A wind resource correction method based on meteorological assimilation

Abstract

The application discloses a wind resource correction method based on meteorological assimilation, and relates to the technical field of wind resource correction, in particular to a wind resource correction method based on meteorological assimilation.

Description

TECHNICAL FIELD

[0001] The application belongs to the technical field of wind resource correction, and particularly relates to a wind resource correction method based on meteorological assimilation. BACKGROUND

[0002] With the advent of the era of parity and large base, the wind power industry is constantly optimizing the model configuration and arrangement scheme of the wind farm in order to obtain the optimal power generation. After the natural wind speed of the wind farm passes through the upwind unit, the wind speed is obviously reduced, the turbulence intensity is obviously increased, the wake effect is produced, the utilization rate of wind resources of most wind farms is not high enough, the actual power generation curve of the wind turbine deviates from the standard curve, and the annual power generation is lower than the design value. Therefore, it is particularly important to accurately and scientifically analyze the influence of the wind resources of the wind farm on the power generation performance. In addition, in the aspects of improving the quality and efficiency of the wind farm equipment and improving the operation safety, the following demands and challenges are faced: (1) there is a lack of effective means for improving the quality and efficiency of the wind turbine, and the technology is mainly controlled by the wind turbine manufacturers; (2) it is difficult to obtain the real wind resource data of the wind farm wind turbine, and the current wind speed correction algorithm and the influence of the nacelle wake on the wind direction measurement make it difficult to carry out quantitative analysis and evaluation; (3) the control strategy of the wind turbine operation is less controlled, especially the effective preventive measures for the operation risk and fault hidden danger of the wind turbine under extreme working conditions;

[0003] In order to accurately and scientifically analyze the influence of the wind resources of the wind farm on the power generation performance, the wind resources need to be corrected first. At present, the wind resource correction methods mainly include: (1) a wind speed correction method based on the OS-ELM algorithm, which adopts the extreme learning machine algorithm to correct the wind speed of the numerical weather prediction, and carries out secondary correction on the predicted power according to the confidence interval of the wind power, but the extreme learning machine is only suitable for single hidden layer neural network, and the application scene is less; (2) a nacelle wind speed correction method based on convolutional neural network, which adopts multi-layer convolution pooling to filter the influence of the wind turbine wake and blade disturbance, highly abstracts the feature variables, and improves the accuracy of the corrected wind speed, but the requirement for the calculation server is high, and it is not suitable for wide use; (3) a wind speed error correction method based on Markov chain, which carries out Markov wind speed error correction on the error sequence based on the NWP wind speed value and the SCADA wind speed value, first divides the wind speed error sequence into model C mean states to calculate the probability distribution of the initial state, then establishes the transition probability matrix between states, predicts the next time wind speed error correction value according to the transition probability matrix, and further obtains the corrected wind speed value, but the Markov chain makes an independence assumption, and the correlation of the meteorological data before and after the time is not considered.

[0004] Therefore, how to correct the wind resources on the basis of considering the correlation of the meteorological data before and after the time, expand the application scene of the correction method, and reduce the requirement for the calculation server has become a key problem in the current research. SUMMARY

[0005] In view of the above problems, the present application provides a meteorological assimilation-based wind resource correction method which at least solves part of the above technical problems.

[0006] The present application provides a meteorological assimilation-based wind resource correction method, comprising:

[0007] S1, obtaining wind resource meteorological data to be corrected through a SCADA system;

[0008] S2, preprocessing the wind resource meteorological data;

[0009] S3, inputting the preprocessed wind resource meteorological data into a wind resource correction model to output corrected wind resource meteorological data.

[0010] Further, the wind resource meteorological data includes wind speed data and wind direction data.

[0011] Further, S2 specifically includes converting the format of the wind resource meteorological data to a preset format and excluding data of unqualified quality.

[0012] Further, in S3, the wind resource correction model is generated by training through the following steps, comprising:

[0013] A large amount of wind resource meteorological data in the same time period is obtained through a laser radar and a SCADA system respectively, denoted as first meteorological data and second meteorological data;

[0014] The first meteorological data and the second meteorological data are preprocessed respectively;

[0015] The preprocessed first meteorological data is corrected based on a meteorological assimilation method;

[0016] The preprocessed second meteorological data is taken as input, and the corrected corresponding first meteorological data is taken as output to train and generate a wind resource correction model.

[0017] Further, the wind resource correction model includes a wind speed correction model and a wind direction correction model;

[0018] The wind speed data in the preprocessed second meteorological data is taken as input, and the corresponding wind speed data in the corrected first meteorological data is taken as output to train and generate the wind speed correction model;

[0019] The wind direction data in the preprocessed second meteorological data is taken as input, and the corresponding wind direction data in the corrected first meteorological data is taken as output to train and generate the wind direction correction model.

[0020] Further, a second-order function algorithm is adopted as the algorithm model of the wind speed correction model or the wind direction correction model, which is expressed as:

[0021] y=w1x 2 +w0

[0022] wherein w0 and w1 represent regression coefficients; x represents an input value; and y represents an output value.

[0023] Compared with the prior art, the wind resource correction method based on meteorological assimilation has the following beneficial effects:

[0024] The method provided by the embodiment of the present application can correct and predict the wind speed measured at the wind power plant, can effectively reduce the influence of unstable wind power generation output on the power grid, realize intelligent load reduction of the wind turbine, improve the safety and reliability of the wind turbine operation, provide a data basis for the control strategy, can provide a power selling quotation guidance for the wind power plant, and reasonably arrange a maintenance plan.

[0025] Other features and advantages of the present application will be set forth in the following description, and in part will become apparent to those skilled in the art from the description, or can be learned by practice of the present application. The objects and other advantages of the present application will be realized and attained by the structure particularly pointed out in the written description and claims hereof as well as the appended drawings.

[0026] The technical solutions of the present application will be described in further detail below with the aid of the accompanying drawings and embodiments. BRIEF DESCRIPTION OF DRAWINGS

[0027] The accompanying drawings are included to provide a further understanding of the present application, and constitute a part of the specification, illustrate the present application and are used to explain the present application together with the embodiments of the present application, and do not constitute a limitation on the present application. In the drawings:

[0028] Figure 1 The flow chart of the wind resource correction method based on meteorological assimilation provided by the embodiment of the present application.

[0029] Figure 2 The flow chart of the wind resource correction model training provided by the embodiment of the present application. DETAILED DESCRIPTION

[0030] Exemplary embodiments of the present disclosure will be described more fully hereinafter with reference to the accompanying drawings, in which exemplary embodiments of the present disclosure are shown. It is to be understood that the present disclosure can be embodied in various forms and should not be limited by the embodiments set forth herein. Rather, these embodiments are provided so that this disclosure will be thorough and complete, and will fully convey the scope of the present disclosure to those skilled in the art.

[0031] Referring to Figure 1 As shown in the figure, the embodiment of the present application provides a wind resource correction method based on meteorological assimilation, which specifically includes the following steps:

[0032] S1, obtaining wind resource meteorological data to be corrected through a SCADA system;

[0033] S2, preprocessing the wind resource meteorological data;

[0034] S3, inputting the preprocessed wind resource meteorological data into a wind resource correction model to output corrected wind resource meteorological data.

[0035] The above steps will be described in detail below.

[0036] In the above step S1, the wind resource meteorological data includes wind speed data and wind direction data.

[0037] In the above step S2, the wind resource meteorological data is preprocessed, that is, the obtained wind resource meteorological data is normalized, including planning of data accuracy and type for convenient later model processing; specifically, the format of the wind resource meteorological data is converted into a preset format, and data with unqualified quality is excluded; the data with unqualified quality includes wind resource meteorological data with serious information missing and inaccurate wind resource meteorological data measured after the anemometer is used for a long time.

[0038] In the above step S3, in the embodiment of the present application, high-precision meteorological data is collected through meteorological assimilation technology, and a wind resource correction model is established combining with SCADA wind resource data to fit the current optimal value, thereby realizing wind resource correction and improving the efficiency of wind turbine control strategy and power generation performance; referring to Figure 2 As shown in the figure, the wind resource correction model is trained and generated through the following steps, including:

[0039] First, a large amount of wind resource meteorological data in a preset time period is obtained through a laser radar, denoted as first meteorological data; at the same time, a large amount of wind resource meteorological data in the same time period is obtained through a SCADA system, denoted as second meteorological data;

[0040] Secondly, the first meteorological data and the second meteorological data obtained are preprocessed; wherein when the first meteorological data obtained by the laser radar is preprocessed, the first meteorological data observed by the laser radar is mainly processed into a data format that can be input into the data assimilation system, and data quality control is needed to exclude data with poor quality in the first meteorological data; when the second meteorological data obtained by the SCADA system is preprocessed, the second meteorological data obtained by the SCADA system is mainly processed into a data format that can be input into the data assimilation system, and data quality control is needed to exclude data with poor quality in the second meteorological data; wherein the data with poor quality is the above unqualified data.

[0041] Thirdly, the first meteorological data after preprocessing is corrected based on the meteorological assimilation method, mainly using the meteorological assimilation technology to correct the data of the laser wind measuring radar; the meteorological assimilation technology takes the analysis field in the meteorological model as the assimilated background field, builds a mesoscale meteorological numerical simulation model and a microscale wind farm atmospheric state simulation model, and builds a corresponding data assimilation system to fuse the observation data with the background field, build a mesoscale meteorological numerical simulation model and a microscale wind farm atmospheric state simulation model; and a corresponding data assimilation system is built to make super-short-term prediction of the wind speed and direction at each wind turbine in the wind farm, so as to obtain a more accurate model analysis field of the meteorological model. In the embodiment of the present application, the first meteorological data after preprocessing is corrected based on the meteorological assimilation method, specifically: using the data obtained by the ground laser radar as the entrance wind profile analysis data; using the mesoscale numerical simulation prediction system WRF as the entrance boundary condition for OpenFOAM; the influence of wind power generation simulated by OpenFOAM on the weather system will be fed back to the mesoscale numerical simulation prediction system WRF through the upscaling technology, and the precise prediction of wind conditions is realized in combination with the wind measuring characteristics of the laser radar.

[0042] Finally, the second meteorological data after preprocessing is taken as input, and the corresponding first meteorological data after correction is taken as output to train a wind resource correction model; specifically: the wind resource correction model includes a wind speed correction model and a wind direction correction model; the algorithm of the correction model adopts a regression model to fit the wind speed and wind direction data to approach the correct value, and obtain the parameters optimized by the best fitting training data; the wind speed data in the second meteorological data after preprocessing is taken as input, and the corresponding wind speed data in the first meteorological data after correction is taken as output to train the wind speed correction model; the wind direction data in the second meteorological data after preprocessing is taken as input, and the corresponding wind direction data in the first meteorological data after correction is taken as output to train the wind direction correction model.

[0043] Considering that a unary linear function is too simple, the model will be limited; therefore, in the embodiment of the application, the order of the polynomial is increased, and a quadratic function algorithm is selected as the algorithm model for wind speed and wind direction correction, which is expressed as:

[0044] y = w1x 2 +w0

[0045] wherein w0 and w1 both represent regression coefficients; x represents an input value; and y represents an output value. That is, in the process of training the wind speed correction model, x represents the wind speed data in the preprocessed second meteorological data; and y represents the corresponding wind speed data in the corrected first meteorological data. In the process of training the wind direction correction model, x represents the wind direction data in the preprocessed second meteorological data; and y represents the corresponding wind direction data in the corrected first meteorological data.

[0046] Through the above wind speed correction model and wind direction correction model, the wind speed and wind direction data in the SCADA system are corrected, the wind speed measured by the wind farm is corrected and predicted, the influence of unstable wind power generation output on the power grid can be effectively reduced, the intelligent load reduction of the wind turbine is realized, the safety and reliability of the wind turbine operation are improved, a data basis is provided for the control strategy, and the wind farm can be provided with a power sale quotation guidance and a reasonable maintenance plan.

[0047] In the embodiment of the application, the least square method or the normal equation is used to solve the regression coefficients, and the determination coefficient R 2 The prediction error of the regression model is compared with the error of the sample and the average value, which is expressed as:

[0048]

[0049] wherein, represents a predicted value; represents a sample average value; and R 2 The greater the value is, the smaller the residual error is, and the better the regression fitting effect is. When the regression fitting effect of the wind speed correction model is evaluated, y i represents the corresponding wind speed data in the corrected first meteorological data.

[0050] The wind resource correction model provided in the embodiment of the application is constructed by combining advanced technologies such as laser radar, meteorological assimilation and big data intelligent mining; based on this, an integrated platform for high-precision collection, correction and prediction of wind resource data can be further built, so that the wind energy absorption efficiency is maximized.

[0051] Obviously, many modifications and variations of the present application are possible in light of the above teachings. It is, therefore, to be understood that within the scope of the appended claims and their equivalents, the application can be practiced otherwise than as specifically described.

Claims

1. A wind resource correction method based on meteorological assimilation, characterized in that, include: S1. Obtain meteorological data of wind resources to be corrected through the SCADA system; S2. Preprocess the wind resource meteorological data; S3. Input the preprocessed wind resource meteorological data into the wind resource correction model and output the corrected wind resource meteorological data. In S3, the wind resource correction model is trained and generated through the following steps: A large amount of wind resource meteorological data within the same time period was acquired using lidar and SCADA systems respectively, and recorded as the first meteorological data and the second meteorological data. The first and second meteorological data are preprocessed respectively. The preprocessed first meteorological data was corrected based on the meteorological assimilation method; The preprocessed second meteorological data is used as input, and the corrected corresponding first meteorological data is used as output to train and generate a wind resource correction model. The correction of the preprocessed first meteorological data based on the meteorological assimilation method specifically includes: using the first meteorological data acquired by lidar as inlet wind profile analysis data; providing inlet boundary conditions for OpenFOAM based on a mesoscale numerical simulation forecasting system; simulating the impact of wind power generation on the weather system through OpenFOAM, and using upscaling technology to feed this impact back to the mesoscale numerical simulation forecasting system to achieve accurate wind condition forecasting and obtain the corrected first meteorological data.

2. The wind resource correction method based on meteorological assimilation as described in claim 1, characterized in that, The wind resource meteorological data includes wind speed data and wind direction data.

3. The wind resource correction method based on meteorological assimilation as described in claim 1, characterized in that, S2 specifically includes: converting the wind resource meteorological data into a preset format and excluding data that is of unqualified quality.

4. The wind resource correction method based on meteorological assimilation as described in claim 1, characterized in that, The wind resource correction model includes a wind speed correction model and a wind direction correction model; The wind speed data in the preprocessed second meteorological data is used as input, and the corresponding wind speed data in the corrected first meteorological data is used as output to train and generate the wind speed correction model. The wind direction correction model is trained by taking the wind direction data from the preprocessed second meteorological data as input and the corresponding wind direction data from the corrected first meteorological data as output.

5. The wind resource correction method based on meteorological assimilation as described in claim 4, characterized in that, The second-order function algorithm is used as the algorithm model for the wind speed correction model or wind direction correction model, and is expressed as follows: y=w1x 2 +w0 Where w0 and w1 both represent regression coefficients; x represents the input value; and y represents the output value.

Images