Method and system for selecting key installation points of a wind farm machine cabin type laser radar, computer and storage medium

Abstract

This invention discloses a method, system, computer, and storage medium for selecting key installation sites for nacelle-type lidar in wind farms. Based on SCADA data, wind speed and direction vectors are constructed to determine the main inflow wind direction. The correlation between the turbine units is evaluated using dot product, correlation coefficient, and random forest models, generating three ranking schemes. Then, the consistency index between the secondary and main inflows is calculated and weighted for comprehensive evaluation, thereby selecting the optimal installation site. This invention improves the scientific rigor and accuracy of nacelle-type lidar installation site selection.

Description

Technical Field

[0001] This invention belongs to the field of wind farm data analysis, specifically relating to a method, system, computer, and storage medium for selecting key installation sites of a wind farm nacelle-type lidar. Background Technology

[0002] In large wind farms, the wake effects between wind turbines are significant. To study the flow field characteristics in wind farms and monitor the incoming wind conditions of wind turbines, nacelle-mounted lidar is typically used to acquire data. Nacelle-mounted lidar can measure wind speed and direction values ​​from tens to hundreds of meters in front of the rotor plane, serving as input values ​​for wind farm power prediction models and wake control strategies. This not only improves the accuracy of model predictions but also ensures the effectiveness of wake control strategies. Although the cost of a single nacelle-mounted lidar unit is not high, the sheer number of wind turbines in a large wind farm makes installing nacelle-mounted lidar on all turbines extremely costly. Furthermore, the installation and subsequent maintenance of nacelle-mounted lidar require considerable expenses. To reduce the cost of nacelle-mounted lidar while maximizing the accuracy of the wind speed and direction values ​​input to the model, it is necessary to study methods for selecting nacelle-mounted lidar installation sites by combining wind speed and direction data from each wind turbine. Summary of the Invention

[0003] The purpose of this invention is to provide a method, system, computer, and storage medium for selecting key installation sites of a wind farm nacelle-type lidar.

[0004] The technical solution to achieve the purpose of this invention is: a method for selecting key installation sites for a wind farm nacelle-type lidar, the specific steps of which are as follows:

[0005] Step 1: Collect data for each wind turbine using the wind farm's SCADA system, including sampling time, wind speed, wind direction, and power, and filter out invalid data;

[0006] Step 2: Establish a Cartesian coordinate system with due north as 0°, construct wind speed and wind direction vectors with wind speed as the modulus and wind direction as the angle, and calculate the coordinate components of the wind speed and wind direction vectors of each wind turbine in the Cartesian coordinate system.

[0007] Step 3: For a single sampling time, superimpose the coordinate components of the wind speed and direction vectors of all wind turbines and take an arithmetic average to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time, and further calculate the wind speed and direction of the main inflow of the wind farm.

[0008] Step 4: Calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Take the arithmetic mean of the point integral of each wind turbine at all sampling times to obtain the comprehensive point integral of each wind turbine. This comprehensive point integral is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and obtain the correlation ranking 1.

[0009] Step 5: Calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the correlation coefficient of each wind turbine at all sampling times is used to obtain the comprehensive correlation coefficient of each wind turbine, which is used as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and the correlation ranking 2 is obtained.

[0010] Step 6: Using the wind speed and direction vectors of the main inflow of the wind farm for the first 80% of all sampling times as the training set, input the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm. Combined with the SHAP value analysis method, calculate the SHAP value of the prediction result of the coordinate components of the wind speed and direction vector of each wind turbine to the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the SHAP value of each wind turbine at all sampling times is used to obtain the comprehensive SHAP value of each wind turbine, which is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and obtain the correlation ranking 3.

[0011] Step 7: Based on the required number of nacelle-type lidar installations, and using the three rankings obtained above as a basis, select the top m wind turbine units in each of the three rankings as the three options for selecting nacelle-type lidar installation sites.

[0012] Step 8: For each scheme, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, calculate the wind speed and direction vector of the secondary inflow of the wind farm for that scheme, and then calculate the consistency index between the wind speed and direction vector of the secondary inflow of the wind farm and the wind speed and direction vector of the main inflow of the wind farm. Add the weight values ​​of wind speed and direction to calculate the comprehensive evaluation index of the scheme at each sampling time, and arithmetically average the comprehensive evaluation index of all sampling times to obtain the final evaluation index of the scheme.

[0013] Step 9: Compare the final evaluation indicators of the three schemes and select the scheme with the highest final evaluation indicator as the installation site scheme for the cabin-type lidar.

[0014] A key site selection system for wind farm nacelle lidar installation includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the key site selection method for wind farm nacelle lidar installation.

[0015] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method for selecting key installation sites for a wind farm nacelle-type lidar.

[0016] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for selecting key installation sites for a wind farm nacelle-type lidar.

[0017] Compared with the prior art, the significant advantages of this invention are:

[0018] 1) The wind speed and direction of a single wind turbine are collected using a SCADA system, converted into coordinate components in a Cartesian coordinate system with 0° north, and then superimposed and arithmetically averaged to obtain the wind speed and direction of the main inflow into the wind farm. By utilizing the measured data of a single wind turbine, an effective and accurate characterization of the overall inflow characteristics of the wind farm is achieved, providing a reliable benchmark for subsequent site selection;

[0019] 2) A multi-dimensional correlation evaluation system is constructed using point integrals, correlation coefficients, and SHAP values ​​to quantify the correlation between the wind speed and direction vectors of a single wind turbine and the wind speed and direction vectors of the main inflow of the wind farm, and to obtain a ranking. Through cross-validation using multiple methods, the limitations of a single evaluation index are avoided, achieving high robustness and high reliability in the selection of nacelle-type lidar installation sites.

[0020] 3) Based on the installation requirements of the nacelle-type lidar, wind turbines with high relevance rankings are selected as installation sites. The wind speed and direction vectors of the secondary inflow of the wind farm are calculated based on their wind speed and direction vectors. The consistency index between the wind speed and direction vectors of the secondary inflow of the wind farm and the wind speed and direction vectors of the main inflow of the wind farm is used as the final evaluation criterion. A weighting coefficient is introduced to comprehensively consider the differences in engineering weights of wind speed and direction, which can flexibly adapt to different scenario requirements and significantly improve the scientificity and accuracy of the selection of nacelle-type lidar installation sites. Attached Figure Description

[0021] Figure 1 A flowchart illustrating the method for selecting key installation sites for nacelle-type lidar in wind farms.

[0022] Figure 2 This is a layout diagram of wind turbine units in a wind farm according to an embodiment of the present invention.

[0023] Figure 3 This is a schematic diagram of the installation sites for the five cabin-type lidar units ultimately selected in this embodiment of the invention. Detailed Implementation

[0024] To make the objectives, technical solutions, and advantages of this application clearer, the following detailed description is provided in conjunction with the accompanying drawings and embodiments. It should be understood that the specific embodiments described herein are merely illustrative and not intended to limit the scope of this application.

[0025] like Figure 1 As shown, a method for selecting key installation sites for a wind farm nacelle-type lidar is described, with the following specific steps:

[0026] In step 1, data for each wind turbine is collected using the wind farm's SCADA system, and invalid data is filtered out. The specific method is as follows:

[0027] The SCADA system is used to collect wind turbine data every 10 minutes, including sampling time, wind speed, wind direction, and power.

[0028] Remove data with power less than 0 from the SCADA data and align the remaining data to the sampling time.

[0029] Determine whether data for all wind turbine units are contained at a given sampling time. If not, remove all data for that sampling time.

[0030] Step 2: Establish a Cartesian coordinate system with true north as 0°. Construct wind speed and wind direction vectors using wind speed as the modulus and wind direction as the angle. Calculate the coordinate components of the wind speed and wind direction vectors for each wind turbine in the Cartesian coordinate system. The specific formula is as follows:

[0031]

[0032] Where i represents the i-th wind turbine in the wind farm, and t represents the t-th sampling time out of all sampling times. Let be the wind speed of wind turbine i at sampling time t. Let t be the wind direction of wind turbine i at sampling time t. Let be the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t in the Cartesian coordinate system.

[0033] Step 3: For a single sampling time, superimpose the coordinate components of the wind speed and direction vectors of all wind turbines and take an arithmetic mean to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time, as well as the wind speed and direction of the main inflow of the wind farm. The specific method is as follows:

[0034] Within a single sampling time t, the coordinate components of the wind speed and direction vectors of all wind turbines are superimposed and arithmetically averaged to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time t. The specific formula is as follows:

[0035]

[0036] in, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t. These are the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at sampling time t;

[0037] Further calculations of the wind speed and direction of the main inflow to the wind farm are expressed as follows:

[0038]

[0039]

[0040]

[0041] in, It represents the wind direction of wind turbine i at sampling time t. It refers to the number of wind turbines in the wind farm. It is the wind speed of the main inflow into the wind farm at sampling time t. The wind direction (range) of the main inflow to the wind farm at sampling time t before the angle conversion. ), It is the wind direction (range) of the main inflow of the wind farm at sampling time t after angle conversion. ).

[0042] Step 4: Calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Then, arithmetically average the point integral of each wind turbine across all sampling times to obtain the comprehensive point integral of each wind turbine. This comprehensive point integral serves as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and a correlation ranking is obtained. The specific method is as follows:

[0043] First, calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The formula is:

[0044]

[0045] In the formula, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t. and These are the coordinate components of the wind speed and direction vector of the main inflow into the wind farm at sampling time t. It is the point integral of wind turbine i at sampling time t.

[0046] Then, the arithmetic mean of the point integrals for each wind turbine at all sampling times is calculated to obtain the comprehensive point integral for each wind turbine. This comprehensive point integral is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. The formula is as follows:

[0047]

[0048] In the formula, The total number of sampling times. Let 1 be the comprehensive point integral of wind turbine i, which is the correlation index.

[0049] Finally, the integral of the comprehensive points is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking is obtained. .

[0050] Step 5: Calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Then, arithmetically average the correlation coefficient of each wind turbine at all sampling times to obtain the comprehensive correlation coefficient of each wind turbine. This comprehensive correlation coefficient serves as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and a correlation ranking 2 is obtained. The specific method is as follows:

[0051] First, calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The formula is:

[0052]

[0053]

[0054]

[0055]

[0056] In the formula, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t, where T represents all sampling times. and It is the average coordinate component of the wind speed and direction vector of wind turbine i over all sampling times. and These are the coordinate components of the wind speed and direction vector of the main inflow wind of the wind farm at sampling time t. and It is the average value of the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at all sampling times. and It is the correlation coefficient between the wind speed and direction vector of wind turbine i at sampling time t and the wind speed and direction vector of the main inflow of the wind farm;

[0057] Then, the correlation coefficient of each wind turbine is calculated by arithmetic mean at all sampling times, thus obtaining the comprehensive correlation coefficient of each wind turbine, as shown in the formula:

[0058]

[0059]

[0060] In the formula, It is the total correlation coefficient of the wind speed and wind direction vectors of wind turbine i at sampling time t. is the comprehensive correlation coefficient of wind turbine i, and T is the number of all sampling times;

[0061] Finally, the comprehensive correlation coefficient is used as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking 2 is obtained. .

[0062] Step 6: Using the wind speed and direction vectors of the main inflow of the wind farm from the first 80% of all sampling times as the training set, input the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm. Combined with the SHAP value analysis method, calculate the SHAP value of the coordinate components of the wind speed and direction vector of each wind turbine to the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the SHAP value of each wind turbine at all sampling times is used to obtain the comprehensive SHAP value of each wind turbine, which is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and the correlation ranking 3 is obtained. The specific method is as follows:

[0063] First, the wind speed and direction vectors of the main inflow of the wind farm in the previous 80% of all sampling times are used as the training set input to the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm.

[0064] Then, using the coordinate components of the wind speed and direction vectors of the main inflow of the wind farm at the last 20% of all sampling times as the test set, the performance of the wind speed and direction vector prediction model for the main inflow of the wind farm was evaluated. If the model If the value is greater than 0.9, then the SHAP value of the predicted coordinate components of the wind speed and direction vectors of each wind turbine is calculated based on the trained model, and the total SHAP value of wind turbine i at sampling time t is calculated based on the sum of squares formula, which is:

[0065]

[0066] In the formula, and It is the SHAP value of the prediction result of the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t against the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. It is the total SHAP value of wind turbine i at sampling time t;

[0067] Secondly, the SHAP value of each wind turbine is calculated by arithmetic mean at all sampling times, thus obtaining the comprehensive SHAP value of each wind turbine. The formula is as follows:

[0068]

[0069] In the formula, T represents the total number of sampling times. This represents the overall SHAP value of wind turbine i.

[0070] Finally, the comprehensive SHAP value is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking 3 is obtained. .

[0071] Step 7: Based on the required number of nacelle-type lidar installations, and using the three rankings obtained above as a basis, select the top m wind turbine units from the three rankings as the three options for selecting nacelle-type lidar installation sites.

[0072] Step 8: For each scheme, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, calculate the wind speed and direction vectors of the secondary inflow of the wind farm for that scheme. Then, calculate the consistency index between the wind speed and direction vectors of the secondary inflow and the wind speed and direction vectors of the main inflow of the wind farm. Add the weight values ​​of wind speed and direction to calculate the comprehensive evaluation index of the scheme at each sampling time. Arithmetically average the comprehensive evaluation index of all sampling times to obtain the final evaluation index of the scheme. The specific method is as follows:

[0073] First, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, the wind speed and direction vector of the secondary inflow to the wind farm in this scheme is calculated using the following formula:

[0074]

[0075]

[0076]

[0077]

[0078] In the formula, p represents one of the three schemes. Let m be the m wind turbines selected as installation sites for the nacelle-type lidar, and j be one of the m wind turbines. and These are the coordinate components of the wind speed and direction vector of wind turbine j at sampling time t. These are the coordinate components of the wind speed and direction vector of the secondary inflow into the wind farm at sampling time t, corresponding to the scheme. It is the wind speed of wind turbine j at sampling time t. It represents the wind speed of the secondary inflow into the wind farm at sampling time t, corresponding to the scheme. This refers to the wind direction (range) of the secondary inflow at the wind farm at sampling time t before the angle conversion, corresponding to the scheme. ), This refers to the wind direction (range) of the secondary inflow at sampling time t after the angle conversion of the corresponding scheme. );

[0079] Secondly, combining the coordinate components of the wind speed and direction vector of the main inflow of the wind farm, the coordinate components and point integral of the wind speed and direction vector of the secondary inflow of the wind farm are calculated. Based on the wind speed and direction of the main inflow and the secondary inflow of the wind farm, a consistency index in wind speed and direction is calculated. The formula is as follows:

[0080]

[0081]

[0082]

[0083] In the formula, It is the dot product of the coordinate components of the wind speed and direction of the main inflow to the wind farm and the coordinate components of the wind speed and direction of the secondary inflow to the wind farm. and These are the coordinate components of the wind speed and direction vector of the main inflow into the wind farm at sampling time t. and These are the coordinate components of the wind speed and direction vector of the secondary inflow into the wind farm at sampling time t for scheme p. It is the wind speed of the main inflow into the wind farm at sampling time t. It is the wind speed of the secondary inflow into the wind farm at sampling time t for scheme p. and These are the consistency in terms of wind direction and wind speed between the secondary inflow wind speed and wind direction vector and the main inflow wind speed and wind direction vector of the wind farm at sampling time t of scheme p.

[0084] Finally, weighted values ​​are added, and a comprehensive evaluation index is calculated for scheme p at sampling time t, combining the wind speed and direction vectors of the secondary inflow and the main inflow. The arithmetic mean of these vectors yields the final evaluation index for scheme p. Schemes with the highest final evaluation index are ranked from largest to smallest, and the scheme with the highest ranking is selected as the final choice for the nacelle-type lidar installation site. The formula is:

[0085]

[0086]

[0087] In the formula, The weighted value for wind direction. The weighted values ​​for wind speed are as follows: , It is the comprehensive correlation index of scheme p at sampling time t. It represents the total number of sampling times. It is the final evaluation index for scheme p.

[0088] In step 9, the final evaluation indicators of the three schemes are compared, and the scheme with the highest final evaluation indicator is selected as the scheme for the installation site of the cabin-type lidar.

[0089] The present invention also proposes a key site selection system for wind farm nacelle lidar installation, including a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the aforementioned key site selection method for wind farm nacelle lidar installation.

[0090] A computer device includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method for selecting key installation sites for a wind farm nacelle-type lidar.

[0091] A computer-readable storage medium storing a computer program, which, when executed by a processor, implements the method for selecting key installation sites for a wind farm nacelle-type lidar.

[0092] Example

[0093] To verify the effectiveness of the present invention, five locations were selected in a wind farm consisting of 33 wind turbine units to install nacelle-type lidar.

[0094] In this embodiment, the locations of the wind turbines within the wind farm are distributed as follows: Figure 2 As shown, the wind turbine units are arranged in a complex manner.

[0095] Step 1) Data from 33 wind turbine units were collected using the SCADA system. The sampling frequency was once every 10 minutes. The data spanned from February 1, 2022 to April 30, 2022. After deleting negative power data and incomplete data within a single sampling time, the total number of data sets was 10065.

[0096] Step 2) Based on the coordinate component method, establish a Cartesian coordinate system with true north as 0°, and calculate the coordinate components in the Cartesian coordinate system using the wind speed and direction data of each wind turbine. For example, if the wind speed of wind turbine 1 at the sampling time of 00:00:00 on February 1, 2022 is 5.82 m / s and the wind direction is 0.3°, then the coordinate components of wind turbine 1 at that sampling time are calculated to be (0.03, 5.82).

[0097] Step 3) Superimpose and average the coordinate components of the wind speed and direction vectors of all wind turbines within a single sampling time to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at the current sampling time, and then calculate the wind speed and direction of the main inflow of the wind farm. The coordinate components of the wind speed and direction vector of the main inflow of the wind farm corresponding to the sampling time of 00:00:00 on February 1, 2022 are (2.47, 3.09), the corresponding wind speed of the main inflow of the wind farm is 5.72 m / s, and the wind direction of the main inflow is 4.69°.

[0098] Step 4) Calculate the point integral of the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean is used to obtain the comprehensive point integral of each wind turbine at all sampling times, which is used as the correlation index 1 with the wind speed and direction of the main inflow. The correlation ranking 1 is obtained. The top 5 turbines in the index ranking are numbered 30, 6, 31, 5, and 17.

[0099] Step 5) Calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Calculate the arithmetic mean of the comprehensive correlation coefficient of each wind turbine at all sampling times, which is used as the correlation index 2 with the wind speed and direction vector of the main inflow of the wind farm. Obtain the correlation ranking 2. The top 5 turbines are numbered 1, 22, 15, 3, and 4.

[0100] Step 6) Using the wind speed and direction vectors of the main inflow of the wind farm corresponding to the first 80% of the total data of 10065 sets of data as the training set input to the random forest model, train the wind speed and direction vector prediction model of the main inflow of the wind farm, and combine the SHAP value analysis method to calculate the comprehensive SHAP value of each wind turbine at all sampling times, as the correlation index 3 of the main inflow wind speed and direction, and obtain the correlation ranking 3. The top 5 turbines are numbered 3, 6, 7, 20 and 15.

[0101] Step 7) Select the 5 units from the above 3 rankings as the three sets of schemes for the installation sites of the cabin-type lidar.

[0102] Step 8) Based on the installation sites of the nacelle-type lidar in the three schemes obtained in Steps 5 to 7 above, extract the coordinate components of the wind speed and direction vectors of the wind turbines at the installation sites. Calculate the coordinate components of the wind speed and direction vectors of the secondary inflow of the wind farm corresponding to the three schemes. Combined with the coordinate components of the wind speed and direction vectors of the main inflow of the wind farm, calculate the consistency index between the coordinate components of the wind speed and direction vectors of the secondary inflow of the wind farm and the coordinate components of the wind speed and direction vectors of the main inflow of the wind farm. Add the weight coefficients of wind speed and direction to calculate the comprehensive evaluation index at the current sampling time. Then, arithmetically average the comprehensive evaluation index of all sampling times to obtain the final evaluation index of each scheme. The wind speed and direction weight values ​​are all 0.5. The final evaluation indices of the three schemes are 0.946, 0.966, and 0.964, respectively.

[0103] Step 9) Comparing the final evaluation indicators of the three schemes, it was found that the second scheme had the highest final evaluation indicator. Therefore, based on the evaluation indicators, the key installation sites for the naval lidar were selected as 1, 22, 15, 3, and 4. The locations of the sites are as follows: Figure 3 As shown.

[0104] The technical features of the above embodiments can be combined in any way. For the sake of brevity, not all possible combinations of the technical features in the above embodiments are described. However, as long as there is no contradiction in the combination of these technical features, they should be considered to be within the scope of this specification.

[0105] The embodiments described above are merely illustrative of several implementation methods of this application, and while the descriptions are specific and detailed, they should not be construed as limiting the scope of this application. It should be noted that those skilled in the art can make various modifications and improvements without departing from the concept of this application, and these modifications and improvements all fall within the protection scope of this application. Therefore, the protection scope of this application should be determined by the appended claims.

Claims

1. A method for selecting key installation sites for a wind farm nacelle-type lidar, characterized in that, The specific steps are as follows: Step 1: Collect data for each wind turbine using the wind farm's SCADA system, including sampling time, wind speed, wind direction, and power, and filter out invalid data; Step 2: Establish a Cartesian coordinate system with due north as 0°, construct wind speed and wind direction vectors with wind speed as the modulus and wind direction as the angle, and calculate the coordinate components of the wind speed and wind direction vectors of each wind turbine in the Cartesian coordinate system. Step 3: For a single sampling time, superimpose the coordinate components of the wind speed and direction vectors of all wind turbines and take an arithmetic average to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time, and further calculate the wind speed and direction of the main inflow of the wind farm. Step 4: Calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Take the arithmetic mean of the point integral of each wind turbine at all sampling times to obtain the comprehensive point integral of each wind turbine. This comprehensive point integral is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and obtain the correlation ranking 1. Step 5: Calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the correlation coefficient of each wind turbine at all sampling times is used to obtain the comprehensive correlation coefficient of each wind turbine, which is used as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and the correlation ranking 2 is obtained. Step 6: Using the wind speed and direction vectors of the main inflow of the wind farm for the first 80% of all sampling times as the training set, input the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm. Combined with the SHAP value analysis method, calculate the SHAP value of the prediction result of the coordinate components of the wind speed and direction vector of each wind turbine to the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the SHAP value of each wind turbine at all sampling times is used to obtain the comprehensive SHAP value of each wind turbine, which is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and obtain the correlation ranking 3. Step 7: Based on the required number of nacelle-type lidar installations, and using the three rankings obtained above as a basis, select the top m wind turbine units in each of the three rankings as the three options for selecting nacelle-type lidar installation sites. Step 8: For each scheme, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, calculate the wind speed and direction vector of the secondary inflow of the wind farm for that scheme, and then calculate the consistency index between the wind speed and direction vector of the secondary inflow of the wind farm and the wind speed and direction vector of the main inflow of the wind farm. Add the weight values ​​of wind speed and direction to calculate the comprehensive evaluation index of the scheme at each sampling time, and arithmetically average the comprehensive evaluation index of all sampling times to obtain the final evaluation index of the scheme. Step 9: Compare the final evaluation indicators of the three schemes and select the scheme with the highest final evaluation indicator as the installation site scheme for the cabin-type lidar.

2. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 2: Establish a Cartesian coordinate system with true north as 0°. Construct wind speed and wind direction vectors using wind speed as the modulus and wind direction as the angle. Calculate the coordinate components of the wind speed and wind direction vectors for each wind turbine in the Cartesian coordinate system. The specific formula is as follows: ； Where i represents the i-th wind turbine in the wind farm, and t represents the t-th sampling time out of all sampling times. Let be the wind speed of wind turbine i at sampling time t. Let t be the wind direction of wind turbine i at sampling time t. Let be the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t in the Cartesian coordinate system.

3. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 3: For a single sampling time, superimpose the coordinate components of the wind speed and direction vectors of all wind turbines and take an arithmetic mean to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time, as well as the wind speed and direction of the main inflow of the wind farm. The specific method is as follows: Within a single sampling time t, the coordinate components of the wind speed and direction vectors of all wind turbines are superimposed and arithmetically averaged to obtain the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at that sampling time t. The specific formula is as follows: ; in, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t. These are the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at sampling time t; The wind speed and direction of the main inflow to the wind farm are calculated as follows: ； ； ； in, It represents the wind direction of wind turbine i at sampling time t. It refers to the number of wind turbines in the wind farm. It is the wind speed of the main inflow into the wind farm at sampling time t. It represents the wind direction of the main inflow to the wind farm at sampling time t before the angle was changed. It is the wind direction of the main inflow of the wind farm at the sampling time t after the angle is converted.

4. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 4: Calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Then, arithmetically average the point integral of each wind turbine across all sampling times to obtain the comprehensive point integral of each wind turbine. This comprehensive point integral serves as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and a correlation ranking is obtained. The specific method is as follows: First, calculate the point integral between the coordinate components of the wind speed and direction vector of each wind turbine at each sampling time and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The formula is: ； In the formula, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t. and These are the coordinate components of the wind speed and direction vector of the main inflow into the wind farm at sampling time t. It is the point integral of wind turbine i at sampling time t; Then, the arithmetic mean of the point integrals for each wind turbine at all sampling times is calculated to obtain the comprehensive point integral for each wind turbine. This comprehensive point integral is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. The formula is as follows: ； In the formula, The total number of sampling times. Let 1 be the integral of the comprehensive points of wind turbine i, which is the correlation index. Finally, the integral of the comprehensive points is used as the correlation index 1 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking is obtained. .

5. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 5: Calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. Then, arithmetically average the correlation coefficient of each wind turbine at all sampling times to obtain the comprehensive correlation coefficient of each wind turbine. This comprehensive correlation coefficient serves as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and a correlation ranking 2 is obtained. The specific method is as follows: First, calculate the correlation coefficient between the coordinate components of the wind speed and direction vector of each wind turbine and the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The formula is: ； ； ； ； In the formula, and These are the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t, where T represents all sampling times. and It is the average coordinate component of the wind speed and direction vector of wind turbine i over all sampling times. and These are the coordinate components of the wind speed and direction vector of the main inflow wind of the wind farm at sampling time t. and It is the average value of the coordinate components of the wind speed and direction vector of the main inflow of the wind farm at all sampling times. and It is the correlation coefficient between the wind speed and direction vector of wind turbine i at sampling time t and the wind speed and direction vector of the main inflow of the wind farm; Then, the correlation coefficient of each wind turbine is calculated by arithmetic mean at all sampling times, thus obtaining the comprehensive correlation coefficient of each wind turbine, as shown in the formula: ； ； In the formula, It is the total correlation coefficient of the wind speed and wind direction vectors of wind turbine i at sampling time t. is the comprehensive correlation coefficient of wind turbine i, and T is the number of all sampling times; Finally, the comprehensive correlation coefficient is used as the correlation index 2 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking 2 is obtained. .

6. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 6: Using the wind speed and direction vectors of the main inflow of the wind farm from the first 80% of all sampling times as the training set, input the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm. Combined with the SHAP value analysis method, calculate the SHAP value of the coordinate components of the wind speed and direction vector of each wind turbine to the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. The arithmetic mean of the SHAP value of each wind turbine at all sampling times is used to obtain the comprehensive SHAP value of each wind turbine, which is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm, and the correlation ranking 3 is obtained. The specific method is as follows: First, the coordinate components of the wind speed and direction vectors of the main inflow of the wind farm in the previous 80% of all sampling times are used as the training set input to the random forest model to train the random forest model and obtain the wind speed and direction vector prediction model of the main inflow of the wind farm. Then, using the coordinate components of the wind speed and direction vectors of the main inflow of the wind farm at the last 20% of all sampling times as the test set, the performance of the wind speed and direction vector prediction model for the main inflow of the wind farm was evaluated. If the model If the value is greater than 0.9, then the SHAP value of the predicted coordinate components of the wind speed and direction vectors of each wind turbine is calculated based on the trained model, and the total SHAP value of wind turbine i at sampling time t is calculated based on the sum of squares formula, which is: ； In the formula, and It is the SHAP value of the prediction result of the coordinate components of the wind speed and direction vector of wind turbine i at sampling time t against the coordinate components of the wind speed and direction vector of the main inflow of the wind farm. It is the total SHAP value of wind turbine i at sampling time t; Secondly, the SHAP value of each wind turbine is calculated by arithmetic mean at all sampling times, thus obtaining the comprehensive SHAP value of each wind turbine. The formula is as follows: ； In the formula, T represents the total number of sampling times. This represents the overall SHAP value of wind turbine i. Finally, the comprehensive SHAP value is used as the correlation index 3 between the wind speed and direction vector of each wind turbine and the wind speed and direction vector of the main inflow of the wind farm. After sorting from largest to smallest, the correlation ranking 3 is obtained. .

7. The method for selecting key installation sites for wind farm nacelle-type lidar according to claim 1, characterized in that, Step 8: For each scheme, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, calculate the wind speed and direction vectors of the secondary inflow of the wind farm for that scheme. Then, calculate the consistency index between the wind speed and direction vectors of the secondary inflow and the wind speed and direction vectors of the main inflow of the wind farm. Add the weight values ​​of wind speed and direction to calculate the comprehensive evaluation index of the scheme at each sampling time. Arithmetically average the comprehensive evaluation index of all sampling times to obtain the final evaluation index of the scheme. The specific method is as follows: First, based on the wind speed and direction vectors of the m wind turbines at each sampling time at the nacelle-type lidar installation site, the wind speed and direction vector of the secondary inflow to the wind farm in this scheme is calculated using the following formula: ； ； ； ； In the formula, p represents one of the three schemes. Let m be the m wind turbines selected as installation sites for the nacelle-type lidar, and j be one of the m wind turbines. and These are the coordinate components of the wind speed and direction vector of wind turbine j at sampling time t. These are the coordinate components of the wind speed and direction vector of the secondary inflow into the wind farm at sampling time t, corresponding to the scheme. It is the wind speed of wind turbine j at sampling time t. It represents the wind speed of the secondary inflow into the wind farm at sampling time t, corresponding to the scheme. This refers to the wind direction of the secondary inflow at the wind farm at sampling time t before the angle conversion, corresponding to the scheme. It is the wind direction of the secondary inflow of the wind farm at sampling time t after the angle conversion of the corresponding scheme; Secondly, combining the coordinate components of the wind speed and direction vector of the main inflow of the wind farm, the coordinate components and point integral of the wind speed and direction vector of the secondary inflow of the wind farm are calculated. Based on the wind speed and direction of the main inflow and the secondary inflow of the wind farm, a consistency index in wind speed and direction is calculated. The formula is as follows: ； ； ； In the formula, Let be the point integral of the coordinate components of the wind speed and direction of the main inflow to the wind farm and the coordinate components of the wind speed and direction of the secondary inflow to the wind farm at sampling time t. and These are the coordinate components of the wind speed and direction vector of the main inflow into the wind farm at sampling time t. and These are the coordinate components of the wind speed and direction vector of the secondary inflow into the wind farm at sampling time t for scheme p. It is the wind speed of the main inflow into the wind farm at sampling time t. It is the wind speed of the secondary inflow into the wind farm at sampling time t for scheme p. and These are the consistency in terms of wind direction and wind speed between the wind speed and wind direction vectors of the secondary inflow and the main inflow of the wind farm at sampling time t of scheme p. Finally, weighted values ​​are added, and the comprehensive evaluation index of scheme p at sampling time t is calculated by combining the wind speed and direction vectors of the secondary inflow and the main inflow of the wind farm. The arithmetic mean is then used to obtain the final evaluation index of scheme p, as shown in the formula: ； ； In the formula, The weighted value for wind direction. The weighted values ​​for wind speed are as follows: , It is the comprehensive correlation index of scheme p at sampling time t. It represents the total number of sampling times. It is the final evaluation index for scheme p.

8. A key site selection system for wind farm nacelle-type lidar installation, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method for selecting key installation sites for wind farm nacelle lidar as described in any one of claims 1-7.

9. A computer device, characterized in that, The system includes a memory, a processor, and a computer program stored in the memory and executable on the processor. When the processor executes the program, it implements the method for selecting key installation sites for wind farm nacelle lidar as described in any one of claims 1-7.

10. A computer-readable storage medium, characterized in that, The computer-readable storage medium stores a computer program, which, when executed by a processor, implements the method for selecting key installation sites of a wind farm nacelle-type lidar as described in any one of claims 1-7.

Images