Multi-angle laser radar vertical two-dimensional wind field inversion method

Abstract

The invention discloses a multi-angle laser radar vertical two-dimensional wind field inversion method. The method comprises the following steps: firstly, setting a measurement center and sampling parameters of a measurement area; acquiring multi-angle radial wind speed data; introducing a multi-point local sampling strategy to each inversion center point, and establishing a local multi-constraint inversion equation set by selecting a plurality of observation points close to the spatial position of the center point and combining the observation angle and radial wind speed information; according to the method, the limitation that a traditional wind field inversion method depends on a large-range overall consistency hypothesis is broken through, a local linear modeling method is adopted, small-scale changes, local non-uniformity and turbulence disturbance existing in an actual wind field can be better adapted, and the adaptability and modeling precision of an inversion model to a complex wind field structure are improved; the anti-noise capability, the error control and the inversion convergence are remarkably improved; the invention further provides a system, equipment and a storage medium for implementing the method.

Description

Technical Field

[0001] This invention relates to the field of atmospheric remote sensing and lidar technology, and in particular to a method for inverting vertical two-dimensional wind fields using multi-angle lidar. Background Technology

[0002] LiDAR (Light Detection and Ranging) has become an important tool for atmospheric wind field detection due to its advantages of high spatiotemporal resolution and non-contact measurement. Doppler lidar obtains the radial wind speed along the line of sight by detecting the Doppler frequency shift of aerosol or molecular scattering signals. Traditional wind field inversion methods (such as geometric analytical methods, least squares methods, and tomography) are usually based on the assumption of a globally uniform wind field, which makes it difficult to accurately capture vertical wind shear and turbulent structures in complex terrain or boundary layers, resulting in large errors in vertical wind speed inversion.

[0003] Current mainstream wind field inversion methods have the following problems: (1) The geometric analytical method relies on the strict coplanarity of the laser path. In actual systems, due to beam divergence and asynchronous scanning, geometric errors will be introduced. Especially in the low signal-to-noise ratio region, the vertical wind speed inversion is easily affected by noise amplification.

[0004] (2) Tomography has high accuracy in theory, but it requires high temporal resolution data support, otherwise it is easily affected by ill-posed problems and has high computational complexity.

[0005] (3) Least squares method, for example, patent application CN112433233B proposes a method for inverting sea surface wind speed using particle swarm algorithm. Although the calculation is simple, when the laser path in the vertical direction is sparse (such as elevation angle > 60°), the normal matrix is ​​close to singular, the inversion result is unstable, and it lacks physical constraints, which can easily lead to false vertical motion.

[0006] (4) Deep learning methods rely on a large amount of training data, have limited generalization ability, and are difficult to adapt to new environments.

[0007] Therefore, a common drawback of existing methods is that they all assume a uniform distribution of the wind field globally. However, actual wind fields (especially in the vertical direction) often exhibit local linear variation characteristics. Traditional methods cannot effectively model such microstructures, resulting in insufficient accuracy in vertical wind speed inversion. Therefore, there is an urgent need for an algorithm for high-precision inversion of wind speed vectors in local areas. Summary of the Invention

[0008] To overcome the shortcomings of traditional wind field inversion methods, this invention aims to provide a multi-angle lidar vertical two-dimensional wind field inversion method. This method introduces a first-order variation term of the wind speed component with respect to the spatial coordinates within a local spatial range, establishing a local linear wind field mathematical model that better conforms to the microscopic characteristics of the actual wind field. Unlike traditional wind field inversion methods based on the assumption of global consistency, the method proposed in this invention can effectively capture the variation law of the wind field within a small scale range, improving the modeling and inversion capabilities for complex non-uniform wind fields.

[0009] To achieve the above objectives, the technical solution of the present invention is as follows: A method for inverting vertical two-dimensional wind fields using multi-angle lidar is as follows: Within the wind field area to be measured, the location of the measurement center point is determined, and the spatial scale parameters of the local inversion area are set. The spatial scale parameters include the sampling radius, elevation scanning range, angular resolution, and distance resolution. The wind field area to be measured is scanned in a multi-elevation fan shape using a multi-angle coherent Doppler lidar scanning system to obtain radial wind speed data containing multiple scanning angles; Taking the center point to be measured as the core, based on the spatial scale parameters, multiple sampling points with uniform spatial distribution and different elevation angles are selected in the vicinity of the center point to be measured; A local linear wind field model is established, and the horizontal wind speed in the local linear wind field model is used as a basis for this model. and vertical wind speed A first-order linear variation model is satisfied, and a set of wind speed observation equations is constructed. The wind speed observation equations were solved using the least squares method and radial wind speed data to obtain the wind speed parameters at the inversion center point. The output inversion result is the local wind field distribution obtained based on the wind speed parameters at the inversion center point.

[0010] The acquisition of radial wind speed data comprising multiple scanning angles specifically involves, at each elevation angle Below, the radar collects a series of radial distance data. The measurement points form a two-dimensional observation matrix. The two-dimensional observation matrix As radial wind speed data.

[0011] Using the center point to be measured as the core, multiple sampling points with uniform spatial distribution and different elevation angles are selected within its vicinity, specifically including: Using the center point to be measured as the core, and constrained by the fact that the spatial distance between each sampling point and the center point to be measured does not exceed the preset maximum spatial scale parameter, sampling points are selected. Among them, the spatial distance between each sampling point does not exceed the set spatial distance resolution, and they are evenly distributed in the elevation direction or spatial coordinates.

[0012] The establishment of the local linear wind field model specifically involves: Within a small area based on the center point, where the small area refers to the coverage area of ​​the center point and multiple selected sampling points, assuming a horizontal wind speed... and vertical wind speed It satisfies a first-order linear change model: in: , These are the horizontal and vertical wind speeds at the center point, respectively. , , , This represents the first-order gradient of wind speed in the spatial direction; The location of the sampling point; This represents the location of the inversion center point.

[0013] The specific steps for constructing the wind speed observation equation set are as follows: Based on the observed elevation angle of each sampling point and spatial location that is, radial distance By combining the local linear wind field model, the wind field prediction formula for each sampling point is obtained: The equations for the selected n sampling points are then reorganized into a matrix form: in: Let the vector be the parameters to be determined. A is a coefficient matrix composed of the location and elevation angle information of each sampling point; b is the radial wind speed observation vector at the sampling point; This is radial wind speed data.

[0014] The wind speed parameters obtained at the inversion center point are specifically as follows: in: , The horizontal and vertical wind speeds at the center point obtained from the inversion; , , , This represents the local wind speed gradient, used to describe the trend of local wind field changes.

[0015] The output inversion result specifically refers to the horizontal wind speed at each inversion center point. and vertical wind speed As the output of the local wind field inversion, combined with local gradient information , , , By interpolating or extrapolating within a small surrounding area, a more complete local wind field distribution can be obtained.

[0016] The present invention also includes: A system comprising hardware components and embedded software programs is capable of performing the entire process of a multi-angle lidar vertical two-dimensional wind field inversion method.

[0017] An apparatus comprising: Memory: Used to store the computer program for the multi-angle lidar vertical two-dimensional wind field inversion method described above; Processor: Used to implement the multi-angle lidar vertical two-dimensional wind field inversion method when executing the computer program.

[0018] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for inverting a vertical two-dimensional wind field using a multi-angle lidar.

[0019] Compared with the prior art, the beneficial effects of the present invention are: (1) The present invention adopts a local sampling strategy, selecting several neighboring sampling points to invert the wind speed at a central point, which has the advantages of enhancing inversion stability and reducing ill-conditioned problems. This makes the inversion process more focused and more resistant to interference, while avoiding problems such as excessively large condition numbers and matrix invertibility that may occur in global inversion.

[0020] (2) The present invention adopts a local linear wind field model, which has the advantages of balancing model accuracy and computational efficiency. It assumes that the wind speed varies linearly along the spatial direction near the center point, which can reflect the small-scale non-uniformity of the wind field and avoid the computational cost and instability caused by using complex nonlinear models. It is suitable for real-time processing under multi-angle observation.

[0021] (3) The present invention constructs a set of wind speed observation equations and solves them using the least squares method, which has the advantages of strong versatility and stable solution. The observation data of each sampling point are organized into a set of linear equations and the parameters are inverted using the classical least squares method, so that the whole process has a good mathematical foundation and robustness, and can still obtain better results when there are errors in wind speed measurement.

[0022] In summary, this invention introduces a multi-point local sampling strategy at each inversion center point. By selecting several observation points that are spatially close to the center point and combining the observation angle and radial wind speed information, a local multi-constraint inversion equation set is established. This can better adapt to the small-scale changes, local non-uniformity and turbulent disturbances in the actual wind field, and improve the adaptability and modeling accuracy of the inversion model to complex wind field structures. Attached Figure Description

[0023] Figure 1 It is a constructed real-time horizontal wind speed map.

[0024] Figure 2 It is a horizontal wind speed map based on the inversion of the local wind field.

[0025] Figure 3 It is a horizontal wind speed map inverted based on the traditional least squares method.

[0026] Figure 4 This is a diagram of the horizontal wind speed inversion error (RMSE) based on local wind field inversion.

[0027] Figure 5 This is a graph showing the horizontal wind speed inversion error (RMSE) based on the traditional least squares method.

[0028] Figure 6 It is a constructed real-time horizontal wind speed map.

[0029] Figure 7 It is a horizontal wind speed map based on the inversion of the local wind field.

[0030] Figure 8 It is a horizontal wind speed map inverted based on the traditional least squares method.

[0031] Figure 9 This is a diagram of the horizontal wind speed inversion error (RMSE) based on local wind field inversion.

[0032] Figure 10 This is a graph showing the horizontal wind speed inversion error (RMSE) based on the traditional least squares method.

[0033] Figure 11 This is a schematic diagram illustrating the principle of selecting local inversion sampling points in step 3.

[0034] Figure 12 This is a schematic diagram illustrating the principle of obtaining radial wind speed using a wind-measuring lidar.

[0035] Figure 13 It is the radial wind speed of the actual wind field.

[0036] Figure 14 It is the radial wind speed detected by a lidar system. Detailed Implementation

[0037] The present invention will now be further described with reference to the accompanying drawings.

[0038] This invention comprehensively considers that the horizontal wind speed in a real wind field is a linear shear, while the vertical wind speed is a wave field model. The linear shear of the horizontal wind speed reflects the characteristic of wind speed increasing with altitude, which is consistent with the actual atmospheric boundary layer structure. Meanwhile, the wave field of the vertical wind speed can simulate the influence of turbulence and terrain disturbance on the wind field, improving the realism of the simulation.

[0039] Simulation: Using an elevation angle range of 20° to 60°, angular resolution... Detection distance Spatial resolution from 30 meters to 150 meters A two-dimensional observation grid was constructed; a real wind field was set up, and the traditional least squares decomposition algorithm and the proposed multi-angle lidar vertical two-dimensional wind field inversion method were used to perform inversion and solution, and the RMSE of the horizontal wind speed and vertical wind speed in the inversion results were calculated respectively.

[0040] The specific operation of the present invention is described in detail below: A method for inverting vertical two-dimensional wind fields using multi-angle lidar, the specific steps of which are as follows: Step 1: Set the measurement area and sampling parameters Within the area of ​​the wind field to be measured, determine the location of the measurement center point and set the spatial scale parameters for the local inversion area: First, the wind field area to be measured and the observation parameters are set. The elevation scanning range is 20° to 60°, the angular resolution is 0.1°, the detection distance range is 30 meters to 150 meters, and the spatial distance resolution is 1.5 meters. Based on the above angular and distance resolutions, the retrieved wind field area is divided into a two-dimensional grid to generate the spatial location and parameters of each sampling point.

[0041] Step 2: Multi-angle radial wind speed data acquisition The wind field area to be measured was systematically scanned using a multi-angle coherent Doppler lidar scanning system according to the aforementioned set elevation angle and distance parameters, and observation data was collected. For each sampling point, its corresponding spatial coordinates were recorded. This is obtained by converting elevation angle to distance. Simultaneously, the observed elevation angle at each sampling point is recorded. and the corresponding radial wind speed observations This matrix reflects the projected wind speed information of the atmosphere under different incident directions, providing sufficient data support for subsequent two-dimensional wind speed field inversion.

[0042] Step 3: Select local inversion sampling points Using the measurement center point as the core, five spatially evenly distributed sampling points with varying elevation angles are selected within its vicinity. The spatial distance between each sampling point and the center point does not exceed a preset maximum spatial scale parameter; that is, the distance between the five sampling points does not exceed 1.5m of the recording resolution, and they maintain equal intervals in spatial coordinates. This strategy ensures that the inversion matrix has a good condition number, thereby improving the numerical stability of the solution process. The principle of local inversion sampling point selection. Figure 11 As shown, when the inversion target is point P3, P1-P5 are selected as sampling points, and when the inversion target is point P4, P2-P6 are selected as sampling points.

[0043] Step 4: Establish a local linear wind field model Within the detection range based on the center point, assuming a horizontal wind speed... and vertical wind speed It satisfies a first-order linear change model: in: , Horizontal and vertical wind speeds at the center point; , , , This represents the first-order gradient of wind speed in the spatial direction; The location of the sampling point; This represents the location of the inversion center point.

[0044] Step 5: Construct the wind speed observation equation set Based on the observed elevation angle of each sampling point and spatial location that is, radial distance By combining the local linear wind field model, the wind field prediction formula for each sampling point is obtained: The equations for the five selected sampling points are then rearranged into a matrix form: in: Let the vector be the parameters to be determined. A is a coefficient matrix composed of the location and elevation angle information of each sampling point; b is the radial wind speed observation vector at the sampling point; This is radial wind speed data.

[0045] Step 6: Solve for local wind field parameters Solve the above system of equations using the least squares method or the weighted least squares method to obtain the wind speed parameters at the inversion center point: in: , The horizontal and vertical wind speeds at the center point obtained from the inversion; , , , This represents the local wind speed gradient, which can be used to describe the trend of local wind field changes.

[0046] Step 7: Output the inversion results The horizontal wind speed at each inversion center point and vertical wind speed As the output of the local wind field inversion, combined with local gradient information , , , By interpolating or extrapolating within a small surrounding area, a more complete local wind field distribution can be obtained.

[0047] Figure 1 To construct a realistic horizontal wind speed map; Figure 2 This is a horizontal wind speed map obtained based on local wind field inversion; Figure 3 This is a horizontal wind speed map retrieved using the traditional least squares method. (Comparison) Figure 2 and Figure 3 ,connect Figure 1 It can be seen that the method of the present invention is closer to the real horizontal wind speed in horizontal wind speed inversion.

[0048] Figure 4 and Figure 5 The figures show the RMSE error distributions of horizontal wind speeds based on local wind field inversion and traditional least squares inversion, respectively. It can be observed that the inversion error of the method described in this invention is significantly lower than that of the traditional least squares inversion method.

[0049] Similarly, Figures 6 to 10 The inversion of vertical wind speed and the comparison of errors are shown, further verifying the universality and high accuracy of the method of the present invention under complex wind field conditions.

[0050] In summary, this invention introduces a multi-point local sampling strategy at each inversion center point. By selecting several observation points spatially close to the center point and combining observation angle and radial wind speed information, a local multi-constraint inversion equation set is established. This overcomes the limitations of traditional wind field inversion methods that rely on the assumption of large-scale overall consistency. By adopting a local linear modeling method, it can better adapt to small-scale changes, local non-uniformity, and turbulent disturbances in actual wind fields, improving the adaptability and modeling accuracy of the inversion model to complex wind field structures. It also shows significant improvements in noise resistance, error control, and inversion convergence.

[0051] The present invention also includes: An apparatus comprising: Memory: Used to store the computer program for the multi-angle lidar vertical two-dimensional wind field inversion method.

[0052] Processor: Used to implement the multi-angle lidar vertical two-dimensional wind field inversion method when executing the computer program.

[0053] A computer-readable storage medium storing a computer program that, when executed by a processor, implements the aforementioned method for inverting a vertical two-dimensional wind field using a multi-angle lidar.

[0054] A system comprising hardware components and embedded software programs is capable of performing the entire process of a multi-angle lidar vertical two-dimensional wind field inversion method.

[0055] Simulation: Using an elevation angle range of 20° to 60°, angular resolution... Detection distance Spatial resolution from 30 meters to 150 meters A two-dimensional observation grid is constructed; the radial wind speed of the inversion region is obtained based on the schematic diagram of the laser system, and the specific operation is as follows: Figure 12 As shown in the schematic diagram of the laser system, the laser emits a narrow-bandwidth continuous laser beam. After passing through an optical fiber beam splitter, part of it serves as the local oscillator, while the other part is emitted into the atmosphere via an optical fiber collimator. The laser beam is backscattered by aerosol particles, carrying Doppler frequency shift information caused by the aerosol's motion. The scattered light is received by a telescope and undergoes beat frequency interference with the local oscillator in a photodetector, forming a beat frequency signal containing the Doppler frequency shift. This signal is converted into an electrical signal by a balanced detector, acquired by a high-speed data acquisition card (such as PXIe-6366), and transmitted to a host computer. A MATLAB program deployed on the host computer performs a Fourier transform on the signal, extracts the center frequency of the spectrum, and thus calculates the radial velocity of the target aerosol (refer to...). Figure 13 It is the radial wind speed of the actual wind field. Figure 14(This is the radial wind speed obtained by the present invention), enabling wind speed measurement at each sampling point in the wind field.

Claims

1. A method for multi-angle lidar vertical two-dimensional wind field retrieval, characterized in that, Specifically as follows: In the region of the wind field to be measured, the position of the measurement center point is determined, and the spatial scale parameters of the local inversion region are set, including the sampling radius, the elevation angle scanning range, the angle resolution, and the distance resolution; The region of the wind field to be measured is scanned by a multi-angle coherent Doppler laser radar scanning system to implement multi-elevation angle sector scanning, and radial wind speed data containing multiple scanning angles are obtained; Taking the center point to be measured as the core, based on the spatial scale parameters, a plurality of sampling points with uniform spatial distribution and different elevation angle coverage are selected within the adjacent spatial range thereof; A local linear wind field model is established, and a first-order linear change model satisfying horizontal wind speed and vertical wind speed in the local linear wind field model is constructed to construct a wind speed observation equation set; The least square method and the radial wind speed data are used to solve the wind speed observation equation set, and the wind speed parameters of the inversion center point are obtained; The inversion result is output, that is, the local wind field distribution based on the wind speed parameters of the inversion center point.

2. The multi-angle laser radar vertical two-dimensional wind field inversion method according to claim 1, characterized in that, The radial wind speed data at multiple scan angles are acquired, specifically, at each elevation angle The radar collects a series of measurement points along the radial distance to form a two-dimensional observation matrix The two-dimensional observation matrix is taken as the radial wind speed data.

3. The multi-angle lidar vertical two-dimensional wind field retrieval method according to claim 1, characterized in that, Taking the center point to be measured as the core, a plurality of sampling points with uniform spatial distribution and different elevation angle coverage are selected within the adjacent spatial range thereof, specifically including: Taking the center point to be measured as the core, the spatial distance between each sampling point and the center point to be measured is not more than a preset maximum spatial scale parameter, and the sampling points are selected, wherein the spatial distance between each sampling point is not more than a set spatial distance resolution, and the sampling points are kept at equal intervals in the elevation angle direction or the spatial coordinates.

4. The multi-angle lidar vertical two-dimensional wind field retrieval method according to claim 1, characterized in that, The local linear wind field model is established, specifically as follows: In a small range with the center point as a reference, the small range refers to a coverage of the center point and the selected multiple sampling points, and assuming that the horizontal wind speed and the vertical wind speed satisfy a first-order linear change model: Wherein: , are the horizontal and vertical wind speed, respectively, of the center point; , , , is the first order gradient of the wind speed in the spatial direction; is the position of the sampling point; is the center point position of the inversion.

5. The multi-angle lidar vertical two-dimensional wind field retrieval method according to claim 1, characterized in that, The wind speed observation equation set is constructed, specifically as follows: According to the observed elevation angle of each sampling point and spatial position , i.e. the radial distance is , combined with the local linear wind field model, the wind field prediction formula of each sampling point is obtained: The equations of the selected n sampling points are unified and arranged into a matrix form: Wherein: is the parameter vector to be found; A is a coefficient matrix composed of the positions and elevation angle information of the sampling points; b is a radial wind speed observation vector of the sampling points; is the radial wind speed data.

6. The multi-angle lidar vertical two-dimensional wind field retrieval method according to claim 1, characterized in that, The wind speed parameters of the inversion center point are obtained, specifically as follows: Wherein: , horizontal and vertical wind speed of the center point obtained by inversion; , , , is a local wind speed gradient, used to describe the trend of local wind field change.

7. The multi-angle lidar vertical two-dimensional wind field retrieval method according to claim 1, characterized in that, The output inversion result is specifically: the horizontal wind speed of each inversion center point and the vertical wind speed As the output result of local wind field inversion, combined with local gradient information 、 、 、 , interpolation or calculation is carried out in a small range around the periphery to obtain a more complete local wind field distribution.

8. A system, characterized by The hardware components and embedded software programs can complete the whole process operation of the multi-angle laser radar vertical two-dimensional wind field inversion method according to any one of claims 1 to 7.

9. An apparatus, comprising: Memory: for storing the computer program of the multi-angle laser radar vertical two-dimensional wind field inversion method according to any one of claims 1 to 7.

10. A computer-readable storage medium, characterized in that, Including: The computer readable storage medium stores a computer program, and the computer program is executed by the processor to realize the multi-angle laser radar vertical two-dimensional wind field inversion method according to any one of claims 1 to 7.

Images