A yaw correction method for intelligent wind turbines
By cleaning and fitting algorithm correction of the historical operation data of the wind turbine, the optimal bias value is obtained and yaw correction is solved, and the problem of large hardware investment and insufficient data cleaning in the prior art is solved, and more efficient yaw correction and power generation efficiency are achieved.
Patent Information
- Application Number
- CN202411702175.7
- Authority / Receiving Office
- CN · China
- Patent Type
- Applications(China)
- Current Assignee / Owner
- Filing Date
- 2024-11-26
- Publication Date
- 2025-04-04
AI Technical Summary
The prior art has problems such as excessive hardware investment, high cost and insufficient data cleaning in wind turbine yaw correction, resulting in insufficient accuracy of correction values.
The yaw correction method of function fitting is adopted, and the historical operation data of the wind turbine is cleaned and fitted with algorithm correction, and the optimal bias value is obtained, and the yaw correction is performed in the PLC to achieve periodic correction.
There is no need to add too many hardware equipment, which improves the accuracy of yaw correction and wind turbine power generation efficiency and reduces costs.
Smart Images
Figure CN119760304A_ABST
Abstract
Description
Technical Field
[0001] The invention relates to the technical field of yaw correction of wind turbine sets, and in particular to an intelligent yaw correction method for wind turbine sets. Background Art
[0002] With the development of wind power technology, increasing power generation is the focus of research in the field of wind power. Accurately aligning wind turbines with the wind and better capturing wind energy are key factors in increasing wind turbine power generation, and the yaw system is the key to adjusting the wind turbine to the correct wind direction. However, due to various external factors or factors of the wind turbine itself, such as the wake generated by the rotation of the impeller affecting the wind direction measurement, the deviation of the wind vane from the reference alignment, the alignment deviation of the commissioning personnel, the influence of terrain and environmental factors, maintenance, adjustment and other reasons, the wind turbine generates a certain wind error during operation, thereby reducing the wind power utilization rate. According to statistical data in this field, yaw error will lead to a decrease in the available wind speed and wind energy utilization efficiency, and the power will drop sharply with the increase of the yaw error angle. The power loss is 1.5% for a yaw error angle of 5°, 5.9% for a yaw error angle of 10°, and as high as 12.9% for a yaw error angle of 15°.
[0003] To address this problem and to ensure that wind turbines can maintain a high power generation efficiency, different yaw correction schemes have been proposed. Existing technologies are mostly based on the arrangement of wind turbines in the target area, with multiple lidars deployed. The lidars are then used to identify the wind conditions in the target area, generate real-time wind condition data, and perform corrections. However, this method requires too much hardware investment and is too costly. Alternatively, corrections are performed based on the historical operating data of the wind turbines, but data cleaning is not perfect, and the correction values are often fuzzy, and a more accurate correction value cannot be provided. Summary of the invention
[0004] The purpose of the present invention is to overcome the shortcomings of the prior art and propose a yaw correction method for an intelligent wind turbine. The yaw correction method adopts a function fitting method, does not require the addition of too much hardware equipment, can clean the data according to the historical operation data of the wind turbine, and uses the functional relationship to optimize the wind energy utilization rate to obtain a more accurate yaw correction value.
[0005] The purpose of the present invention is achieved through the following technical solution: A method for yaw correction of an intelligent wind turbine generator system, comprising the following steps:
[0006] S1. Collecting the operation data of wind turbines;
[0007] S2. Clean the operation data to obtain cleaned operation data;
[0008] S3, according to the running data after cleaning, the fitting algorithm is calibrated to obtain the optimal bias value;
[0009] S4, inputting the optimal bias value into the wind turbine PLC to perform yaw correction;
[0010] S5, setting a calibration period, and after reaching the preset calibration period, repeating steps S1 to S4.
[0011] Further, the step S1 comprises:
[0012] The operating data of the wind turbine set within the preset cycle is collected. The operating data is second-level data and includes time, wind speed, actual power of the set, relative wind direction, air density, set operating mode, yaw mode, blade angle, power limit flag and yaw angle offset value.
[0013] Further, the step S2 comprises the following steps:
[0014] S2.1. Primary screening: Eliminate abnormal operation data of wind turbines, power-limited data and missing data;
[0015] S2.2. Data calculation: Calculate the value S of the energy absorbed by the wind turbine from natural wind energy cp , standard power and wind direction range;
[0016] S2.3, Secondary screening: Eliminate outliers in the operating data and remove the effect of wind speed on S cp impact.
[0017] Further, the step S2.1 comprises:
[0018] Convert the operating data into 2-minute average data, optimize the wind direction jump, and then eliminate the operating data with the following problems:
[0019] 1) Eliminate missing operation data;
[0020] 2) Eliminate the operation data of wind turbines whose operation mode is not the normal operation mode;
[0021] 3) Eliminate the operation data with power limit flag set as power limit;
[0022] 4) Eliminate the operating data with blade angle greater than 2°;
[0023] 5) Except for the operating data in the wind speed range [3m / s, 9m / s] and the operating data of wind turbines at full power or limited power at wind speeds above 9m / s, other wind speed operating data are excluded;
[0024] 6) Eliminate operating data with power less than 0.1 times the rated power and greater than 0.9 times the rated power.
[0025] Further, the step S2.2 comprises the following steps:
[0026] S2.2.1、S cp calculate:
[0027] The optimal analysis of yaw correction requires the calculation of wind turbine S cp Value, according to the wind power calculation formula:
[0028] P=1 / 2(ρsv 3 )C p (1);
[0029] Where P is the actual power of the unit; ρ is the air density; s is the wind swept area of the impeller; v is the wind speed; C p is the wind energy utilization rate;
[0030] S cp The calculation formula is as follows:
[0031] s*C p / 2=P / (ρv 3 ) (2);
[0032] remember:
[0033] S cp =s*C p / 2
[0034] =P / (ρv 3 ) (3)
[0035] S2.2.2, Standard power calculation:
[0036] Convert the wind turbine power to the standard power P at 0℃ 0
[0037] P 0 =P / ρ*1.293 (4);
[0038] S2.2.3. Relative wind direction angle interval division: each 1° is divided into one interval.
[0039] Further, the step S2.3 includes the following steps:
[0040] S2.3.1. Screening outliers:
[0041] Use the Tiger formula and the least squares method to fit the relative wind direction-S cp curve, obtain the functional relationship between them, and then use this function to calculate the fitting S in the relative wind direction. cp Value, remove S cp Fitting S at 0.2 times cp and 1.8 times the fitting S cpThe data other than the above were then screened again using 0.4x and 1.6x factors;
[0042] S2.3.2. Eliminate the influence of power on cp:
[0043] Fitting Power-S using Taylor formula polynomial and least squares method cp curve, get the functional relationship between them, and use S cp The fitted S calculated with this function at the same power cp The ratio of S cp , eliminating standard S cp is a value less than 0.2 and greater than 2.4;
[0044] S2.3.3. Compare the data before and after data cleaning, and compare the discreteness before and after data cleaning.
[0045] Further, the step S3 comprises the following steps:
[0046] S3.1. Determine the relative wind direction range, including the data range during fitting and the credible interval range of the bias value result;
[0047] S3.2. Fitting relative wind direction and S cp Function, find the fitting curve function;
[0048] S3.3. Calculate the optimal offset value based on the relative wind direction range and the fitting curve function.
[0049] Further, the step S3.1 comprises:
[0050] Data range during fitting: Since the relative wind direction edge data density is very sparse, in order to reduce the impact of edge data during fitting, the data with relative wind direction intervals not greater than 5 are removed before fitting;
[0051] Calculate the credible interval range of the bias value result: take the interval with wind direction interval data greater than 200 as the credible interval range.
[0052] Further, the step S3.2 includes:
[0053] Find the fitting curve function: Use Taylor's formula and least squares method to fit the scattered points and curve graph, determine the order of the polynomial, which represents the degree of fitting, and use the determined order polynomial to fit the relative wind direction and S cp function, and obtain the fitting function.
[0054] Further, the step S3.3 includes:
[0055] Find the optimal bias value: According to the range of the bias value credible interval and the relative wind direction and S cpThe function relationship of the bias value is obtained within the range of the credible interval, that is, S cp The maximum value and its corresponding relative wind direction are the optimal offset value required for wind turbine correction. The wind turbine is corrected according to the optimal offset value.
[0056] Compared with the prior art, the present invention has the following advantages and beneficial effects:
[0057] The present invention utilizes the Taylor formula fitting curve function to screen data without the need for new hardware investment, thereby solving the problem of data cleaning in automated programs and making the results obtained more accurate. The accuracy of yaw error correction for wind turbines is greatly improved, thereby improving the power generation efficiency of wind turbines. BRIEF DESCRIPTION OF THE DRAWINGS
[0058] Figure 1 It is a flow chart of the present invention.
[0059] Figure 2 Relative wind direction of the original data - S cp Scatter plot.
[0060] Figure 3 Relative wind direction after one screening - S cp Scatter plot.
[0061] Figure 4 Relative wind direction after secondary screening - S cp Scatter plot.
[0062] Figure 5 It is a curve graph of the wind turbine fitting function within the credible wind direction range. DETAILED DESCRIPTION
[0063] The present invention will be further described below in conjunction with specific embodiments.
[0064] Example 1
[0065] See also Figure 1 As shown, the yaw correction method of the intelligent wind turbine provided in this embodiment uses a method of scatter point fitting curve function based on the Taylor formula principle and the least square method. Function fitting is used when performing data screening and finding the optimal relative wind direction. The principle of this fitting curve is as follows:
[0066] Relative wind direction and wind energy utilization have a certain functional relationship f(x) within a certain wind direction. This functional relationship is unknown, but based on experience, it can be concluded that near the optimal relative wind direction, it is a convex function, and the angle corresponding to the highest point should be the optimal offset angle. The essence of Taylor expansion is polynomial approximation, that is, the function value of the function f(x) in the neighborhood of a certain point can be fitted by accumulating low-order to high-order polynomials, that is, using polynomial functions to approximate smooth functions. The relative wind direction and wind energy utilization obtained according to the operating data of the wind turbine set are non-uniquely mapped. Therefore, the curve to be found should use the least squares method to obtain the polynomial function of this curve to approximate the function f(x). The method includes the following steps:
[0067] S1. Collect the operating data of the wind turbine; collect the operating data of the wind turbine in the past two months. The operating data is second-level data, including time, wind speed, actual power of the unit, relative wind direction, air density, unit operating mode, yaw mode, blade angle, power limit flag and yaw angle offset value.
[0068] S2. Clean the operation data. In order to ensure the rationality and objectivity of the yaw error correction, it is necessary to clean some of the wind turbine operation data, filter out the messy data that affects the analysis results, and retain the data required for subsequent processing and calculation. Obtaining the cleaned operation data includes the following steps:
[0069] S2.1. Primary screening: Eliminate abnormal operation data of wind turbines, power-limited data and missing data, including:
[0070] Convert the operating data into 2-minute average data, optimize the wind direction jump, and then eliminate the operating data with the following problems:
[0071] 1) Eliminate missing operation data;
[0072] 2) Eliminate the operation data of wind turbines whose operation mode is not the normal operation mode;
[0073] 3) Eliminate the operation data with power limit flag set as power limit;
[0074] 4) Eliminate the operating data with blade angle greater than 2°;
[0075] 5) Except for the operating data in the wind speed range [3m / s, 9m / s] and the operating data of wind turbines at full power or limited power at wind speeds above 9m / s, other wind speed operating data are excluded;
[0076] 6) Eliminate operating data with power less than 0.1 times the rated power and greater than 0.9 times the rated power.
[0077] S2.2. Data calculation: Calculate the value S of the energy absorbed by the wind turbine from natural wind energy cp , standard power and wind direction range, including the following steps:
[0078] S2.2.1、S cp calculate:
[0079] The optimal analysis of yaw correction requires the calculation of wind turbine S cp Value, according to the wind power calculation formula:
[0080] P=1 / 2(ρsv 3 )C p (1);
[0081] Among them, P is the actual power of the unit; ρ is the air density; s is the wind swept area of the impeller; v is the wind speed; Cp is the wind energy utilization rate;
[0082] Among them, variables such as ρ, s, and v directly affect the power. For the same wind turbine, only the relative wind direction affects the size of the wind swept area s of the impeller. In order to remove the influence of factors such as air density and wind speed, P and ρv are used. 3 The ratio of is used as the evaluation value of the wind energy utilization rate of the wind turbine, which is called S cp , and C p The functions are the same, both represent the extent to which wind turbines absorb energy from natural wind energy. To facilitate unified calculation, S is used in subsequent processing. cp Replace C p .
[0083] S cp The calculation formula is as follows:
[0084] s*C p / 2=P / (ρv 3 ) (2);
[0085] remember:
[0086] S cp =s*C p / 2
[0087] =P / (ρv 3 ) (3)
[0088] S2.2.2, Standard power calculation:
[0089] Convert the wind turbine power to the standard power P at 0℃ 0
[0090] P 0 =P / ρ*1.293 (4);
[0091] S2.2.3. Relative wind direction angle interval division: each 1° is divided into one interval.
[0092] S2.3, Secondary screening: Eliminate outliers in the operating data and remove the effect of wind speed on S cp The impact includes the following steps:
[0093] S2.3.1. Screening outliers:
[0094] Use the Tiger formula and the least squares method to fit the relative wind direction-S cp curve, obtain the functional relationship between them, and then use this function to calculate the fitting S in the relative wind direction. cp Value, remove S cp Fitting S at 0.2 times cp and 1.8 times the fitting S cp The data other than the above were then screened again using 0.4x and 1.6x factors;
[0095] S2.3.2. Eliminate the influence of power on cp:
[0096] Fitting Power-S using Taylor formula polynomial and least squares method cp curve, get the functional relationship between them, and use S cp The fitted S calculated with this function at the same power cp The ratio of S cp , eliminating standard S cp is a value less than 0.2 and greater than 2.4;
[0097] S2.3.3, Compare the data before and after cleaning, and compare the discreteness before and after data cleaning. Figure 2 Original data relative wind direction-S cp Scatter plot, Figure 3 Relative wind direction after one screening - S cp Scatter plot, Figure 4 Relative wind direction after secondary screening-S cp Scatter plot, by Figures 2 to 4 As shown, it can be clearly seen that the fan data after data cleaning becomes more convergent, and the functional relationship obtained by iterative calculation using polynomials becomes closer to the theoretical results.
[0098] S3, according to the running data after cleaning, the fitting algorithm is calibrated to obtain the optimal bias value, including the following steps:
[0099] S3.1. Determine the relative wind direction range, including the data range during fitting and the credible interval range of the bias value results, including:
[0100] Data range during fitting: Since the relative wind direction edge data density is very sparse, in order to reduce the impact of edge data during fitting, the data with relative wind direction intervals not greater than 5 are removed before fitting;
[0101] Calculate the credible interval range of the bias value result: take the interval with wind direction interval data greater than 200 as the credible interval range.
[0102] S3.2. Fitting relative wind direction and S cp Function, find the fitting curve function, including:
[0103] Find the fitting curve function: Use Taylor's formula and least squares method to fit the scattered points and curve graph, determine the order of the polynomial, which represents the degree of fitting. For data of two months or more, the 30th order is used. Use the 30th order polynomial to fit the relative wind direction and S cp function, and obtain the fitting function.
[0104] S3.3. Calculate the optimal offset value based on the relative wind direction range and the fitting curve function, including:
[0105] Find the optimal bias value: According to the range of the bias value credible interval and the relative wind direction and S cp The function relationship of the bias value is obtained within the range of the credible interval, that is, S cp The maximum value of and its corresponding relative wind direction, the corresponding relative wind direction is the optimal offset value required for wind turbine calibration, and the wind turbine is calibrated according to the optimal offset value. Figure 5 As shown in FIG. 1 , the curve of the wind turbine fitting function within the credible wind direction range has the highest point of the image being the optimal bias value for the current wind turbine yaw correction.
[0106] S4, inputting the optimal bias value into the wind turbine PLC to perform yaw correction;
[0107] S5. Set the calibration period to 2 months. After 2 months, repeat steps S1 to S4 to reduce the long-term yaw error of the wind turbine.
[0108] After executing the above method, according to actual measurements at the wind farm, the power generation of some wind turbines in the wind farm increased by more than 5%.
[0109] Example 2
[0110] This embodiment discloses a non-transitory computer-readable medium storing instructions. When the instructions are executed by a processor, the steps of the yaw correction method for an intelligent wind turbine set according to Embodiment 1 are executed.
[0111] The non-temporary computer-readable medium in this embodiment may be a disk, an optical disk, a computer memory, a read-only memory (ROM), a random access memory (RAM), a USB flash drive, a mobile hard disk, or other media.
[0112] Example 3
[0113] This embodiment discloses a computing device, including a processor and a memory for storing a program executable by the processor. When the processor executes the program stored in the memory, the yaw correction method for the intelligent wind turbine set described in Embodiment 1 is implemented.
[0114] The computing device described in this embodiment may be a desktop computer, a laptop computer, a smart phone, a PDA handheld terminal, a tablet computer, a programmable logic controller (PLC), or other terminal devices with a processor function.
[0115] The embodiments described above are only preferred embodiments of the present invention and are not intended to limit the scope of implementation of the present invention. Therefore, all changes made according to the shape and principle of the present invention should be included in the protection scope of the present invention.
Claims
1. A yaw correction method for an intelligent wind turbine, characterized in that: The following steps are involved: S1. Collecting the operation data of wind turbines; S2. Clean the operation data to obtain cleaned operation data; S3, according to the running data after cleaning, the fitting algorithm is calibrated to obtain the optimal bias value; S4, inputting the optimal bias value into the wind turbine PLC to perform yaw correction; S5, setting a calibration period, and after reaching the preset calibration period, repeating steps S1 to S4.
2. The method for yaw correction of an intelligent wind turbine according to claim 1, characterized in that: The step S1 comprises: The operating data of the wind turbine set within the preset cycle is collected. The operating data is second-level data and includes time, wind speed, actual power of the set, relative wind direction, air density, set operating mode, yaw mode, blade angle, power limit flag and yaw angle offset value.
3. The method for yaw correction of an intelligent wind turbine according to claim 1, characterized in that: The step S2 comprises the following steps: S2.
1. Primary screening: Eliminate abnormal operation data of wind turbines, power-limited data and missing data; S2.
2. Data calculation: Calculate the value S of the energy absorbed by the wind turbine from natural wind energy cp , standard power and wind direction range; S2.3, Secondary screening: Eliminate outliers in the operating data and remove the effect of wind speed on S cp impact.
4. The method for yaw correction of an intelligent wind turbine according to claim 3, characterized in that: The step S2.1 comprises: Convert the operating data into 2-minute average data, optimize the wind direction jump, and then eliminate the operating data with the following problems: 1) Eliminate missing operation data; 2) Eliminate the operation data of wind turbines whose operation mode is not the normal operation mode; 3) Eliminate the operation data with power limit flag set as power limit; 4) Eliminate the operating data with blade angle greater than 2°; 5) Except for the operating data in the wind speed range [3m / s, 9m / s] and the operating data of wind turbines at full power or limited power at wind speeds above 9m / s, other wind speed operating data are excluded; 6) Eliminate operating data with power less than 0.1 times the rated power and greater than 0.9 times the rated power.
5. The method for yaw correction of an intelligent wind turbine according to claim 3, characterized in that: The step S2.2 comprises the following steps: S2.2.1、S cp calculate: The optimal analysis of yaw correction requires the calculation of wind turbine S cp Value, according to the wind power calculation formula: P=1 / 2(ρsv 3 )C p (1); Where P is the actual power of the unit; ρ is the air density; s is the wind swept area of the impeller; v is the wind speed; C p is the wind energy utilization rate; S cp The calculation formula is as follows: s*C p / 2=P / (ρv 3 ) (2); remember: S cp =s*C p / 2 =P / (ρv 3 ) (3) S2.2.2, Standard power calculation: Convert the wind turbine power to the standard power P0 at 0℃ P0=P / ρ*1.293 (4); S2.2.
3. Relative wind direction angle interval division: each 1° is divided into one interval.
6. The method for yaw correction of an intelligent wind turbine according to claim 3, characterized in that: The step S2.3 comprises the following steps: S2.3.
1. Screening outliers: Use the Tiger formula and the least squares method to fit the relative wind direction-S cp curve, obtain the functional relationship between them, and then use this function to calculate the fitting S in the relative wind direction. cp Value, remove S cp Fitting S at 0.2 times cp and 1.8 times the fitting S cp The data other than the above were then screened again using 0.4x and 1.6x factors; S2.3.
2. Eliminate the influence of power on cp: Fitting Power-S using Taylor formula polynomial and least squares method cp curve, get the functional relationship between them, and use S cp The fitted S calculated with this function at the same power cp The ratio of S cp , eliminating standard S cp is a value less than 0.2 and greater than 2.4; S2.3.
3. Compare the data before and after data cleaning, and compare the discreteness before and after data cleaning.
7. The method for yaw correction of an intelligent wind turbine according to claim 1, characterized in that: The step S3 comprises the following steps: S3.
1. Determine the relative wind direction range, including the data range during fitting and the credible interval range of the bias value result; S3.
2. Fitting relative wind direction and S cp Function, find the fitting curve function; S3.
3. Calculate the optimal offset value based on the relative wind direction range and the fitting curve function.
8. The method for yaw correction of an intelligent wind turbine according to claim 7, characterized in that: The step S3.1 comprises: Data range during fitting: Since the relative wind direction edge data density is very sparse, in order to reduce the impact of edge data during fitting, the data with relative wind direction intervals not greater than 5 are removed before fitting; Calculate the credible interval range of the bias value result: take the interval with wind direction interval data greater than 200 as the credible interval range.
9. The method for yaw correction of an intelligent wind turbine according to claim 7, characterized in that: The step S3.2 comprises: Find the fitting curve function: Use Taylor's formula and least squares method to fit the scattered points and curve graph, determine the order of the polynomial, which represents the degree of fitting, and use the determined order polynomial to fit the relative wind direction and S cp function, and obtain the fitting function.
10. The method for yaw correction of an intelligent wind turbine according to claim 7, characterized in that: The step S3.3 comprises: Find the optimal bias value: According to the range of the bias value credible interval and the relative wind direction and S cp The function relationship of the bias value is obtained within the range of the credible interval, that is, S cp The maximum value and its corresponding relative wind direction are the optimal offset value required for wind turbine correction. The wind turbine is corrected according to the optimal offset value.