A method for filtering wind data based on correlation of process lines

Abstract

The application provides a wind measurement data screening method based on process line correlation, comprising the following steps: S1. Drawing all wind speed process lines and wind direction process lines of all same periods at each height; S2. Marking a first wind speed abnormal period T i，1 and a first wind direction abnormal period T j，1 according to the process lines in S1 respectively; S3. Selecting all same period wind speeds of any two adjacent heights, and establishing a regression model according to the least square method; S4. Performing a significance test on the regression model in S3, if the test is passed, entering S5, if the test is not passed, returning to S3 to reselect the height; S5. Calculating the residual of the wind speed of each time point and the regression model, and obtaining a second wind speed abnormal period T i，2 ; S6. Selecting all same period wind directions of any two adjacent heights, drawing wind direction rose diagrams respectively, analyzing the dominant wind direction, calculating the difference angle of the dominant wind direction, and marking a second wind direction abnormal period T j，2 if the difference angle is greater than a first threshold; S7. Obtaining an abnormal wind speed period T i =∪(T i，1 , T i，2 ), an abnormal wind direction period T j =∪(T j，1 , T j，2 ), and deleting the corresponding abnormal data, thereby solving the technical problems of the prior art that the existing data screening method is not intuitive and has low accuracy.

Description

Technical Field

[0001] This invention relates to the field of wind farm technology, and in particular to a method for screening and washing wind measurement data based on process line correlation. Background Technology

[0002] To clarify the characteristics of wind energy resources in wind farm construction areas, ensure the quality of basic wind resource data, and reduce investment risks, a thorough investigation of the construction conditions of wind farm projects is necessary to obtain reliable basic data. The basic elements of wind measurement data include wind speed and wind direction, primarily obtained through at least one year of on-site observation using anemometers or lidar in the wind farm area. However, due to the influence of the accuracy, condition, and power supply of on-site observation instruments, as well as the environmental conditions of the wind farm, a certain proportion of unreasonable observation data from anemometers or lidar may exist. This data needs to be filtered and removed based on experience. Current standards recommend verifying wind measurement data by judging whether the difference in wind speed and direction at various heights is less than a certain threshold. However, since wind speed variations are inherently random, and wind speed is affected not only by the atmospheric system but also by the topography of the observation area, relying solely on differences in wind speed and direction at adjacent heights to filter unreasonable data is not intuitive. Furthermore, some valid data may be mistakenly deleted, and the resulting data, after interpolation, can introduce unnecessary errors into wind energy resource assessments. Therefore, it is necessary to propose a more intuitive and convenient method for screening wind measurement data, which will help to more accurately assess the wind energy resources of wind farm projects. Summary of the Invention

[0003] The purpose of this invention is to provide a process-line related wind measurement data screening method to solve the technical problems of non-intuitive and low-accuracy screening methods in the prior art.

[0004] This invention provides a method for screening and filtering wind measurement data based on process line correlation, comprising the following steps: S1. Drawing all concurrent wind speed and wind direction process lines at each altitude; S2. Marking the first wind speed anomaly period T according to the process lines in S1. i，1 and the first abnormal wind direction period T j，1 S3. Select all concurrent wind speeds at any two adjacent heights, plot a scatter plot, and establish a regression model using the least squares method; S4. Perform a significance test on the regression model in S3. If the test passes, proceed to S5; otherwise, return to S3 and select a new height; S5. Calculate the residual between the wind speed at each time point and the regression model to obtain the second wind speed anomaly period T. i，2 S6. Select all concurrent wind directions at any two adjacent heights, draw wind rose diagrams for each, analyze the prevailing wind direction, calculate the difference angle between the prevailing wind directions, and if the difference angle is greater than the first threshold, the corresponding time period is the second abnormal wind direction time period T. j，2S7. Obtain the period T of abnormal wind speed. i =∪(T) i，1 T i，2 Abnormal wind direction period T j =∪(T) j，1 T j，2 ), delete the corresponding abnormal data.

[0005] Furthermore, the first threshold is 22.5°.

[0006] Furthermore, in the second abnormal wind speed period T in S5 i，2 The calculation steps are as follows: S51. Calculate the residuals of wind speed and regression model at each time point; S52. Calculate the rate of change of the absolute value of residuals between adjacent time points; S53. The time period T between the time corresponding to the time of the maximum absolute value of residuals and the time when the rate of change of the absolute value of residuals between two adjacent time points first exceeds the second threshold is the second wind speed anomaly period. i，2 .

[0007] Furthermore, the second threshold is 10%.

[0008] Furthermore, the abnormal period in S2 is a period in which the wind speed or wind direction change trend at a certain altitude is inconsistent with that at other altitudes.

[0009] This invention provides a process-line-related wind measurement data filtering method. It collects hourly observation data from wind measurement towers or lidar in the sea area where the wind farm is located, filters out unreasonable data based on the process line, and determines the unreasonable data period by combining the sum of squared errors of any two height-wind-speed linear regression models with the wind rose diagram of a certain period. This completes the filtering of the original wind measurement data, solving the technical problems of non-intuitive and low-accuracy data filtering methods in the prior art, and helps to more accurately assess the wind energy resources of wind farm projects. Attached Figure Description

[0010] To more clearly illustrate the specific embodiments of the present invention or the technical solutions in the prior art, the drawings used in the description of the specific embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are some embodiments of the present invention. For those skilled in the art, other drawings can be obtained from these drawings without creative effort.

[0011] Figure 1 A flowchart of a process-line correlated wind measurement data screening method provided in this embodiment;

[0012] Figure 2 The wind speed data process lines at heights of 80m and 70m, and the wind direction data process lines at heights of 75m and 65m provided in this embodiment;

[0013] Figure 3 Scatter plots of wind speeds observed simultaneously at heights of 80m and 70m for this embodiment;

[0014] Figure 4 The wind rose diagrams for the period from 0:00 on October 1, 2015 to 23:00 on October 31, 2015 at heights of 75m and 65m provided in this embodiment;

[0015] Figure 5 The wind rose diagrams for heights of 75m and 65m provided in this embodiment, from 0:00 to 23:00 on October 24, 2015. Detailed Implementation

[0016] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0017] This embodiment provides a method for screening and washing wind measurement data based on process line correlation, such as... Figure 1 As shown, it includes the following steps:

[0018] S1. Draw the process lines for wind speed data at heights of 80m and 70m, and wind direction data at heights of 75m and 65m, from 0:00 on October 1, 2015 to 23:00 on October 31, 2015. Figure 2 As shown.

[0019] S2. By combining the trends of wind speed and wind direction changes at adjacent time points, the periods of abnormal wind speed and wind direction in this example can be determined from the process line:

[0020] The wind speed at a height of 70m remained stagnant at 0.3m / s from 00:00 to 17:00 on October 28, 2015 (the wind speed trend remained unchanged during this period), which is inconsistent with the wind speed trend at a height of 80m during the same period (first increasing and then decreasing). Therefore, the initial screening indicates an abnormal wind speed period T. i，1 The wind direction at a height of 65m remained stagnant at 112.2° from 00:00 to 23:00 on October 24, 2015 (the wind direction trend remained unchanged during this period), which is inconsistent with the wind direction trend (decreasing) at a height of 75m during the same period. Therefore, it was initially screened as a period of abnormal wind direction, T. j，1 .

[0021] S3. Draw a scatter plot of wind speeds observed at heights of 80m and 70m during the same period, such as... Figure 3As shown, a regression model is established based on the least squares method:

[0022]

[0023] Using the F-test, the P-value of the regression model was calculated to be 1.02 × 10⁻⁶. -15 It passed the significance test at a confidence level of 5%.

[0024] S4. Calculate the residuals between the wind speed at each time point and the regression model in the scatter plots of wind speed at 80m and 70m:

[0025]

[0026] |e k Sort from largest to smallest, starting with |e k The rate of change of the absolute value of the residual from the moment of maximum value to two adjacent points, Q e The time period included in the first time greater than 10% |e k The values ​​are 7.89, 7.88, 7.85, 7.76, 7.70, and 7.24 respectively. e The percentages were 1.09%, 3.02%, 1.27%, 0.73%, 6.00%, and 11.08%, respectively, thus yielding T. i，2 The corresponding times include: 5:00 on October 28, 2015, 2:00 on October 28, 2015, 6:00 on October 28, 2015, 0:00 on October 28, 2015, 1:00 on October 28, 2015, and 4:00 on October 28, 2015.

[0027] S5. Draw the wind rose diagram for the period from 00:00 on October 1, 2015 to 23:00 on October 31, 2015. In this example, the prevailing easterly wind direction at 75m altitude is E, and the prevailing easterly wind direction at 65m altitude is ESE. The difference between the wind rose diagrams at the two altitudes is greater than 22.5°. Figure 4 As shown.

[0028] Further analysis of the wind rose diagram during the easterly wind period revealed that from 00:00 to 23:00 on October 24, 2015, the prevailing wind direction at heights of 75m and 65m was greater than 22.5°. Figure 5 As shown.

[0029] Except for the period from 0:00 to 23:00 on October 24, 2015, when the difference in prevailing wind direction between the 75m and 65m altitudes was greater than 22.5°, the difference in prevailing wind direction between the two altitudes was less than 22.5° during the other periods. Therefore, the corresponding time periods include: from 0:00 to 23:00 on October 24, 2015.

[0030] S6. Combining the wind speed and direction process lines, wind speed scatter plots, and wind rose diagrams, identify the wind speed and direction data periods T that satisfy both trend anomalies and scatter plot anomalies. i , T j ,:

[0031] T i =∪(T) i，1 T i，2 )

[0032] T j =∪(T) j，1 T j，2 )

[0033] In this example, T i From 0:00 on October 28, 2015 to 17:00 on October 28, 2015, T j The period is from 0:00 on October 24, 2015 to 23:00 on October 24, 2015.

[0034] In this embodiment, combining wind speed and direction process lines, wind speed scatter plots, and wind rose diagrams can easily filter out unreasonable wind speed and direction data, improving work efficiency by approximately 20%. Furthermore, compared to the traditional method of filtering unreasonable data based on wind speed differences of less than 1 m / s at 80m and 70m heights, and wind direction differences of less than 45° at 75m and 65m heights, this method is more intuitive and avoids mistakenly deleting reasonable and valid data from the process lines, thus laying a foundation for accurately assessing the wind energy resources of the site.

[0035] Finally, it should be noted that the above embodiments are only used to illustrate the technical solutions of the present invention, and not to limit them; although the present invention has been described in detail with reference to the foregoing embodiments, those skilled in the art should understand that modifications can still be made to the technical solutions described in the foregoing embodiments, or equivalent substitutions can be made to some or all of the technical features; and these modifications or substitutions do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of the present invention.

Claims

1. A method for screening and washing wind measurement data based on process line correlation, characterized in that, Includes the following steps: S1. Draw the wind speed and wind direction process lines for all contemporaneous periods at each altitude; S2. Mark the first abnormal wind speed period according to the process line in S1. and the first abnormal wind direction period ; S3. Select all concurrent wind speeds at any two adjacent heights, plot a scatter plot, and establish a regression model using the least squares method; S4. Perform a significance test on the regression model in S3. If the test passes, proceed to S5. If the test fails, return to S3 and select a new height. S5. Calculate the residuals between the wind speed and the regression model at each time point to obtain the second period of wind speed anomalies. ; S6. Select all concurrent wind directions at any two adjacent altitudes, draw wind rose diagrams for each, analyze the prevailing wind direction, and calculate the angle of difference between the prevailing wind directions. If the angle of difference is greater than the first threshold, the corresponding time period is considered the second abnormal wind direction period. ; S7. Obtain the period of abnormal wind speed. Abnormal wind direction period Delete the corresponding abnormal data; The abnormal period in S2 is a period in which the wind speed or wind direction change trend at a certain altitude is inconsistent with that at other altitudes. The second abnormal wind speed period in S5 The calculation steps are as follows: S51. Calculate the residuals between the wind speed and the regression model at each time point; S52. Calculate the rate of change of the absolute value of the residuals at adjacent time points; S53. The time interval between the moment corresponding to the moment when the absolute value of the residual reaches its maximum value and the moment when the rate of change of the absolute value of the residual first exceeds the second threshold between two adjacent time points is defined as the second wind speed anomaly period. .

2. The method for screening and washing wind measurement data based on process line correlation according to claim 1, characterized in that, The first threshold is 22.5°.

3. The method for screening and washing wind measurement data based on process line correlation according to claim 1, characterized in that, The second threshold is 10%.

Images