A method for optimizing parameters of weather radar wind field inversion

Abstract

The present invention provides a method for optimizing the wind field inversion parameters of a weather radar. The method comprises the following steps: obtaining radial wind speed data output by a weather radar; obtaining wind profile data output by a laser wind measuring radar; calculating wind field inversion parameters based on the radial wind speed data; and optimizing the wind field inversion parameters based on the wind profile data. The present invention optimizes the parameters of a wind field inversion algorithm based on radial data of a weather radar by using actual wind field data detected by a laser wind measuring radar, so that the wind field inverted by the optimized parameters is closer to the true value.

Description

Technical Field

[0001] The invention belongs to the technical field of meteorological radar detection, and in particular relates to a method for optimizing wind field inversion parameters of a weather radar. Background Art

[0002] The radial velocity detected by weather radar contains horizontal and vertical wind field information. Since it only detects radial velocity distribution and cannot directly obtain information on the three-dimensional wind field structure, theoretically, at least three radars are required to detect the three-dimensional wind field structure of the atmosphere simultaneously. However, joint detection by multiple radars requires synchronous observation of the area. This detection mode is not only costly, but also has a limited detection range. Common methods for inverting wind fields using a single weather radar include the velocity azimuth display (VAD) method and volume velocity processing (VVP) based on the assumption of a uniform wind field. The wind field model of the VAD method is not suitable for the inversion of complex wind fields. The VVP inversion method can theoretically obtain rich and complete three-dimensional wind field information, but the integrity of the wind field model and the complexity of the coefficient matrix during inversion will significantly affect the inversion results, thereby limiting the inversion of vertical wind speed. When more inversion parameters are selected to establish a wind field model, although the influence of model errors is reduced, the VVP algorithm will amplify the calculation error and observation error due to the excessive condition number of the complex coefficient matrix, thereby affecting the inversion accuracy. Similarly, if fewer inversion parameters are selected, the calculation errors caused by too many condition numbers can be avoided, but the wind field model is too simple, which will also limit its application in complex wind field inversion. There are various errors in the wind field data inverted using only the radial data of weather radar. There are many different inversion algorithms and improvements to these algorithms, but they are all based on the detection data of a single weather radar, and the inverted wind field may be different from the real wind field.

[0003] In view of this, the present invention proposes a weather radar wind field inversion parameter optimization method, which optimizes the parameters of the inversion algorithm by using the real wind field data detected by the laser wind radar, so that the inversion result can be closer to the real wind field. Summary of the invention

[0004] In order to solve the above problems existing in the prior art, the present invention provides a method for optimizing wind field inversion parameters of a weather radar.

[0005] In order to achieve the above object, the present invention adopts the following technical solutions.

[0006] A method for optimizing weather radar wind field inversion parameters comprises the following steps:

[0007] Obtain radial wind speed data output by weather radar;

[0008] Obtain wind profile data output by laser wind radar;

[0009] Calculating wind field inversion parameters based on the radial wind speed data;

[0010] The wind field inversion parameters are optimized based on the wind profile data.

[0011] Furthermore, the radial wind speed data is radial wind speeds at different pitch angles, different azimuth angles and different distances obtained by the weather radar when performing multi-pitch angle PPI scanning.

[0012] Furthermore, the wind profile data are wind speeds at different heights obtained when a laser wind radar performs wind profile scanning, including horizontal wind speeds at different heights, angles between horizontal wind directions and due north, and vertical wind speeds.

[0013] Furthermore, the wind profile data of the laser wind measuring radar is transformed into coordinates and the wind field is decomposed to obtain the wind speed data at different spatial positions, as well as the coordinates (x, y, z) and wind speed components u, v, and w of the particles in the spatial rectangular coordinate system O-XYZ; the origin O of the spatial rectangular coordinate system is the location of the weather radar, and the positive directions of the X, Y, and Z axes are due north, due east, and vertically upward, respectively; u, v, and w are the components of the wind speed in the directions of the X, Y, and Z axes, respectively; the calculation formulas for the coordinates (x, y, z) and the wind speed components u, v, and w are:

[0014] x=x lidar , y=y liadar , z=z lidar +h;

[0015] u=V h sinθ h , v = V h cosθ h , w=V v ;

[0016] In the formula, (x lidar ,y liadar ,z lidar ) is the position coordinate of the laser wind radar under O-XYZ, h is the height of the wind profile data relative to the laser wind radar, V h is the horizontal wind speed, θ h V is the angle between the horizontal wind direction and the north direction. v is the vertical wind speed.

[0017] Furthermore, the wind field inversion parameters are VVP wind field inversion algorithm parameters, including: particle velocity components u0, v0, w0 at the center of the inversion analysis area, and the coefficients of change of wind speed components u, v, w in the X, Y, and Z axis directions with distance u x 、v x 、w x 、u y、v y 、w y 、u z 、v z 、w z .

[0018] Furthermore, the method for calculating the VVP wind field inversion algorithm parameters includes:

[0019] The radial velocity of the particles observed by the weather radar is expressed as:

[0020]

[0021] in:

[0022] u=u0+u x (x-x0)+u y (y-y0)+u z (z-z0) (2)

[0023] v=v0+v x (x-x0)+v y (y-y0)+v z (z-z0) (3)

[0024] w=w0+w x (x-x0)+w y (y-y0)+w z (z-z0) (4)

[0025] Where V r is the radial velocity, (x0, y0, z0) is the coordinate of the center of the inversion analysis area, θ, are the azimuth and elevation angles of the weather radar detection direction respectively;

[0026] Substituting equations (2), (3), and (4) into equation (1), we obtain:

[0027]

[0028] In the formula,

[0029] X1=u0,X2=v0,X3=w0,X4=u x , X5=v x ,

[0030] X6=w x , X7=u y , X8=v y , X9=w y , X 10 =u z, X 11 =v z , X 12 =w z ;

[0031] Get n different radial velocities V r And the corresponding azimuth angle θ, pitch angle Substituting the coordinates (x, y, z) into equation (5) yields the following matrix equation:

[0032]

[0033] make Formula (6) can be written as The solution is: Thus, the VVP wind field inversion algorithm parameters are obtained.

[0034] Furthermore, the gradient descent method is used to optimize the parameters of the VVP wind field inversion algorithm.

[0035] Furthermore, the gradient descent method is used to invert the VVP wind field algorithm parameters u0, u x 、u y 、u z Methods for optimization include:

[0036] Construct the cost function:

[0037] e u =[u-(u0+u x (x-x0)+u y (y-y0)+u z (z-z0))] 2 (7)

[0038] In the formula, e u is the cost function value, u is the north direction velocity component of the particle with coordinates (x, y, z) obtained by the laser wind radar;

[0039] u0 and u obtained by VVP wind field inversion algorithm x 、u y 、u z is the initial value of the iteration, and u0 and u are updated as follows: x 、u y 、u z :

[0040]

[0041] In the formula, u′0, u′ x , u′ y , u′ z are u0 and u0 of the previous iteration step respectively. x 、uy 、u z The updated value, α is the learning rate, de u / du0、de u / du x 、de u / du y 、de u / du z To pass e u Find the gradient value obtained by partial differentiation;

[0042] When e u When the value of is less than the set threshold or the number of iteration steps is less than the set threshold, the iteration is stopped and the optimized u0, u x 、u y 、u z .

[0043] Compared with the prior art, the present invention has the following beneficial effects.

[0044] The present invention obtains radial wind speed data output by a weather radar, obtains wind profile data output by a laser wind radar, calculates wind field inversion parameters based on the radial wind speed data, and optimizes the parameters based on the wind profile data, thereby realizing the optimization of the weather radar wind field inversion parameters based on the laser wind radar data. The present invention optimizes the wind field inversion algorithm parameters based on the radial data of the weather radar by using the actual wind field data detected by the laser wind radar, so that the wind field inverted by the optimized parameters is closer to the true value. BRIEF DESCRIPTION OF THE DRAWINGS

[0045] Figure 1 The present invention is a flowchart of a method for optimizing wind field inversion parameters of a weather radar according to an embodiment of the present invention.

[0046] Figure 2 It is a block diagram of the hardware structure of the present invention. DETAILED DESCRIPTION

[0047] In order to make the purpose, technical solution and advantages of the present invention clearer and more understandable, the present invention is further described below in conjunction with the accompanying drawings and specific implementation methods. 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 making creative work are within the scope of protection of the present invention.

[0048] Figure 1 This is a flow chart of a method for optimizing wind field inversion parameters of a weather radar according to an embodiment of the present invention, comprising the following steps:

[0049] Step 101, obtaining radial wind speed data output by a weather radar;

[0050] Step 102, obtaining wind profile data output by a laser wind radar;

[0051] Step 103, calculating wind field inversion parameters based on the radial wind speed data;

[0052] Step 104: Optimize wind field inversion parameters based on the wind profile data.

[0053] This embodiment provides a method for optimizing the inversion parameters of a weather radar wind field. The hardware equipment for implementing the method mainly includes a weather radar, a laser wind radar, and a client computer. Figure 2 As shown. The weather radar model is 713, and the laser wind radar model is FC-Ⅲ. The equipment is deployed in an open space (such as an airport). Both radars are connected to a 4G router via a network cable, and the router is connected to the Internet. When working, the weather radar performs repeated multi-pitch angle PPI scanning, and the laser wind radar performs repeated wind profile scanning. The client computer receives the data uploaded by the two radars through the Internet, and calculates and optimizes the wind field inversion parameters based on the data.

[0054] In this embodiment, step 101 is mainly used to obtain radial wind speed data output by weather radar. Weather radar is a type of meteorological radar. Weather radar is the main tool for monitoring and warning severe convective weather. Its working principle is to detect the spatial distribution and vertical structure of precipitation by emitting a series of pulsed electromagnetic waves and utilizing the scattering and absorption of electromagnetic waves by precipitation particles such as clouds, rain, and snow. The wavelengths commonly used by weather radars are mostly in the range of 1 to 10 cm. Due to the low attenuation of the 10 cm wavelength, typhoons, heavy rains and hail can be better detected. Weather radar is an important means of detecting small and medium-scale precipitation, which helps to understand the wind field structure in the precipitation area.

[0055] In this embodiment, step 102 is mainly used to obtain wind profile data output by the laser wind radar. The radial velocity detected by the weather radar contains horizontal wind field and vertical wind field information. Since it only detects radial velocity distribution, it cannot directly obtain information about the three-dimensional wind field structure. Therefore, in theory, at least three radars are required to detect the three-dimensional wind field structure of the atmosphere at the same time. However, the joint detection of multiple radars requires synchronous observation areas. This detection mode is not only expensive, but also has a limited detection range. To this end, this embodiment combines a weather radar and a laser wind radar to obtain information about the three-dimensional wind field structure. The laser wind radar uses the Doppler shift principle to derive wind speed and wind direction. First, the laser signal is emitted into the air at a fixed wavelength and frequency. When the laser signal contacts the precipitation particles moving with the wind in the air, the signal is scattered back and received by the wind radar. The frequency of the returned signal will change compared with the emission frequency, which is the Doppler shift. There is a corresponding functional relationship between the frequency difference of the round-trip signal and the airflow velocity, thereby deriving the radial moving speed of the airflow along the laser. The laser wind radar emits multiple signal sources at the same time, first decomposing each radial velocity horizontally and vertically into components, and finally synthesizing all horizontal components into a horizontal wind speed vector with wind direction, and all vertical components into a vertical wind speed vector. The wind profile (wind speed vs. height curve) mode of the laser wind radar can detect the wind profile above the radar, and the data obtained is the horizontal wind speed, horizontal wind direction and vertical wind speed of each point at different heights.

[0056] In this embodiment, step 103 is mainly used to calculate the wind field inversion parameters based on the radial wind speed data. There are many methods for inverting the wind field using a single weather radar, such as the velocity azimuth display (VAD) method based on the uniform wind field assumption, volume velocity processing (VVP), vorticity-divergence method, VAP extended application method for inverting the near-center wind field of a tropical cyclone, reflectivity factor and radial velocity joint constraint method, etc., based on various wind field inversion algorithms under simple assumptions. Different inversion algorithms obtain different inversion parameters, and this embodiment does not limit specific inversion algorithms and inversion parameters.

[0057] In this embodiment, step 104 is mainly used to optimize the inversion parameters obtained in step 103 based on the wind profile data. Since there are various errors in the wind field data inverted using only the radial data of the weather radar, there are many different inversion algorithms and improvements to these algorithms, but they are all based on the detection data of a single weather radar, and there is a gap between the inverted wind field and the real wind field. In this embodiment, the real wind field data detected by the laser wind radar is used to optimize the parameters of the inversion algorithm, so that the inversion result can be closer to the real wind field.

[0058] As an optional embodiment, the radial wind speed data is radial wind speeds at different pitch angles, different azimuth angles and different distances obtained by a weather radar when performing multi-pitch angle PPI scanning.

[0059] This embodiment limits the radial wind speed data. The weather radar works in the PPI (Plane Position Indicator) scanning mode, and its scanning method is to perform circular scanning detection at a fixed elevation angle. Through multi-pitch angle PPI scanning, radial wind speeds at different pitch angles, different azimuths and different distances can be obtained.

[0060] As an optional embodiment, the wind profile data is wind speeds at different heights obtained when a laser wind measuring radar performs wind profile scanning, including horizontal wind speeds at different heights, angles between horizontal wind directions and due north, and vertical wind speeds.

[0061] This embodiment defines the wind profile data. The wind profile is a curve of wind speed changing with height, which is used to study the wind speed law in the atmospheric boundary layer. The wind profile mode of the laser wind radar can detect the wind profile above the radar, and the data obtained is the horizontal wind speed, horizontal wind direction and vertical wind speed of each point at different heights.

[0062] As an optional embodiment, the wind profile data of the laser wind measuring radar is subjected to coordinate transformation and wind field decomposition to obtain wind speed data at different spatial positions, as well as the coordinates (x, y, z) and wind speed components u, v, and w of the particle in the spatial rectangular coordinate system O-XYZ; the origin O of the spatial rectangular coordinate system is the location of the weather radar, and the positive directions of the X, Y, and Z axes are due north, due east, and vertically upward, respectively; u, v, and w are the components of the wind speed in the directions of the X, Y, and Z axes, respectively; the calculation formulas for the coordinates (x, y, z) and the wind speed components u, v, and w are:

[0063] x=x lidar , y=y liadar , z=z lidar +h;

[0064] u=V h sinθ h , v = V h cosθ h , w=V v ;

[0065] In the formula, (x lidar ,y liadar ,z lidar ) is the position coordinate of the laser wind radar under O-XYZ, h is the height of the wind profile data relative to the laser wind radar, V h is the horizontal wind speed, θ h V is the angle between the horizontal wind direction and the north direction.v is the vertical wind speed.

[0066] This embodiment provides a technical solution for coordinate transformation and wind field decomposition of the wind profile data of the laser wind measuring radar. Since the wind profile data of the laser wind measuring radar is used to optimize the wind field inversion parameters of the weather radar, it is necessary to convert the wind profile data of the laser wind measuring radar to the coordinate system O-XYZ of the weather radar. The origin O of the coordinate system O-XYZ is the location of the weather radar, and the positive directions of the X, Y, and Z axes are due north, due east, and vertical upward, respectively. According to the position coordinates (x lidar ,y liadar ,z lidar ) and the wind profile data height h, it is easy to obtain the coordinates (x, y, z) of the particle in the spatial rectangular coordinate system O-XYZ: x = x lidar , y=y liadar , z=z lidar +h. According to the horizontal wind speed V h , the angle θ between the horizontal wind speed direction and the due north direction h and vertical wind speed V v , it is easy to get the wind speed components u, v, w in the X, Y, and Z directions: u = V h sinθ h , v = V h cosθ h , w=V v .

[0067] As an optional embodiment, the wind field inversion parameters are VVP wind field inversion algorithm parameters, including: particle velocity components u0, v0, w0 at the center of the inversion analysis area, and the coefficients of change of wind speed components u, v, w in the X, Y, and Z axis directions with distance u x 、v x 、w x 、u y 、v y 、w y 、u z 、v z 、w z .

[0068] This embodiment provides a specific wind field inversion algorithm and its corresponding inversion algorithm parameters. This embodiment adopts the VVP wind field inversion algorithm. The VVP algorithm divides the radar scanning data into multiple small fan-shaped analysis areas. Under the assumption that the wind field in each analysis area is linear, multiple regression analysis is used to estimate the average three-dimensional wind, divergence, deformation and other wind parameters of each analysis area. The VVP algorithm uses multiple PPI scanning data of weather radar for inversion, and takes a three-dimensional space composed of radial, tangential and vertical directions as the analysis area. Since the selected analysis object is a three-dimensional spatial area, it is suitable for high elevation angles. The following assumptions are made in the selected analysis area: (1) The wind field structure is linearly distributed; (2) The wind field does not change with time during the radar scan. There are 12 VVP wind field inversion algorithm parameters, namely, the particle velocity components u0, v0, w0 in the center of the inversion analysis area, the change coefficient u of the wind speed components u, v, w in the X, Y, and Z axis directions with distance x 、v x 、w x 、u y 、v y 、w y 、u z 、v z 、w z .

[0069] As an optional embodiment, the method for calculating the VVP wind field inversion algorithm parameters includes:

[0070] The radial velocity of the particles observed by the weather radar is expressed as:

[0071]

[0072] in:

[0073] u=u0+u x (x-x0)+u y (y-y0)+u z (z-z0) (2)

[0074] v=v0+v x (x-x0)+v y (y-y0)+v z (z-z0) (3)

[0075] w=w0+w x (x-x0)+w y (y-y0)+w z (z-z0) (4)

[0076] Where V r is the radial velocity, (x0, y0, z0) is the coordinate of the center of the inversion analysis area, θ, are the azimuth and elevation angles of the weather radar detection direction respectively;

[0077] Substituting equations (2), (3), and (4) into equation (1), we obtain:

[0078]

[0079] In the formula,

[0080] X1=u0,X2=v0,X3=w0,X4=u x , X5=v x ,

[0081] X6=w x , X7=u y , X8=v y , X9=w y , X 10 =u z , X 11 =v z , X 12 =w z ;

[0082] Get n different radial velocities V r And the corresponding azimuth angle θ, pitch angle Substituting the coordinates (x, y, z) into equation (5) yields the following matrix equation:

[0083]

[0084] make Formula (6) can be written as The solution is: Thus, the VVP wind field inversion algorithm parameters are obtained.

[0085] This embodiment provides a technical solution for calculating the parameters of the VVP wind field inversion algorithm. First, the radial velocity V of the particles observed by the weather radar is listed. r With wind speed components u, v, w and azimuth angle θ, pitch angle The relationship between u, v and w is given in equation (1), and the formulas for calculating u, v and w based on the 12 inversion parameters are given in equations (2) to (4). Equations (2) to (4) are obtained based on the assumption that the wind field in the analysis area is linearly distributed. Substituting equations (2) to (4) into equation (1) yields V r ,θ, The linear relationship between the particle position coordinates (x, y, z) and the 12 inversion parameters is shown in the simplified expression (5). i is the i-th parameter Xi The coefficients are i=1,2,...,12. n groups of V measured by weather radar are r ,θ, Substituting the (x, y, z) values ​​into equation (5), we get a system of n equations, which can be written as a matrix equation: Solving this matrix equation yields X, and thus the VVP wind field inversion algorithm parameters.

[0086] As an optional embodiment, the VVP wind field inversion algorithm parameters are optimized using a gradient descent method.

[0087] This embodiment uses the gradient descent method for parameter optimization. The gradient descent method is an iterative optimization algorithm used to find the minimum value of the objective function. The core idea is to gradually reduce the value of the objective function by continuously adjusting the value of the parameter. The basic steps of the gradient descent method include selecting the initial parameter value, calculating the gradient, updating the parameter, iterating and outputting the result. The gradient is the partial derivative of the objective function in each parameter dimension. When updating the parameters, the parameters are adjusted in the opposite direction of the gradient at a certain learning rate. The learning rate determines the step size of each parameter update. The advantage of the gradient descent method is that it is applicable to large-scale data sets and can find the local minimum of the objective function.

[0088] As an optional embodiment, the VVP wind field inversion algorithm parameters u0, u x 、u y 、u z Methods for optimization include:

[0089] Construct the cost function:

[0090] e u =[u-(u0+u x (x-x0)+u y (y-y0)+u z (z-z0))] 2 (7)

[0091] In the formula, e u is the cost function value, u is the north direction velocity component of the particle with coordinates (x, y, z) obtained by the laser wind radar;

[0092] u0 and u obtained by VVP wind field inversion algorithm x 、u y 、u z is the initial value of the iteration, and u0 and u are updated as follows: x 、u y 、u z :

[0093]

[0094] In the formula, u′0, u′ x , u′ y , u′ z are u0 and u0 of the previous iteration step respectively. x 、u y 、u z The updated value, α is the learning rate, de u / du0、de u / du x 、de u / du y 、de u / du z To pass e u Find the gradient value obtained by partial differentiation;

[0095] When e u When the value of is less than the set threshold or the number of iteration steps is less than the set threshold, the iteration is stopped and the optimized u0, u x 、u y 、u z .

[0096] This embodiment gives the parameters u0, u x 、u y 、u z Among the 12 parameters to be optimized, the parameters of the u, v, and w components each account for 4. The 4 parameters of the same component are optimized using the same iterative process, and the optimization methods of the 4 parameters of different components are the same. Therefore, this embodiment only gives the 4 parameters u0, u x 、u y 、u z Optimization scheme.

[0097] First, construct the cost function, as shown in formula (7). The value of the cost function is equal to the u value measured by the laser wind radar and the u value based on the parameters to be optimized u0, u x 、u y 、u z The square of the difference in u values ​​calculated. The optimization process is to make the cost function small enough by continuously changing the value of the parameter to be optimized, that is, to make the error between the measured value and the calculated value small enough to meet the accuracy requirements. Then determine the initial value of the iteration, and change the parameter to be optimized according to the update formula at each iteration step until the value of the cost function is less than the set threshold or reaches the set maximum number of iterations, and the iteration ends to obtain the optimized parameters. In this embodiment, the parameter value obtained by the VVP wind field inversion algorithm is used as the initial value of the iteration, and the parameter update formula is designed according to the gradient descent method, such as formulas (8) to (11). The update step size in the updated formula is equal to the product of the learning rate and the gradient of the corresponding parameter. The gradient value is obtained by the cost function e uAccording to formula (7), the cost function e u It is a multivariate quadratic function of the parameters to be optimized. The method of calculating partial differentials is very simple, and the specific expression of the gradient is not given here.

[0098] The above is only a specific embodiment of the present invention, but the protection scope of the present invention is not limited thereto. Any changes or substitutions that can be easily thought of by a person skilled in the art within the technical scope disclosed by the present invention should be included in the protection scope of the present invention. Therefore, the protection scope of the present invention shall be subject to the protection scope of the claims.

Claims

1. A method for optimizing weather radar wind field inversion parameters, characterized in that: The following steps are involved: Obtain radial wind speed data output by weather radar; Obtain wind profile data output by laser wind radar; Calculating wind field inversion parameters based on the radial wind speed data; The wind field inversion parameters are optimized based on the wind profile data.

2. The weather radar wind field inversion parameter optimization method according to claim 1, characterized in that: The radial wind speed data are radial wind speeds at different pitch angles, different azimuth angles and different distances obtained by the weather radar when performing multi-pitch angle PPI scanning.

3. The weather radar wind field inversion parameter optimization method according to claim 1, characterized in that: The wind profile data are wind speeds at different heights obtained when a laser wind radar performs wind profile scanning, including horizontal wind speeds at different heights, angles between horizontal wind directions and the north direction, and vertical wind speeds.

4. The weather radar wind field inversion parameter optimization method according to claim 3, characterized in that: The wind profile data of the laser wind radar is transformed into coordinates and the wind field is decomposed to obtain the wind speed data at different spatial positions, as well as the coordinates (x, y, z) and wind speed components u, v, and w of the particles in the spatial rectangular coordinate system O-XYZ; the origin O of the spatial rectangular coordinate system is the location of the weather radar, and the positive directions of the X, Y, and Z axes are due north, due east, and vertically upward, respectively; u, v, and w are the components of the wind speed in the directions of the X, Y, and Z axes, respectively; the calculation formulas for the coordinates (x, y, z) and the wind speed components u, v, and w are as follows: x=x lidar ,y=y liadar ,z=z lidar +h; u=V h sinθ h ,v=V h cosθ h ,w=V v ; In the formula, (x lidar ,y liadar ,z lidar ) is the position coordinate of the laser wind radar under O-XYZ, h is the height of the wind profile data relative to the laser wind radar, V h is the horizontal wind speed, θ h V is the angle between the horizontal wind direction and the north direction. v is the vertical wind speed.

5. The weather radar wind field inversion parameter optimization method according to claim 4, characterized in that: The wind field inversion parameters are the VVP wind field inversion algorithm parameters, including: particle velocity components u0, v0, w0 at the center of the inversion analysis area, and the coefficient of change of wind speed components u, v, w in the X, Y, and Z axis directions with distance u x 、v x 、w x 、u y 、v y 、w y 、u z 、v z 、w z .

6. The weather radar wind field inversion parameter optimization method according to claim 5, characterized in that: Methods for calculating VVP wind field inversion algorithm parameters include: The radial velocity of the particles observed by the weather radar is expressed as: in: u=u0+u x (x-x0)+u y (y-y0)+u z (z-z0) (2) v=v0+v x (x-x0)+v y (y-y0)+v z (z-z0) (3) w=w0+w x (x-x0)+in y (y-y0)+w z (z-z0) (4) Where V r is the radial velocity, (x0, y0, z0) is the coordinate of the center of the inversion analysis area, θ, are the azimuth and elevation angles of the weather radar detection direction respectively; Substituting equations (2), (3), and (4) into equation (1), we obtain: In the formula, X1=u0,X2=v0,X3=w0,X4=u x ,X5=v x , X6=w x ,X7=u y ,X8=v y ,X9=w y ,X 10 =u z ,X 11 =v z ,X 12 =w z ; Get n different radial velocities V r And the corresponding azimuth angle θ, pitch angle Substituting the coordinates (x, y, z) into equation (5) yields the following matrix equation: make Formula (6) can be written as The solution is: Thus, the VVP wind field inversion algorithm parameters are obtained.

7. The weather radar wind field inversion parameter optimization method according to claim 6, characterized in that: The gradient descent method is used to optimize the parameters of the VVP wind field inversion algorithm.

8. The weather radar wind field inversion parameter optimization method according to claim 7, characterized in that: The gradient descent method is used to invert the VVP wind field algorithm parameters u0, u x 、u y 、u z Methods for optimization include: Construct the cost function: yes u =[u-(u0+u x (x-x0)+u y (y-y0)+u z (z-z0))] 2 (7) In the formula, e u is the cost function value, u is the north direction velocity component of the particle with coordinates (x, y, z) obtained by the laser wind radar; u0 and u obtained by VVP wind field inversion algorithm x 、u y 、u z is the initial value of the iteration, and u0 and u are updated as follows: x 、u y 、u z : In the formula, u′0, u′ x , u′ y , u′ z are u0 and u0 of the previous iteration step respectively. x 、u y 、u z The updated value, α is the learning rate, de u / du0、de u / du x 、de u / du y 、de u / du z To pass e u Find the gradient value obtained by partial differentiation; When e u When the value of is less than the set threshold or the number of iteration steps is less than the set threshold, the iteration is stopped and the optimized u0, u x 、u y 、u z .

Images