Gust front identification method combining wind laser radar and Doppler weather radar

Abstract

The present application discloses a gust front identification method combining a wind laser radar and a Doppler weather radar, and relates to the field of meteorological technology. The gust front identification method establishes a gust front motion state model according to the gust front detection results obtained by two consecutive scans of the gust front to be identified outside the detection range of the wind laser radar by the Doppler weather radar, and then determines the gust front characteristic parameters when the gust front to be identified enters the detection range of the wind laser radar according to the gust front motion state model, and obtains the gust front detection result of the wind laser radar according to the gust front characteristic parameters and the radial velocity data observed in real time by the wind laser radar; the gust front detection results of the two radars are fused to obtain the final identification result. The method can give play to the near-field detection advantage of the wind laser radar while retaining the far-field detection advantage of the Doppler weather radar for the gust front, and has a good gust front identification effect and a low false alarm misjudgment.

Description

Technical Field

[0001] The present application relates to the field of meteorological technology, and in particular to a gust front identification method combining a wind laser radar and a Doppler weather radar. Background Art

[0002] The downdraft in a strong convective cell brings the high-altitude cold air to the ground to form a cold pool, and forms a divergent wind field near the ground that diffuses to the surroundings. Affected by the cell's moving direction, terrain, and weather background, this diffuse airflow forms a weak cold front at the front of the strong convection, which becomes a gust front. Gust fronts often occur near mature storm cells or squall lines, caused by thunderstorms. Since the near-ground outflow of the downdraft often has extremely high wind speeds, the gust front manifests itself locally as sudden strong gusts before the arrival of a thunderstorm, forming significant strong wind shear, which often causes serious meteorological disasters. When the gust front passes through this field, it will cause significant wind shear near this field, which poses a serious threat to the takeoff and landing of aircraft. Therefore, it is of great significance to accurately and effectively identify the gust front.

[0003] At present, Doppler weather radar is mainly used to detect gust fronts. The main methods include (1) identification based on the convergence lines in the radial wind field; (2) the gust front appears as one or more narrow weak echoes in the shape of arcs at the front of the strong convection in the echo image of the Doppler weather radar, so it can also be identified based on the narrow-band weak echoes in the echo image. Or a combination of the above two methods can be used for identification. However, the near-field detection of gust fronts by Doppler weather radar has long been a difficult problem. Doppler weather radar has poor recognition effect and serious deformation problems for near-field gust fronts, especially within 10 km. The main reasons include: (1) From the echo characteristics, the narrow-band weak echoes of the gust front will be submerged in the clear sky clutter in the near field, which objectively increases the risk of misjudgment. (2) From the perspective of velocity field characteristics, the radial velocity convergence characteristics of the gust front are only significant within the range of about ±45 degrees in azimuth angle, while the signal-to-noise ratio of Doppler weather radar is low and the wind field consistency is poor, making it more difficult to effectively obtain the gust front identification image in the near field. (3) The scanning cycle of Doppler weather radar is long (typically 6 minutes), making it difficult to perform refined detection and tracking of fast-moving gust fronts (about 1-2 km / min) and accurate approach warning.

[0004] At present, some practices use wind laser radar to detect gust fronts. Gust fronts show characteristics similar to fronts in the radial wind field, so they can be detected based on the identification method of convergence lines. Compared with Doppler weather radar, wind laser radar has a high signal-to-noise ratio and better velocity field consistency, so it can perform near-field identification of gust fronts in a larger azimuth angle range. On the other hand, wind laser radar has higher spatiotemporal resolution, so it can provide more accurate approaching warnings for gust fronts. However, the disadvantages of wind laser radar are: (1) Due to the limited detection range, it is difficult to monitor gust fronts at a long distance; (2) Also due to the limited detection range and detection methods, it is difficult to detect and confirm strong convective cells related to gust fronts, thereby increasing the risk of false alarms and misjudgments; (3) The extrapolation algorithm requires historical records of front observations of more than two adjacent times. Therefore, the gust front needs to be several kilometers deep into the field of view of the wind laser radar to complete the accumulation of historical records and achieve a good identification effect, which greatly limits the timeliness of approaching warnings. Summary of the invention

[0005] In response to the above problems and technical requirements, this application proposes a gust front identification method combining wind laser radar and Doppler weather radar. The technical solution of this application is as follows:

[0006] A gust front identification method combining a wind laser radar and a Doppler weather radar, the gust front identification method comprising:

[0007] The gust front detection result GFW is obtained according to the radial velocity data observed in real time by the Doppler weather radar, and the gust front motion state model is established according to the gust front detection result GFW obtained by two consecutive scans. The gust front to be identified corresponding to the gust front detection result GFW is outside the detection range of the wind laser radar.

[0008] According to the gust front motion state model, the characteristic parameters of the gust front when the gust front to be identified enters the detection range of the wind laser radar are determined;

[0009] According to the characteristic parameters of the gust front and the radial velocity data observed in real time by the wind laser radar, the gust front detection result GFL of the wind laser radar to be identified is obtained;

[0010] The gust front detection result GFL of the gust front to be identified by the wind laser radar is fused with the gust front detection result GFW of the gust front to be identified by the Doppler weather radar to obtain the identification result of the gust front to be identified.

[0011] A further technical solution is that the characteristic parameters of the gust front include the azimuth and time when the gust front to be identified enters the detection range of the wind measuring laser radar.

[0012] A further technical solution is to obtain the gust front detection result GFL of the wind laser radar for the gust front to be identified, including:

[0013] The convergence zone is identified based on the radial velocity data observed in real time by the wind laser radar, and the convergence zone at the time and azimuth indicated by the characteristic parameters of the gust front is used as the gust front detection result GFL of the gust front to be identified by the wind laser radar.

[0014] A further technical solution is to fuse the gust front detection result GFL with the gust front detection result GFW, including:

[0015] The observation times of the two gust front detection results are fused, the gust front intensities detected by the two gust front detection results are fused, and the gust front shapes detected by the two gust front detection results are fused.

[0016] A further technical solution is to fuse the observation times of the two gust front detection results, and to fuse the gust front intensities detected by the two gust front detection results, including:

[0017] When the observation times of two gust front detection results are different, the two gust front detection results are recorded as the gust front detection results of the gust front to be identified at two different observation times;

[0018] The larger value of the gust front intensity in the two gust front detection results is recorded as the gust front intensity of the gust front to be identified, and the smaller gust front intensity in the other gust front detection result is updated to the larger value.

[0019] A further technical solution is to fuse the gust front forms detected by the two gust front detection results, including:

[0020] When the gust front detection result GFL is fitted to obtain the front line curve The length of the front curve obtained by fitting the gust front detection result GFW When the length is small and the difference reaches the difference threshold, the front curve Pan to the front curve After that, the front line curve Discrete points and front curves in The discrete points in the are merged and refitted to obtain the updated front line curve .

[0021] A further technical solution is to establish a gust front motion state model based on the gust front detection result GFW, including:

[0022] Determine the Doppler weather radar The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , and in The gust front detection result obtained by the last scan is completed at the moment. The front line curve obtained by GFW fitting ;

[0023] Calculate the forward curve To the front curve The moving vector of Determine the moving speed of the gust front to be identified , the direction of the moving vector is taken as the moving direction of the gust front to be identified.

[0024] A further technical solution is to determine the characteristic parameters of the gust front when the gust front to be identified enters the detection range of the wind laser radar according to the gust front motion state model, including:

[0025] According to Doppler weather radar The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , extrapolated to get any future The forward curve at the moment With the front curve The curve shape is the same as that of the front line curve The moving distance along the moving direction of the gust front to be identified is ;

[0026] When it is confirmed that It is determined that the gust front to be identified will enter the detection range of the wind laser radar and will appear for the first time. of The time is taken as the time when the gust front to be identified enters the detection range of the wind laser radar, and the front line curve at this time is determined Azimuth relative to the wind lidar; where, is the position of the wind lidar, is the detection radius of the wind lidar.

[0027] A further technical solution is that the gust front identification method further includes:

[0028] After each scan, the Doppler weather radar uses the gust front detection results obtained from the two most recent scans of the Doppler weather radar to update the gust front motion state model, and uses the updated gust front motion state model to update the front curve at any time in the future to achieve early warning of the gust front.

[0029] A further technical solution is that the gust front identification method further includes:

[0030] After the gust front to be identified enters the detection range of the wind measuring lidar, when either the Doppler weather radar or the wind measuring lidar completes a scan, the gust front detection results obtained from the two most recent radar scans are used to update the gust front motion state model, and the updated gust front motion state model is used to update the front line curve at any time in the future, so as to realize the early warning of the approaching gust front.

[0031] The beneficial technical effects of this application are:

[0032] The present application discloses a gust front identification method combining a wind laser radar and a Doppler weather radar. The method uses the wind laser radar and the Doppler weather radar to jointly observe and warn of gust fronts. While giving full play to the advantage of good wind field consistency of the wind laser radar and greatly improving the 0-10km near-field detection effect of the Doppler weather radar, the method effectively uses the advance detection information of the Doppler weather radar, greatly improves the gust front detection and warning efficiency of the wind laser radar, and can give full play to the near-field detection advantage of the wind laser radar while retaining the far-field detection advantage of the Doppler weather radar for gust fronts, thereby having a good gust front identification effect and a low false alarm misjudgment.

[0033] Traditional methods require wind laser radar and the accumulation of more than two relatively complete scans of gust fronts before extrapolated warnings can be carried out. However, the method of this application uses Doppler weather radar to provide early warning information, so that extrapolated warnings can be established in the first scanning cycle when the gust front enters the detection range of the wind laser radar, which significantly improves the timeliness of warnings and ultimately achieves high-precision and refined detection of gust fronts and approaching warnings. This application is used for early warning protection of gust front-related strong wind disasters in high-impact areas such as airports and ports, and has good industrial application prospects. BRIEF DESCRIPTION OF THE DRAWINGS

[0034] Figure 1 This is a flow chart of a gust front identification method according to an embodiment of the present application. DETAILED DESCRIPTION

[0035] The specific implementation of the present application is further described below in conjunction with the accompanying drawings.

[0036] This application discloses a gust front identification method combining wind laser radar and Doppler weather radar, please refer to Figure 1 As shown in the flowchart, the gust front identification method includes:

[0037] Step 1: Obtain the gust front detection result GFW according to the radial velocity data observed in real time by the Doppler weather radar.

[0038] In the application scenario of the method of this application, Doppler weather radar and wind laser radar are used to monitor the surrounding gust fronts. The detection range of Doppler weather radar is greater than that of wind laser radar. The main body of the gust front is below 1km in height. Affected by the curvature of the earth, the theoretical maximum detection distance of Doppler weather radar for gust fronts is about 150km. In fact, since the gust front is often lower than 500m and the echo is very weak, Doppler weather radar can often detect the gust front better within 60km. Therefore, this application considers the typical detection radius of Doppler weather radar to be 75km and the conservative detection radius to be 45km. The detection radius of wind laser radar is typically around 10km.

[0039] Since the detection range of the Doppler weather radar is larger than that of the wind laser radar, the gust front to be identified corresponding to the gust front detection result GFW initially obtained by the Doppler weather radar is outside the detection range of the wind laser radar and cannot be detected by the wind laser radar temporarily.

[0040] Based on the strong convective cell identification algorithm and the radial wind field convergence zone detection algorithm (GFDA algorithm), the identification and detection of gust fronts can be realized according to the radial velocity data observed in real time by the Doppler weather radar: through the radial wind field convergence zone detection algorithm, the convergence line can be identified based on the radial velocity data observed in real time by the Doppler weather radar, and then it is determined whether there is a strong convective cell within a predetermined range on one side of the strong wind area of ​​the identified convergence line (that is, behind the front line). When it is determined that there is a strong convective cell in the direction of the convergence line, it is determined that the identified convergence line is a gust front caused by strong convective weather, that is, the gust front to be identified in this application is detected. When it is determined that there is no strong convective cell in the direction of the convergence line, the convergence line continues to be identified. When the convergence lines identified for multiple consecutive times are stable and continue to move at a predetermined speed, it is determined that the identified convergence line is a gust front caused by far-field strong convective weather. In addition, if it is determined that the identified convergence line does not belong to the gust front, detection continues. The strong convective cell identification algorithm and the radial wind field convergence zone detection algorithm are both existing mature algorithms, and will not be described in detail in this embodiment.

[0041] Step 2: Establish a gust front motion state model based on the gust front detection results GFW obtained from two consecutive scans.

[0042] Determine the Doppler weather radar The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , and in The gust front detection result obtained by the last scan is completed at the moment. The front line curve obtained by GFW fitting .

[0043] Calculate the forward curve To the front curve The moving vector of Determine the moving speed of the gust front to be identified , the direction of the moving vector is used as the moving direction of the gust front to be identified. Each front curve can be regarded as a discrete point set, so in one embodiment, the front curve is calculated The corresponding discrete point set and front curve The median Hausdorff distance vector of the corresponding discrete point set is used as the front curve To the front curve The moving vector of .

[0044] The wind field characteristics (dominant wind speed, wind direction) in the strong wind area of ​​the gust front can be obtained using the GFDA algorithm, which will not be repeated here.

[0045] Step 3: Determine the gust front characteristic parameters when the gust front to be identified enters the detection range of the wind measuring laser radar according to the gust front motion state model.

[0046] The characteristic parameters of the gust front include the azimuth and time when the gust front to be identified enters the detection range of the wind laser radar. Optionally, the gust front intensity when the gust front to be identified enters the detection range of the wind laser radar is also included.

[0047] As in step 2 above, the gust front motion state model is obtained based on the results of the last two consecutive scans, thereby determining the moving direction and speed of the gust front to be identified. Then, the front curve can be extrapolated according to the moving direction and speed. The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , extrapolated to get any future The forward curve at the moment With the front curve The curve shape is the same as that of the front line curve The moving distance along the moving direction of the gust front to be identified is .

[0048] Check if there is a moment Can satisfy ,in, is the position of the wind lidar, is the detection radius of the wind lidar.

[0049] If there is a moment If the above inequality holds, it means that the gust front to be identified will move into the detection range of the wind laser radar, and the first of The moment is taken as the moment when the gust front to be identified enters the detection range of the wind laser radar, and the front curve at this time is determined Relative to the azimuth of the wind laser radar, the characteristic parameters of the gust front to be identified are obtained when the gust front enters the detection range of the wind laser radar.

[0050] Step 4: According to the gust front characteristic parameters and the radial velocity data observed in real time by the wind laser radar, the gust front detection result GFL of the gust front to be identified by the wind laser radar is obtained.

[0051] The radial wind field convergence zone detection algorithm can be used to identify the convergence zone based on the radial velocity data observed in real time by the wind laser radar. The convergence zone at the time and azimuth indicated by the characteristic parameters of the gust front is directly identified as the gust front detection result GFL for the gust front to be identified, and there is no requirement for the length of the convergence zone. For other convergence zones, it is required that the length of the convergence zone reaches the gust front length threshold before it is identified as a gust front.

[0052] Step 5: The gust front detection result GFL of the gust front to be identified by the wind laser radar is merged with the gust front detection result GFW of the gust front to be identified by the Doppler weather radar to obtain the identification result of the gust front to be identified.

[0053] What is fused here is the most recent gust front detection result GFL of the wind laser radar and the most recent gust front detection result GFW of the Doppler weather radar, including: the fusion of the observation time of the two gust front detection results, the fusion of the gust front intensity detected by the two gust front detection results, and the fusion of the gust front shape detected by the two gust front detection results. They are introduced as follows:

[0054] (1) Fusion of observation times. Since the scanning cycles of Doppler weather radar and wind laser radar are different, the gust front detection results GFL and gust front detection results GFW may have different observation times. When the observation times of two gust front detection results are different, the two gust front detection results are recorded as the gust front detection results of the gust front to be identified at two different observation times.

[0055] (2) Fusion of gust front intensity: The maximum outflow intensity of the gust front is usually located near the ground, and the wind speed decreases rapidly with increasing altitude. Due to the different detection elevation angles of the Doppler weather radar and the wind measurement lidar and the influence of clutter, the detected gust front intensities are usually different. The larger value of the gust front intensity in the two gust front detection results is recorded as the gust front intensity of the gust front to be identified. In addition, considering that a weak GFW may need to match the two GFLs before and after, the smaller gust front intensity in the two gust front detection results is also updated to the larger value.

[0056] (3) Fusion of gust front shape: Since the detection ranges of Doppler weather radar and wind laser radar are different, the length of the gust front is greatly affected by the detection range, so the detected front lengths usually differ. The length of the front curve obtained by fitting the gust front detection result GFW When the length is small and the difference reaches the difference threshold, the front curve Pan to the front curve After that, the front line curve Discrete points and front curves in The discrete points in the are merged and refitted to obtain the updated front line curve On the contrary, if the lengths of the two front curves are similar, the shape of the gust front detection result GFL will not be corrected.

[0057] In addition, the gust front motion state model used in the above steps 2 and 3 is not fixed, but is continuously and dynamically updated. After each scan, the Doppler weather radar will use the gust front detection results obtained by the Doppler weather radar in the last two scans to update the gust front motion state model, and use the updated gust front motion state model to update the front curve at any time in the future according to the method of step 3, so that the trend of the front curve can be dynamically updated to achieve gust front early warning. Estimated at the typical moving speed of gust front of 1.5km / min, the Doppler weather radar can correspond to a warning lead time of about 30 minutes in the case of a detection and warning range of 45km, achieving a gust front early warning with a time resolution of 6 minutes and a spatial resolution of 150m.

[0058] Not only before the gust front to be identified enters the detection range of the wind laser radar, the Doppler weather radar will continuously update the gust front motion state model, but also after the gust front to be identified enters the detection range of the wind laser radar, the Doppler weather radar will continuously update the gust front motion state model according to the above method. And after the gust front to be identified enters the detection range of the wind laser radar, the wind laser radar will also detect the gust front to be identified. After the wind laser radar completes a scan, it will also use the gust front detection results obtained by the wind laser radar in the last two scans to update the gust front motion state model. Then, according to the method similar to the above step 3, the front line curve at the future moment is fitted according to the gust front detection result GFL obtained by the wind laser radar after completing the current scan, so as to realize the gust front approach warning. Estimated with the typical moving speed of 1.5km / min of the gust front, under the detection radius of 10km of the wind laser radar, the gust front approach warning with a time resolution of 2 minutes and a spatial resolution of 60m can be realized, corresponding to a warning lead time of about 6 minutes.

[0059] The above is only a preferred embodiment of the present application, and the present application is not limited to the above embodiments. It is understood that other improvements and changes directly derived or associated by those skilled in the art without departing from the spirit and concept of the present application should be considered to be included in the protection scope of the present application.

Claims

1. A gust front identification method combining wind laser radar and Doppler weather radar, characterized in that: The gust front identification method comprises: The gust front detection result GFW is obtained according to the radial velocity data observed by the Doppler weather radar in real time, and the gust front motion state model is established according to the gust front detection result GFW obtained by two consecutive scans. The gust front to be identified corresponding to the gust front detection result GFW is outside the detection range of the wind laser radar. The gust front motion state model is established according to the gust front detection result GFW, including: determining the Doppler weather radar in The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , and in The gust front detection result obtained by the last scan is completed at the moment. The front line curve obtained by GFW fitting , calculate the front curve To the front curve The moving vector of Determine the moving speed of the gust front to be identified , taking the direction of the moving vector as the moving direction of the gust front to be identified; Determining the gust front characteristic parameters when the gust front to be identified enters the detection range of the wind laser radar according to the gust front motion state model includes: The gust front detection result obtained by the current scan is completed at the moment. The front line curve obtained by GFW fitting , extrapolated to get any future The forward curve at the moment With the front curve The curve shape is the same as that of the front line curve The moving distance along the moving direction of the gust front to be identified is ; When it is determined that there is It is determined that the gust front to be identified will enter the detection range of the wind laser radar and will appear for the first time. of The time is taken as the time when the gust front to be identified enters the detection range of the wind laser radar, and the front line curve at this time is determined Azimuth relative to the wind lidar; where, is the position of the wind lidar, is the detection radius of the wind lidar; According to the gust front characteristic parameters and the radial velocity data observed in real time by the wind laser radar, a gust front detection result GFL of the wind laser radar for the gust front to be identified is obtained; The gust front detection result GFL of the gust front to be identified by the wind laser radar is fused with the gust front detection result GFW of the gust front to be identified by the Doppler weather radar to obtain the identification result of the gust front to be identified; The fusion of the gust front detection result GFL and the gust front detection result GFW includes: when the observation times of the two gust front detection results are different, the two gust front detection results are recorded as the gust front detection results of the gust front to be identified at two different observation times; the larger value of the gust front intensity in the two gust front detection results is recorded as the gust front intensity of the gust front to be identified, and the smaller gust front intensity in the other gust front detection result is updated to the larger value; when the front line curve obtained by fitting the gust front detection result GFL The length of the front curve obtained by fitting the gust front detection result GFW When the length is small and the difference reaches the difference threshold, the front curve Pan to the front curve After that, the front line curve Discrete points and front curves in The discrete points in the merged and refitted to get the updated front curve , thereby fusing the gust front shapes detected by the two gust front detection results.

2. The gust front identification method according to claim 1, characterized in that: The gust front characteristic parameters include the azimuth and time when the gust front to be identified enters the detection range of the wind measuring laser radar.

3. The gust front identification method according to claim 2, characterized in that: Obtaining the gust front detection result GFL of the wind laser radar for the gust front to be identified includes: The convergence zone is identified according to the radial velocity data observed in real time by the wind laser radar, and the convergence zone at the time and azimuth indicated by the characteristic parameters of the gust front is used as the gust front detection result GFL of the wind laser radar for the gust front to be identified.

4. The gust front identification method according to claim 1, characterized in that: The gust front identification method further includes: After the Doppler weather radar completes each scan, the gust front detection results obtained by the Doppler weather radar in the two most recent scans are used to update the gust front motion state model, and the updated gust front motion state model is used to update the front curve at any time in the future to achieve early warning of the gust front.

5. The gust front identification method according to claim 1, characterized in that: The gust front identification method further includes: After the gust front to be identified enters the detection range of the wind measuring laser radar, when either the Doppler weather radar or the wind measuring laser radar completes a scan, the gust front detection results obtained by the two most recent scans of the radar are used to update the gust front motion state model, and the updated gust front motion state model is used to update the front line curve at any time in the future, so as to realize the early warning of the approaching gust front.

Images