factors further includes: Performing a matrix multiplication operation on the meteorological standard matrix and the eigenvector matrix to obtain a principal component analysis matrix, wherein each column in the principal component analysis matrix represents a principal component direction, and each row under the column represents a principal component score of a time node; According to the principal component scores, the coordinated change patterns of meteorological factors at the same time node are identified, and the time series change information of meteorological factors is obtained.

9. A meteorological data quality control system for new energy plants based on EOF, characterized in that: The system comprises: A meteorological data preprocessing module collects meteorological observation information of new energy plants and stations, extracts meteorological factor data, and preprocesses some of the meteorological factor data to obtain a meteorological observation sequence; An abnormal data elimination module performs data anomaly detection on the meteorological observation sequence, determines and eliminates comprehensive abnormal records, and obtains a meteorological correction subsequence; A data correction and integration module performs data correction on the comprehensive abnormality record, generates a meteorological replacement subsequence, and integrates the meteorological replacement subsequence with the meteorological correction subsequence to obtain a meteorological correction sequence; The collaborative anomaly optimization module collaboratively analyzes the meteorological correction sequence, obtains collaborative analysis results, identifies potential anomalies based on the collaborative analysis results, and optimizes the potential anomalies.

10. The EOF-based new energy plant meteorological data quality control system according to claim 9 is characterized in that: The collaborative exception optimization module includes: A covariance matrix construction unit, which constructs a covariance matrix based on the meteorological correction sequence, wherein the matrix rows of the covariance matrix represent time series and the matrix columns represent meteorological factors; an eigenvalue vector analysis unit, for respectively calculating the eigenvalues ​​of the covariance matrix and the eigenvectors corresponding to the eigenvalues, wherein the eigenvalues ​​are used to represent the weight ratios corresponding to the meteorological factors, and the eigenvectors are used to indicate the mutual relationships and change trends of the meteorological factors; A feature vector screening unit, which sets a cumulative contribution threshold according to historical meteorological observation information, screens the feature values ​​and corresponding feature vectors based on the cumulative contribution threshold, obtains screening results, and forms a feature vector matrix with the screened feature vectors; The principal component anomaly optimization unit performs principal component analysis on the eigenvector matrix to obtain the time series change information of the meteorological factors, and combines the statistical anomaly detection method to identify potential anomaly records and optimize the potential anomaly records.

Images