A method for optimizing diffusion parameters of CALPUFF numerical model

Abstract

The present invention relates to a method for optimizing diffusion parameters of a CALPUFF numerical model, comprising the following steps: obtaining atmospheric stability according to meteorological observation data; calculating turbulence observation standard deviation according to turbulence observation data, and establishing a relationship between the atmospheric stability and turbulence observation standard deviation; obtaining turbulence observation standard deviation corresponding to the meteorological observation data according to the relationship, and using it as input data for diffusion parameters of the CALPUFF model. The method for optimizing diffusion parameters of the CALPUFF numerical model of the present invention: using conventional meteorological gradient observations to obtain wind temperature data and part of the turbulence sample data obtained from turbulence observation work carried out at the same time, a set of turbulence observation standard deviations under various stability conditions with universality in the simulated area is established as input data for the atmospheric diffusion parameter scheme 1 of the CALPUFF numerical model, thereby obtaining an atmospheric diffusion parameter scheme that meets the actual weather stability conditions.

Description

Technical Field

[0001] The present invention specifically relates to a method for optimizing a diffusion parameter scheme of a CALPUFF numerical model based on conventional meteorological and turbulence observation data. Background Art

[0002] CALPUFF is a non-steady-state Lagrangian puff model system that simulates the transport, transformation, and removal of pollutants in the atmosphere when the three-dimensional flow field changes with time and space.

[0003] There are five options for atmospheric diffusion parameter schemes in the CALPUFF numerical model:

[0004] 1. Directly use the turbulence observation standard deviation (σ v , σ w );

[0005] 2. Based on the simulated surface thermal and dynamic fluxes, the CALMET meteorological model is used to calculate σ using the similarity theory principle. v , σ w ;

[0006] 3. Use the Pasquill-Gifford (PG) diffusion parameter provided by the CALPUFF numerical model;

[0007] 4. Use the McElroy-Pooler (MP) diffusion parameters provided by the CALPUFF numerical model;

[0008] 5. Use the diffusion equation based on the complex terrain diffusion model (CTDM) provided by the CALPUFF numerical mode for calculation.

[0009] Option 1 and Option 5 require the input of hourly turbulence measurement data, Option 5 also requires refined scale terrain data, and for unstable weather, its diffusion parameters are still calculated using the diffusion scheme Option 3. The diffusion parameter schemes of Options 2 to 4 are all associated with the atmospheric stability conditions embedded in the CALPUFF numerical model.

[0010] The atmospheric stability classification method embedded in the CALPUFF numerical model is the Pasquill-Gifford-Turner method (PGT method) (reference: Boxin. CALPUFF model calculation method and application [M]. Beijing, China Environmental Press, 2016: 5-6). At present, the distribution of domestic ground meteorological observation stations including cloud meteorological element observation stations is relatively sparse, and there are only a few fixed-time observation data per day. In the actual application of the CALPUFF numerical model, cloud data from the simulation results of small and medium-scale weather numerical models (WRF, MM5) are usually used as input data, but this cloud data is a large-scale grid interpolation result, and the representativeness of atmospheric stability in small-scale regions is relatively poor. Therefore, when using the atmospheric diffusion parameter schemes 2 to 4 in the CALPUFF numerical model to simulate regional pollutant diffusion, it often cannot truly reflect the diffusion characteristics under actual atmospheric stability conditions. Option 5 is usually not considered because it requires the results of other external models. Therefore, option 1, as the atmospheric diffusion parameter scheme selected by the model, has good operability and simulation representativeness.

[0011] When the atmospheric diffusion parameter scheme 1 is selected, the standard deviation data of hourly turbulence observations is required as the basic data input, but this type of data is difficult to obtain and the cost is high. Usually, only a certain number of observations are carried out in typical months, and long-series continuous observations and data acquisition are rarely carried out like conventional meteorological element observations. This brings difficulties to the use of hourly turbulence observation standard deviation data as the atmospheric diffusion parameter scheme of the CALPUFF numerical model. Therefore, there is an urgent need for an optimization method for the diffusion parameters of the CALPUFF numerical model, which can obtain atmospheric diffusion parameters that meet the conditions of actual weather stability. Summary of the invention

[0012] In view of this, in order to overcome the defects of the prior art and achieve the above-mentioned purpose, an object of the present invention is to provide a method for optimizing diffusion parameters of a CALPUFF numerical model.

[0013] In order to achieve the above object, the present invention adopts the following technical solutions:

[0014] A method for optimizing diffusion parameters of a CALPUFF numerical model comprises the following steps: obtaining atmospheric stability according to meteorological observation data; calculating a turbulence observation standard deviation according to turbulence observation data, and establishing a relationship between the atmospheric stability and the turbulence observation standard deviation; obtaining the turbulence observation standard deviation corresponding to the meteorological observation data according to the relationship, and using it as input data for diffusion parameters of a CALPUFF model.

[0015] According to some preferred implementation aspects of the present invention, the relationship between atmospheric stability and turbulence observation standard deviation is established according to the following steps:

[0016] The standard deviation of turbulence observation σ obtained by using turbulence observation data statistics v , σ w The temperature gradient and wind speed data obtained from the corresponding meteorological observations are fit by binary linear regression, and the basic equation form is:

[0017] σ v =A0+A1×ΔT+A2×U

[0018] σ w =A0`+A1`×ΔT+A2`×U

[0019] In the formula, σ v , σ w is the standard deviation of lateral or vertical turbulence observations; ΔT is the temperature gradient; U is the wind speed at 10 m; A0, A1, A2, A0 ` 、A1 ` 、A2 ` is the coefficient of the equation.

[0020] According to some preferred implementation aspects of the present invention, taking the calculation of equation coefficients A0, A1, and A2 as an example, the following steps are used to calculate:

[0021] 1) Classify the atmospheric stability by the temperature gradient wind speed method, and classify the σ under different atmospheric stability v , σ w It is divided into three groups: unstable, neutral, and stable. The sample size of each group is set to n1, n2, and n3;

[0022] 2) For a set of samples, use the multivariate regression least squares method, let σ v 、w=Yy i , A0=b0, A1=b1, A2=b2, ΔT=X i1 , U=X i2 , in the predicted values ​​of Y and the measured values ​​of the factor variables with a sample size of n, the binary linear regression equation is satisfied:

[0023] Yy i =b0+b1X i1 +b2X i2 (i=1,2,……n)

[0024] The regression coefficients b0, b1, and b2 should be such that the sum of squares of the differences between all observed values ​​and the regression estimates is minimized; that is, Q is minimized:

[0025]

[0026] For a set of sample data, the estimate of the predicted value is regarded as a vector, denoted as

[0027]

[0028] Satisfy the binary linear regression equation and rewrite it into matrix form: y = Xb, where X is the X defined above. i1 , X i2 The factor matrix is ​​composed of, b is the regression coefficient vector, that is

[0029]

[0030] Q is a non-negative quadratic expression of b0, b1, and b2, so the minimum value must exist. According to the extreme value principle of differential calculus, we have:

[0031]

[0032] That is X ’ Xb=X ’ y

[0033] The above formula can be expanded by matrix multiplication to obtain the general form of the standard equation system:

[0034]

[0035] According to the general linear algebra method of solving systems of equations, b0, b1, and b2 can be solved. The three numbers obtained correspond to the equation coefficients A0, A1, and A2 respectively.

[0036] According to some preferred implementation aspects of the present invention, the turbulence observation standard deviation is calculated by the following formula:

[0037]

[0038] Where M is the observed sample value of v and w, n is the number of observed samples, and σ is the standard deviation of turbulence observation.

[0039] According to some preferred implementation aspects of the present invention, the original observation data obtained by turbulence observation is corrected according to the steps of data group screening, outlier elimination, coordinate transformation, and trend regression to obtain observation sample values.

[0040] According to some preferred implementation aspects of the present invention, the wild point elimination is performed according to the following steps: in a set of turbulence observation raw data, if the deviation of a certain data exceeds 3 times the standard deviation, the value is eliminated and the adjacent data points are interpolated to fill in the points.

[0041] According to some preferred implementation aspects of the present invention, the wild point elimination complies with the following rules: after each wild point is eliminated, a judgment should be performed again until all wild points are eliminated; if multiple wild points appear at the same time, the largest one should be eliminated first, and then a judgment should be performed; when the number of wild points in a group of turbulent raw data exceeds 1%, the group of data is discarded.

[0042] According to some preferred implementation aspects of the present invention, the coordinate transformation is performed according to the following steps: during the data processing process, the data must be rotated, and the x direction after rotation is the average wind direction, called the downwind wind direction, the y direction is the crosswind direction, and the z direction is the vertical wind direction. The average values ​​of the wind speed components in the y direction and the z direction are both zero, and the x, y, and z directions follow the right-hand rule.

[0043] According to some preferred implementation aspects of the present invention, the trend regression is performed according to the following steps: the noise introduced by slow weather system changes or zero drift of the instrument system, as well as the noise introduced by the battery power supply voltage drop, is eliminated by the trend regression method.

[0044] According to some preferred implementation aspects of the present invention, the atmospheric stability is statistically analyzed using a temperature gradient wind speed method.

[0045] Due to the adoption of the above technical scheme, compared with the prior art, the advantages of the present invention are: the optimization method of the diffusion parameters of the CALPUFF numerical model of the present invention: using conventional meteorological gradient observations to obtain wind temperature data and part of the turbulence sample data obtained from the turbulence observation work carried out in the same period, a set of turbulence observation standard deviations under various stability conditions with universal applicability in the simulation area is established as input data of the CALPUFF numerical model atmospheric diffusion parameter scheme 1, thereby obtaining an atmospheric diffusion parameter scheme that meets the actual weather stability conditions. BRIEF DESCRIPTION OF THE DRAWINGS

[0046] In order to more clearly illustrate the technical solutions in the embodiments of the present invention, the following briefly introduces the drawings required for use in the description of the embodiments. Obviously, the drawings described below are only some embodiments of the present invention. For ordinary technicians in this field, other drawings can be obtained based on these drawings without creative work.

[0047] Figure 1 This is a flow chart of turbulence observation data preprocessing in a preferred embodiment of the present invention. DETAILED DESCRIPTION

[0048] In order to enable those skilled in the art to better understand the technical solution of the present invention, the technical solution in the embodiments of the present invention will be clearly and completely described below in conjunction with the drawings in the embodiments of the present invention. Obviously, the described embodiments are only part of the embodiments of the present invention, not all of the embodiments. Based on the embodiments of the present invention, all other embodiments obtained by ordinary technicians in this field without creative work should fall within the scope of protection of the present invention.

[0049] The present invention can optimize the CALPUFF numerical model's ability to analyze and study the diffusion and distribution characteristics of airborne pollutants under normal operating conditions of nuclear power plants. The present invention uses conventional gradient meteorological observation data and turbulence observation standard deviation data of nuclear power plants to establish a set of atmospheric diffusion characteristics that can more accurately reflect the atmospheric stability state of the plant site area, and optimizes the atmospheric diffusion parameter scheme in the CALPUFF numerical model to improve the accuracy and representativeness of the CALPUFF numerical model's simulation results of atmospheric pollutant diffusion in nuclear power plants.

[0050] The optimization method of the diffusion parameters of the CALPUFF numerical model in this embodiment includes the following steps: obtaining atmospheric stability based on meteorological observation data; calculating the turbulence observation standard deviation based on the turbulence observation data, and establishing a relationship between the atmospheric stability and the turbulence observation standard deviation; obtaining the turbulence observation standard deviation corresponding to the meteorological observation data based on the relationship, and using it as input data for the diffusion parameters of the CALPUFF model.

[0051] Specifically, each step is described as follows:

[0052] 1. Obtain atmospheric stability

[0053] Conventional meteorological observation data of nuclear power plants include gradient wind temperature observations from a 100-meter meteorological tower and ground meteorological observation data. This embodiment uses the gradient wind temperature observation data from a 100-meter meteorological tower to carry out statistical analysis of atmospheric stability.

[0054] In addition to the PGT method, the more traditional atmospheric stability classification methods include the temperature gradient method, the temperature gradient wind speed method, the wind direction fluctuation standard deviation method, the boundary layer turbulence parameter method, etc. (References: Jiang Weimei, Sun Jianning, Cao Wenjun, et al. Air Pollution Meteorology Course [M]. Beijing, Meteorological Press, 2004: 85-87). Usually, the temperature gradient wind speed method is used in the nuclear power plant site area to conduct statistical analysis and research on the atmospheric stability of the plant site area, which is also the method used in this embodiment.

[0055] The temperature gradient is calculated by using the hourly average temperature data at the lower 10m and the upper 80m of the 100-meter meteorological tower, and the wind speed is calculated by using the hourly average wind speed data at the 10m height of the 100-meter meteorological tower. The performance indicators of the sensors used for wind temperature meteorological element observations must meet the relevant requirements in Table 1. The meteorological element collection indicators must meet the relevant requirements in Table 2. According to "Atmospheric Dispersion Problems in Site Selection of Nuclear Power Plants" (HAD101 / 02), Table 3 gives the rules for determining stability using the temperature gradient wind speed method.

[0056] Table 1 Sensor performance indicators

[0057]

[0058] Table 2 Meteorological element collection indicators

[0059]

[0060]

[0061] Table 3 Temperature gradient wind speed method to determine stability rules

[0062]

[0063] 2. Obtaining the standard deviation of turbulence observations

[0064] Turbulence observation is mainly carried out by three-dimensional ultrasonic wind temperature instrument, and its observation equipment is usually arranged on the iron tower with the same height as the pollutant emission. The three measuring probes of the three-dimensional ultrasonic wind temperature instrument correspond to each other and are distributed in an upper and lower triangle. The acoustic path of the probe is 10cm. The wind speed measurement range is required to be 0-40m / s, with a resolution of 0.01m / s and an accuracy of ±0.01m / s; the temperature measurement range is -20-+50℃, the temperature pulsation resolution is 0.01℃, the accuracy is ±0.01℃, and the average temperature accuracy is ±2℃. In order to minimize the bit error in the data transmission process, the SDM data line is used for data transmission.

[0065] Before conducting ultrasonic wind and temperature instrument observations, the equipment needs to be calibrated before leaving the factory. The electronic control box contains the calibration information for the corresponding probes, and a closed space static wind test is performed. During the equipment observation, the sensor is oriented toward the dominant wind direction of the time period according to historical meteorological data statistics to minimize the interference of the probe and supporting structure on the flow field. The instrument is installed at the end of the horizontal arm extending out of the tower to reduce the influence of the tower body. The sensor is installed vertically to reduce measurement errors.

[0066] In the process of ultrasonic wind temperature instrument obtaining turbulence data, external signal interference, data acquisition system error, instrument system drift, data recording error, etc. will inevitably cause various noises or interferences to be mixed in the original turbulence signal. Changes in the weather system itself will also cause the original turbulence signal to have certain trends. Therefore, after obtaining the original observation data, the following preprocessing is required to convert and correct the data.

[0067] Specifically, Figure 1 As shown, the preprocessing of raw observation data includes data group screening, outlier elimination, coordinate transformation, and trend regression.

[0068] 1) Data set screening

[0069] Eliminate the blank values ​​or missing values ​​in the original observation data and select the normal data group.

[0070] 2) Wild point elimination

[0071] According to the random error probability distribution theory, in general, in a set of equal-precision measurements, the probability of a large error is extremely small, and the probability that the absolute value of the error exceeds 3 times the standard is only 0.2%. For a set of turbulence observations of a certain atmospheric physical quantity, the accuracy is equal. Therefore, it can be considered that in a set of turbulence raw data, if the deviation of a certain data exceeds 3 times the standard deviation, the data point is regarded as a "bad value", commonly known as a "wild point", and the value is eliminated and the adjacent data points are interpolated to fill the point.

[0072] At the same time, the following rules must be followed when removing outliers: (1) After each outlier is removed, a judgment should be performed again until all outliers are removed; (2) Multiple outliers may appear at the same time, and the largest one should be removed first, and then the judgment should be performed; (3) When the number of outliers in a set of turbulent data exceeds 1%, the data set should be discarded.

[0073] 3) Coordinate transformation

[0074] During the turbulence observation test, once the instrument is installed, the u, v, and w directions are fixed. During the data processing process, the data must be rotated. The x direction after rotation is the average wind direction, called the downwind wind direction, the y direction is the crosswind direction, and the z direction is the vertical wind direction. The average values ​​of the wind speed components in the y and z directions are both zero, and the x, y, and z directions follow the right-hand rule.

[0075] 4) Trend Reversion

[0076] For the slow changes in weather systems or zero drift of instrument systems, as well as the noise introduced by the drop in battery power supply voltage, the trend regression method is used to eliminate them. Generally, the first-order regression method can effectively remove the zero drift of the observation system or the slow changes in the weather system. However, although its frequency range is very low, it may still overlap with the low-frequency band of the turbulence spectrum. In this case, the quadratic regression equation is used to remove the trend term.

[0077] After the original observation data are processed through the above steps, the standard deviation of turbulence observations (σ v , σ w ):

[0078]

[0079] Where M is the turbulence observation sample value of v and w, n is the number of observation samples, and the turbulence observation data 30 minutes before the exact time are usually taken as a group for analysis; σ is the turbulence observation standard deviation (i.e. the required turbulence observation standard deviation).

[0080] 3. Establish the relationship between atmospheric stability and standard deviation of turbulence observations

[0081] 1) Since temperature gradient and wind speed are the indicators of atmospheric stability, the standard deviation of turbulence observation (σ v , σ w ) has a clear linear relationship with wind speed and temperature gradient. v , σ w ) is used to perform binary linear regression fitting with the temperature gradient and wind speed data obtained from the corresponding conventional meteorological observations. The basic equation is as follows:

[0082] σ v =A0+A1×ΔT+A2×U (2-1)

[0083] σ w =A0`+A1`×ΔT+A2`×U (2-2)

[0084] In the formula, σ v , σ w is the standard deviation of lateral or vertical turbulence observations;

[0085] ΔT is the temperature gradient (take the gradient from 10m to 80m);

[0086] U is the wind speed at 10m; A0, A1, A2 (A0`, A1`, A2`) are the coefficients of the equation.

[0087] 2) Calculate the coefficients A0, A1, A2 (the method of A0`, A1`, A2` is the same as before, now with σ v For example)

[0088] The atmospheric stability is classified by the temperature gradient wind speed method, and the σ under different atmospheric stability is v , σ w It is divided into three groups: unstable (A to C), neutral (D), and stable (E to F). The sample capacity of each group is assumed to be n1, n2, and n3.

[0089] For a set of samples, by using the multivariate regression least squares method, let σ v =Yy i , A0=b0, A1=b1, A2=b2, ΔT=X i1 , U=X i2 , in the predicted values ​​of Y and the measured values ​​of the factor variables with a sample size of n, the binary linear regression equation is satisfied:

[0090] Yy i =b0+b1X i1 +b2X i2 (i=1,2,……n) (3)

[0091] The regression coefficients b0, b1, and b2 should be such that the sum of squares of the differences between all observed values ​​and the regression estimates is minimized. That is, Q is minimized:

[0092]

[0093] For a set of sample data, the estimate of the predicted value can be regarded as a vector, denoted as

[0094]

[0095] The regression equation that satisfies formula (3) can also be written in matrix form as y = Xb, where X is the X defined above. i1 , X i2 The factor matrix is ​​composed of, b is the regression coefficient vector, that is

[0096]

[0097] Q is actually a non-negative quadratic expression of b0, b1, and b2, so the minimum value must exist. According to the extreme value principle of differential calculus, we have:

[0098]

[0099] That is, X'Xb=X'y (8)

[0100] Formula (8) can be expanded by matrix multiplication to obtain the general form of the standard equation system:

[0101]

[0102] According to the general linear algebra method of solving the system of equations, b0, b1, and b2 can be solved. The three numbers obtained correspond to the equation coefficients A0, A1, and A2 respectively. This method can be used to obtain the coefficients of the three sets of equations respectively. Therefore, according to Formula 2-1, the standard deviation of turbulence observations under various atmospheric stability characteristics can be obtained.

[0103] 4. CALPUFF atmospheric diffusion parameter scheme optimization

[0104] The CALPUFF numerical model requires input data including plant site ground observation data, high-altitude observation data or mesoscale three-dimensional meteorological observation data, high-resolution land use data and refined grid schemes, and turbulence observation data required for the atmospheric diffusion parameter scheme.

[0105] According to the relationship between atmospheric stability and turbulence observation standard deviation obtained in step 3, in practical applications, the turbulence fluctuation standard deviation (σ v , σ w), write it into the PROFILE.DAT file under the format specified by the CALPUFF model, and finally select the atmospheric diffusion parameter option 1 for simulation calculation when performing CALPUFF diffusion calculation, so as to achieve a refined match between the measured atmospheric stability and the atmospheric diffusion parameters.

[0106] The present invention uses the universal turbulence observation standard deviation data based on stability conditions obtained by statistical analysis of conventional wind temperature observation data and turbulence observation data as the basic data for option 1 in the CALPUFF atmospheric diffusion parameter scheme. This data accurately reflects the atmospheric diffusion characteristics under the actual weather conditions of the observation area to a certain extent, and can be used in continuous simulation research on atmospheric diffusion laws within a certain range of the observation area. It has the following advantages: 1. Conventional data and turbulence observation data are used to more accurately reflect the atmospheric stability characteristics of the simulation area; 2. The correlation between atmospheric stability and turbulence observation data in the simulation area is established, and the relationship between the turbulence observation standard deviation that accurately reflects the atmospheric stability characteristics of the simulation area and conventional wind temperature observation data is constructed. Eliminate the dependence of the CALPUFF numerical model on high-precision turbulence observations; 3. The turbulence observation standard deviation calculated based on conventional wind temperature observation data is used as the input data for option 1 of the diffusion parameter scheme of the CALPUFF model, which can accurately reflect the atmospheric diffusion characteristics under various weather stability conditions in the plant site area.

[0107] The above embodiments are only for illustrating the technical concept and features of the present invention, and their purpose is to enable people familiar with this technology to understand the contents of the present invention and implement them accordingly. They cannot be used to limit the protection scope of the present invention. Any equivalent changes or modifications made according to the spirit of the present invention should be included in the protection scope of the present invention.

Claims

1. A method for optimizing diffusion parameters of a CALPUFF numerical model, characterized in that: The method comprises the following steps: obtaining atmospheric stability according to meteorological observation data; calculating the turbulence observation standard deviation according to turbulence observation data, and establishing a relationship between the atmospheric stability and the turbulence observation standard deviation; According to the relationship, the turbulence observation standard deviation corresponding to the meteorological observation data is obtained and used as the input data of the diffusion parameter of the CALPU FF model; The relationship between atmospheric stability and the standard deviation of turbulence observations is established according to the following steps: The turbulence observation standard deviation obtained by using the turbulence observation data statistics is used to perform binary linear regression fitting with the temperature gradient and wind speed data obtained from the corresponding meteorological observations. The basic equation form is: s v =A0+A1×ΔT+A2×U σ w =A0`+A1`×ΔT+A2`×U In the formula, σ v , σ w are the standard deviations of lateral or vertical turbulence observations, respectively; ΔT is the temperature gradient; U is the wind speed at 10 m; A0, A1, A2, A0 ` 、A1 ` 、A2 ` is the coefficient of the equation; The equation coefficients A0, A1, and A2 are calculated as follows: 1) The atmospheric stability is classified by the temperature gradient wind speed method, and the σ under different atmospheric stability is v , σ w It is divided into three groups: unstable, neutral, and stable. The sample size of each group is set to n1, n2, and n3; 2) For a set of samples, use the multivariate regression least squares method, let σ v、w =Yy i , A0=b0, A1=b1, A2=b2, ΔT=X i1 , U=X i2 , in the predicted values ​​of Y and the measured values ​​of the factor variables with a sample size of n, the binary linear regression equation is satisfied: Yy i =b0+b1X i1 +b2X i2 (i=1,2,……n) The regression coefficients b0, b1, and b2 should be such that the sum of squares of the differences between all observed values ​​and the regression estimates is minimized; that is, Q is minimized: For a set of sample data, the estimate of the predicted value is regarded as a vector, denoted as Satisfy the binary linear regression equation and rewrite it into matrix form: y = Xb, where X is the X defined above. i1 , X i2 The factor matrix is ​​composed of, b is the regression coefficient vector, that is Q is a non-negative quadratic expression of b0, b1, and b2, so the minimum value must exist. According to the extreme value principle of differential calculus, we have: i.e., X ’ Xb = X ’ y The above formula can be expanded by matrix multiplication to obtain the general form of the standard equation system: According to the general linear algebra method of solving systems of equations, b0, b1, and b2 can be solved. The three numbers obtained correspond to the equation coefficients A0, A1, and A2 respectively.

2. The optimization method according to claim 1, characterized in that: The turbulence observation standard deviation is calculated by the following formula Where M is the observed sample value of v and w, n is the number of observed samples, and σ is the standard deviation of turbulence observation.

3. The optimization method according to claim 2, characterized in that: The original observation data obtained from turbulence observations are corrected according to the steps of data group screening, wild point elimination, coordinate transformation, and trend regression to obtain observation sample values.

4. The optimization method according to claim 3, characterized in that: The outlier elimination is performed according to the following steps: in a set of raw turbulence observation data, if the deviation of a certain data exceeds 3 times the standard deviation, the value is eliminated and the adjacent data points are interpolated to fill in the gaps.

5. The optimization method according to claim 4, characterized in that: The wild point elimination follows the following rules: after each wild point is eliminated, a judgment should be performed again until all wild points are eliminated; if multiple wild points appear at the same time, the largest one should be eliminated first, and then the judgment should be performed; when the number of wild points in a group of turbulent raw data exceeds 1%, the group of data is discarded.

6. The optimization method according to claim 3, characterized in that: The coordinate transformation is performed according to the following steps: during the data processing, the data must be rotated. After the rotation, the x direction is the average wind direction, called the downwind wind direction, the y direction is the crosswind direction, and the z direction is the vertical wind direction. The average values ​​of the wind speed components in the y and z directions are both zero, and the x, y, and z directions follow the right-hand rule.

7. The optimization method according to claim 3, characterized in that: The trend regression is performed according to the following steps: the noise introduced by the slow weather system change or the zero drift of the instrument system, as well as the noise introduced by the battery power supply voltage drop, is eliminated by the trend regression method.

8. The optimization method according to claim 1, characterized in that: The atmospheric stability is statistically analyzed using the temperature gradient wind speed method.

Images