Flood control channel flow monitoring method fused with flow velocity and area method

Abstract

The invention discloses a flood control channel flow monitoring method fused with a flow velocity area method, and belongs to the technical field of water conservancy flow monitoring. The method comprises the following steps: acquiring a water-passing section contour at a monitored section through a three-dimensional scanning technology, dispersing the water-passing section contour into n small units by adopting Delaunay triangulation, and recording areas and positions; acquiring distribution flow velocity data at the monitoring section through the two-dimensional flow velocity measurement matrix, and distributing the distribution flow velocity data to each small unit through an inverse distance weighted interpolation method; aggregating the small units into large units, calculating the total area and the average flow velocity of the large units, evaluating the flow velocity deviation by taking the area weighted flow velocity of all the small units as the reference flow velocity, identifying vortex regions based on the flow velocity deviation and counting the number of the vortex regions, and dynamically adjusting the flow velocity distribution coefficient according to the number of the vortex regions. And calculating the total flow of the flood control channel through partition integration. The accuracy and reliability of flow monitoring in the complex flow state are improved, and the actual flow state requirement of the flood control channel is met.

Description

Technical Field

[0001] This invention relates to the field of water conservancy flow monitoring technology, specifically to a method for monitoring flood control channel flow that integrates the velocity-area method. Background Technology

[0002] Flood control channels are critical infrastructure in flood control and disaster reduction systems. Real-time and accurate monitoring of their flow is essential for flood forecasting, water conservancy project scheduling, and the safety of life and property along the riverbanks. The velocity-area method, due to its intuitive principle and strong feasibility, has become the mainstream technology for flood control channel flow monitoring. Its core principle is based on calculating the total flow rate using the formula "flow rate = velocity × cross-sectional area". However, traditional methods often rely on water level sensors to obtain single water level data, combining this data with pre-defined rectangular, trapezoidal, or other regular cross-sectional shapes to estimate the cross-sectional area. During flood season, the actual cross-sectional shape of flood control channels is often irregular, and water levels vary at different locations, leading to significant discrepancies between the calculated and actual cross-sectional area values. This directly affects the basic accuracy of flow monitoring.

[0003] Existing technologies have several shortcomings: sparsely deployed velocity sensors and simple interpolation methods cannot fully capture the velocity differences between the center and boundaries of a cross-section, especially for complex flow regimes such as vortices formed during floods; traditional fixed velocity distribution coefficients cannot dynamically adapt to changes in vortex regions, easily leading to velocity estimation errors; cross-section discretization often uses equal-area division without adjusting unit density based on cross-sectional curvature and velocity characteristics, resulting in the loss of crucial flow regime information. These problems make existing methods insufficient to meet the high-precision and high-reliability requirements of flood control scheduling for flow monitoring. Therefore, developing a flow monitoring method capable of accurately acquiring irregular cross-sections, comprehensively characterizing cross-sectional velocity distribution, and dynamically adapting to complex flow regimes has become an urgent technical problem to be solved in the field of water conservancy flow monitoring. Summary of the Invention

[0004] The purpose of this invention is to provide a flood control channel flow monitoring method that integrates the velocity-area method, so as to solve the problems mentioned in the background art.

[0005] To achieve the above objectives, the present invention provides the following technical solution:

[0006] A method for monitoring flood control channel flow that integrates the velocity-area method includes the following steps:

[0007] S1. The outline of the water-passing section at the monitoring section of the flood control channel is obtained in real time through three-dimensional scanning technology, and the outline of the water-passing section is discretized into n small units, and the corresponding area and position of each small unit are recorded.

[0008] S2. Obtain the distributed velocity data at the monitoring section of the flood control channel, and distribute the distributed velocity data to each small unit using the inverse distance weighted interpolation method to obtain the velocity of each small unit;

[0009] S3. Set a flow rate tolerance threshold. Based on the flow rate of the small units, aggregate n small units into m large units and calculate the total area and average flow rate of each large unit. At the same time, use the area-weighted flow rate of all small units as the benchmark flow rate to evaluate the flow rate deviation of each large unit.

[0010] S4. Identify vortex zones based on the velocity deviation of each large unit and count the number of vortex zones. Dynamically adjust the velocity distribution coefficient according to the number of vortex zones. Based on this, combine the average velocity and total area of ​​each large unit to calculate the total flow of the flood control channel through zonal integration.

[0011] Preferably, the method for obtaining the water-passing cross-section contour is as follows: 2D lidar is symmetrically deployed along both banks of the flood control channel monitoring section to collect three-dimensional point cloud data of the water surface at the flood control channel monitoring section in real time. The three-dimensional point cloud data of the water surface is then unified into a three-dimensional rectangular coordinate system with the fixed reference point of the flood control channel as the origin, the X-axis parallel to the water flow direction as the Y-axis perpendicular to the water flow direction as the Z-axis, and the vertical direction as the Z-axis. The point cloud is then aligned with the pre-stored flood control channel contour point cloud during the dry season using the ICP registration algorithm to obtain a fused point cloud. The fixed reference point of the flood control channel is three stainless steel marker points pre-buried on both banks of the flood control channel monitoring section.

[0012] An instantaneous cross-section is extracted from the fused point cloud along the direction perpendicular to the water flow in the flood control channel. The water surface point set and the channel boundary point set in the instantaneous cross-section are extracted. The alpha shape algorithm is used to fit the water surface point set and the channel boundary point set to generate a two-dimensional water cross-section profile. The alpha value of the alpha shape algorithm is set to a range of 0.3-0.5. The specific alpha value is adjusted according to the geometric complexity of the flood control channel monitoring cross-section. If the flood control channel monitoring cross-section has a regular shape, the alpha value is 0.4-0.5; if the flood control channel monitoring cross-section has irregular protrusions or pits, the alpha value is 0.3-0.4.

[0013] Preferably, the discretization of the water-passing cross-sectional profile is achieved by using the water-passing cross-sectional profile as the profile boundary and employing Delaunay triangulation to divide the region within the profile boundary into n small units, where the number of small units n is 50-300. This is dynamically adjusted according to the cross-sectional curvature of the water-passing cross-sectional profile. Based on the water-passing cross-sectional profile, the cross-sectional curvature of each segment of the profile is calculated using the coordinate difference method of adjacent three points, with a curvature threshold set to 0.1. If the cross-sectional curvature of a certain contour segment is >0.1 Therefore, the area of ​​the small units within the inner region of the cross-sectional profile corresponding to that segment should be controlled to be 0.1-0.3. If the cross-sectional curvature of a certain profile is ≤0.1 Then, the area of ​​the small unit within the inner region of the cross-sectional profile corresponding to that segment is set to 0.3-0.5. Based on this, the unit density of the inner region of the water cross-section contour is dynamically adjusted by radiating the cross-sectional curvature to the inner region of the water cross-section contour. The method of differing coordinates of adjacent three points is as follows: ; in, , and These are the coordinates of three consecutive contour points on the water cross-section profile. The curvature of the cross section.

[0014] Preferably, the method for obtaining the distributed flow velocity data is as follows: several equally spaced vertical lines are arranged horizontally on the cross section of the flood control channel monitoring section. The vertical lines are made of stainless steel with an impact resistance strength ≥300MPa and are reinforced by anchor seats. Layered flow velocity sensors are set at equal intervals along the vertical direction on each vertical line to form a two-dimensional flow velocity measurement matrix. The layered flow velocity sensors are any one of propeller-type flow velocity sensors, electromagnetic flow velocity sensors, and ultrasonic flow velocity sensors.

[0015] Distributed flow velocity data is obtained by synchronously sampling all layered flow velocity sensors in the two-dimensional flow velocity measurement matrix via LoRa wireless. ,in, y represents the lateral position of the stratified flow velocity sensor, and y represents the vertical depth of the stratified flow velocity sensor. The velocity at the stratified velocity sensor is used to intuitively characterize the two-dimensional velocity distribution across the water cross section.

[0016] Preferably, the method for obtaining the flow rate of the small unit is as follows: for each small unit All distances to this small unit The straight-line distance is less than the preset distance threshold. The stratified flow velocity sensor serves as the stratified flow velocity sensor collection for this small unit. , The number of stratified flow velocity sensors in the stratified flow velocity sensor set; the preset distance threshold is set according to the deployment density of the stratified flow velocity sensors; the straight-line distance is the number of small units. The straight-line distance between the corresponding position and the coordinate position of the stratified flow velocity sensor;

[0017] The spatial straight-line distance between the position of this small unit and each stratified flow velocity sensor in the stratified flow velocity sensor set is calculated using the spatial straight-line distance formula. The inverse distance of the linear distance in space is used as the weight and substituted into the inverse distance weighted interpolation formula to calculate the value of the small unit. The flow rate;

[0018] The formula for the straight-line distance in space is:

[0019] ;

[0020] in, Small unit Location, For stratified flow velocity sensor The coordinates of the location;

[0021] The inverse distance weighted interpolation formula is:

[0022] ;

[0023] in, Small unit The flow rate, Small unit Location and Layered Flow Velocity Sensor Set spatial straight-line distance reciprocal ;

[0024] If a small unit does not have a hierarchical flow velocity sensor set, the flow velocity of that small unit is supplemented by linear interpolation based on the flow velocities of the three small units adjacent to the shared edges or vertices of that small unit.

[0025] Preferably, the aggregation uses the DBSCAN density clustering algorithm, with a flow velocity tolerance threshold as the cluster radius. It traverses n small units, grouping adjacent small units whose flow velocity difference is less than or equal to the flow velocity tolerance threshold into one class, aggregating them into m large units, denoted as mm. And m≤n;

[0026] The large unit Total area The sum of the areas of the smaller units, and the larger unit average flow velocity It is the area-weighted average of the flow velocities of the constituent sub-units;

[0027] The formula for calculating the average flow velocity is:

[0028] ;

[0029] in, For large units All small units flow rate Perform area-weighted calculation. Small unit The area.

[0030] Preferably, the velocity deviation is calculated by substituting the velocity and area of ​​all small units into the area-weighted average velocity formula to calculate the area-weighted velocity. That is, the reference flow rate Based on this, the reference flow rate is determined. and the average flow velocity of large units Substitute into the velocity deviation formula to calculate the velocity deviation of the large unit. ;

[0031] The formula for the area-weighted average flow velocity is: ;

[0032] The formula for the flow velocity deviation is: .

[0033] Preferably, the method for identifying the vortex region is as follows: setting a threshold for abnormal flow velocity deviation. traverse all large units If the flow velocity deviation of a large unit is greater than the flow velocity abnormality deviation threshold, then the large unit is marked as a flow velocity abnormality unit.

[0034] If there are three or more spatially adjacent large units that are also flow velocity anomalies, and the area ratio between the area of ​​the smallest circumscribed circle of the clustered region calculated by the minimum enclosing circle algorithm and the area of ​​the clustered region is greater than 0.3, then the clustered region is determined to be a vortex region. The clustered region is a continuous region jointly formed by the flow velocity anomaly unit and all flow velocity anomalies spatially adjacent to the flow velocity anomaly unit.

[0035] Preferably, the velocity distribution coefficient is obtained by inputting the vortex feature vector into a pre-trained gradient boosting regression tree. The vortex feature vector includes the number of vortex regions, vortex region area features, and channel flow characteristics. The vortex region area features include the vortex region area ratio and the largest vortex region area ratio. The channel flow characteristics include the width-to-depth ratio of the flood control channel and the cross-sectional Reynolds number. The vortex region area ratio is the ratio between the sum of the areas of all vortex regions and the area of ​​the cross-section. The largest vortex region area ratio is the ratio between the vortex region with the largest vortex region area and the area of ​​the cross-section. The width-to-depth ratio of the flood control channel is the ratio between the water surface width and the water depth of the cross-section. The area, water surface width, and water depth of the cross-section are obtained from the cross-sectional profile.

[0036] The formula for calculating the Reynolds number of the cross section is: ;

[0037] in, The cross-sectional Reynolds number is... The water depth at the cross-section of the water passage. Let be the kinematic viscosity coefficient of water.

[0038] Preferably, the formula for calculating the total flow of the flood control channel is:

[0039] ;

[0040] in, The total flow rate of the flood control channel, The velocity distribution coefficient is... The total area of ​​the large unit. The average flow velocity of the large unit.

[0041] Compared with the prior art, the present invention has the following beneficial effects:

[0042] 1. This invention acquires three-dimensional point clouds of the water surface by symmetrically deploying 2D lidar, aligns them with pre-stored point clouds during the dry season using the ICP registration algorithm, and then fits the profile of the water-passing cross section using the alpha shape algorithm with adaptive alpha value. At the same time, it dynamically adjusts the density of small units based on the cross section curvature, thereby achieving accurate restoration and discretization of irregular water-passing cross sections and improving the accuracy of cross section morphology acquisition.

[0043] 2. This invention uses a high-strength stainless steel vertical line to construct a two-dimensional flow velocity measurement matrix. It acquires flow velocity data including lateral position and vertical depth through LoRa wireless synchronous sampling, and then uses inverse distance weighted interpolation combined with linear supplementation of adjacent units to comprehensively capture the cross-sectional flow velocity distribution and avoid information blind spots in sparse sampling.

[0044] 3. This invention adapts the velocity anomaly deviation threshold according to the cross-sectional width-to-depth ratio, marks the velocity anomaly units, and determines the vortex zone by combining the number of adjacent anomaly units with the ratio of the minimum circumscribed circle area of ​​the aggregation area. This achieves quantitative identification of complex vortex flow patterns during flood season and solves the problem that traditional methods are difficult to characterize complex flow patterns.

[0045] 4. This invention inputs feature vectors containing the number of vortices, cross-sectional Reynolds number, etc., into a pre-trained gradient boosting regression tree to obtain dynamic velocity distribution coefficients. Then, it combines the product of the area of ​​the large cell and the average velocity to calculate the total flow rate, avoiding the defect of poor adaptability of fixed coefficients and improving the accuracy of flow rate monitoring under complex flow conditions. Attached Figure Description

[0046] To more clearly illustrate the technical solutions and advantages in the embodiments of the present invention or the prior art, the drawings used in the description of the embodiments or the prior art will be briefly introduced below. Obviously, the drawings described below are only some embodiments of the present invention. For those skilled in the art, other drawings can be obtained based on these drawings without creative effort.

[0047] Figure 1 This is a flowchart of the method steps of the present invention. Detailed Implementation

[0048] To make the objectives, technical solutions, and advantages of this invention clearer, the technical solutions of this invention will be described in detail below. Obviously, the described embodiments are merely some embodiments of this invention, and not all embodiments. Based on the embodiments of this invention, all other implementation methods obtained by those skilled in the art without creative effort are within the scope of protection of this invention.

[0049] Examples, such as Figure 1 As shown, a flood control channel flow monitoring method integrating velocity-area method includes the following steps:

[0050] S1. The outline of the water-passing section at the monitoring section of the flood control channel is obtained in real time through three-dimensional scanning technology, and the outline of the water-passing section is discretized into n small units, and the corresponding area and position of each small unit are recorded.

[0051] S2. Obtain the distributed velocity data at the monitoring section of the flood control channel, and distribute the distributed velocity data to each small unit using the inverse distance weighted interpolation method to obtain the velocity of each small unit;

[0052] S3. Set a flow rate tolerance threshold. Based on the flow rate of the small units, aggregate n small units into m large units and calculate the total area and average flow rate of each large unit. At the same time, use the area-weighted flow rate of all small units as the benchmark flow rate to evaluate the flow rate deviation of each large unit.

[0053] S4. Identify vortex zones based on the velocity deviation of each large unit and count the number of vortex zones. Dynamically adjust the velocity distribution coefficient according to the number of vortex zones. Based on this, combine the average velocity and total area of ​​each large unit to calculate the total flow of the flood control channel through zonal integration.

[0054] Furthermore, the working principle of the present invention will be illustrated below through embodiments:

[0055] This embodiment uses a flood control canal in a plain area of ​​North my country as the monitoring object. The design flow rate of the flood control canal is 45. -50 At the monitoring section, the channel bottom width is 10m and the bank slope is 1:2.5. During the flood season, local vortex flow often forms due to water turbulence.

[0056] Two 2D lidar units with a scanning frequency of 15Hz and a ranging accuracy of ±2mm were symmetrically deployed on both banks of the flood control channel monitoring section to ensure that the scanning range covered the water surface width of approximately 12m and the bank slope area during the flood season. The fixed reference points for the flood control channel were three stainless steel markers buried in the monitoring section, whose coordinates were calibrated using GNSS static measurement. The 3D point cloud data of the water surface collected by the 2D lidar was unified into a 3D rectangular coordinate system with the reference point as the origin, the X-axis parallel to the water flow direction as the X-axis, the Y-axis perpendicular to the water flow direction as the Y-axis, and the vertical direction as the Z-axis. The point cloud was then aligned with the pre-stored point cloud of the flood control channel outline during the dry season using an ICP registration algorithm to obtain a fused point cloud. The water flow direction was calibrated using a water flow direction sensor with a measurement accuracy of ±1°. Instantaneous cross-sections of the merged point cloud were captured at a frequency of 1 minute per segment along the vertical direction of water flow. Based on the Z-axis elevation, the water surface point set and the channel boundary point set were separated from the instantaneous cross-section. Due to the presence of local depressions in the instantaneous cross-section, an alpha value of 0.35 was adopted. The alphashape algorithm was used to fit the water surface point set and the channel boundary point set to generate a closed cross-sectional profile. Using this cross-sectional profile as the boundary, Delaunay triangulation was used for discretization. The cross-sectional curvature of each segment of the cross-sectional profile was calculated using the adjacent three-point coordinate difference method. For example, the coordinates of the three points of the profile at the depression were (2.0, 31.5), (2.5, 31.8), and (3.0, 32.0). The cross-sectional curvature calculated by the adjacent three-point coordinate difference method was 0.12. Set the curvature threshold to 0.1. For sections with a curvature > 0.1 The area of ​​the inner region of the cross-sectional profile corresponding to the concave area, i.e., the area of ​​the small unit of the concave region, is controlled to be 0.2. The area of ​​each small unit in the canal bottom region is set to 0.4. Finally, it is discretized into 180 small units, and each small unit records the corresponding area and the coordinates of the geometric center point, which is the position.

[0057] Six vertical lines are laid out laterally at 2m intervals along the cross-section of the monitoring section. Three ultrasonic flow velocity sensors are installed vertically at 0.8m intervals along each vertical line, forming a 6×3 two-dimensional flow velocity measurement matrix. All ultrasonic flow velocity sensors in this matrix are sampled synchronously via LoRa wireless to obtain distributed flow velocity data including lateral position, vertical depth, and flow velocity. For example, the lateral position of the second ultrasonic flow velocity sensor on the third vertical line is 5.0m, the vertical depth is 1.6m, and the flow velocity is 1.8m / s. A preset distance threshold of 1.2m is set. For each small unit, ultrasonic flow velocity sensors whose straight-line distance from their geometric center point coordinates is less than the distance threshold are selected to form a layered flow velocity sensor set. For example, the geometric center point coordinates of the 50th small unit are (4.5m, 1.5m). For example, three ultrasonic flow sensors that meet the criteria are selected (x=5.0m, y=1.6m, v=1.8m / s; x=3.0m, y=1.6m, v=1.6m / s; x=5.0m, y=0.8m, v=1.5m / s). The spatial straight-line distance between the 50th small unit and each ultrasonic flow sensor is calculated using the spatial straight-line distance formula. Using the reciprocal of the spatial straight-line distance as the weight, the flow velocity of the 50th small unit is calculated to be approximately 1.71m / s using the inverse distance weighted interpolation formula. For the 120th small unit in the non-layered flow sensor set, based on the flow velocities of its three adjacent small units of 1.4m / s, 1.3m / s, and 1.5m / s, the flow velocity of the 120th small unit is calculated to be 1.4m / s using linear interpolation.

[0058] A velocity tolerance threshold of 0.2 m / s was set. The DBSCAN density clustering algorithm was used to traverse 180 small cells, grouping adjacent cells with a velocity difference ≤ the velocity tolerance threshold into one class, ultimately aggregating them into 25 large cells. The total area of ​​each large cell is the sum of the areas of its constituent small cells, and the average velocity is the area-weighted average of the velocities of its constituent small cells. For example, the 8th large cell contains 12 small cells with a total area of ​​3.6 m / s. The average flow velocity is 1.6 m / s. Substituting the flow velocities and areas of all small units into the area-weighted average flow velocity formula, the baseline flow velocity is calculated to be 1.55 m / s. Then, substituting the baseline flow velocity and the average flow velocity of each large unit into the flow velocity deviation formula, the flow velocity deviation of each large unit is calculated. Taking the 8th large unit as an example, its flow velocity deviation is 3.2%.

[0059] A flow velocity anomaly threshold of 55% was set. Twenty-five large cells were traversed. If the flow velocity deviation of a large cell exceeded the threshold, that large cell was marked as a flow velocity anomaly cell. The 12th, 13th, and 14th large cells were marked in total.

[0060] This is a velocity anomaly unit. Within this unit, the 12th large unit contains two spatially adjacent velocity anomaly units. The clustered region is a continuous area comprised of the 12th, 13th, and 14th large units, with an area of ​​10.0. The minimum circumscribed circle area of ​​this cluster region, calculated using the minimum enclosing circle algorithm, is 12.5. Since 12.5 / 10.0 = 1.25 > 0.3, the aggregation area is determined to be one vortex region, meaning there is only one vortex region. A vortex feature vector is constructed, which includes the number of vortex regions N = 1, vortex region area characteristics (vortex region area percentage 46.3%, largest vortex region area percentage 46.3%), and channel flow characteristics (flood control channel width-to-depth ratio 6.7, cross-sectional Reynolds number). The vortex feature vector is input into a pre-trained gradient boosting regression tree. The training dataset of the pre-trained gradient boosting regression tree contains 500 sets of correspondences between vortex feature vectors and velocity distribution coefficients under different flow regimes. The output velocity distribution coefficient is 0.98. Finally, the total flow rate of the flood control channel is calculated using the formula for calculating the total flow rate of the flood control channel, which is 47.2. This result is consistent with the actual flow capacity of the channel, which is 45. -50 The matching verified the accuracy of this method.

[0061] The above-described embodiments are only used to illustrate the technical solutions of this application, and are not intended to limit them; modifications to the technical solutions described in the foregoing embodiments, or equivalent substitutions of some of the technical features, do not cause the essence of the corresponding technical solutions to deviate from the scope of the technical solutions of the embodiments of this application, and should all be included within the protection scope of this application.

Claims

1. A method for monitoring flood control channel flow using a fusion velocity-area method, characterized in that, Includes the following steps: S1. The outline of the water-passing section at the monitoring section of the flood control channel is obtained in real time through three-dimensional scanning technology, and the outline of the water-passing section is discretized into n small units, and the corresponding area and position of each small unit are recorded. S2. Obtain the distributed velocity data at the monitoring section of the flood control channel, and distribute the distributed velocity data to each small unit using the inverse distance weighted interpolation method to obtain the velocity of each small unit; S3. Set a flow rate tolerance threshold. Based on the flow rate of the small units, aggregate n small units into m large units and calculate the total area and average flow rate of each large unit. At the same time, use the area-weighted flow rate of all small units as the benchmark flow rate to evaluate the flow rate deviation of each large unit. S4. Identify vortex zones based on the velocity deviation of each large unit and count the number of vortex zones. Dynamically adjust the velocity distribution coefficient according to the number of vortex zones. Based on this, combine the average velocity and total area of ​​each large unit to calculate the total flow of the flood control channel through zonal integration.

2. The flood control channel flow monitoring method based on the velocity-area method according to claim 1, characterized in that, The method for obtaining the water flow cross-section contour is as follows: using lidar to collect three-dimensional point cloud data of the water surface at the monitoring section of the flood control channel, and unifying the three-dimensional point cloud data of the water surface to a coordinate system with the fixed reference point of the flood control channel as the origin, and then aligning the point cloud with the pre-stored point cloud of the flood control channel contour during the dry season through the ICP registration algorithm to obtain the fused point cloud. An instantaneous cross-section is extracted from the fused point cloud along the direction of water flow perpendicular to the flood control channel. The water surface point set and the channel boundary point set in the instantaneous cross-section are extracted, and the alpha shape algorithm is used to fit the water surface point set and the channel boundary point set to generate the water passage cross-section profile.

3. The flood control channel flow monitoring method based on the velocity-area method according to claim 2, characterized in that, The discretization of the cross-sectional profile of the water passage is achieved by using the cross-sectional profile of the water passage as the profile boundary and employing Delaunay triangulation to divide the region within the profile boundary into n small units. The number of small units n is 50-300, and is dynamically adjusted according to the cross-sectional curvature of the cross-sectional profile of the water passage.

4. The flood control channel flow monitoring method based on the velocity-area method according to claim 1, characterized in that, The method for obtaining the distributed flow velocity data is as follows: several equally spaced vertical lines are laid out along the transverse direction on the cross section of the flood control channel monitoring section. Layered flow velocity sensors are set at equal intervals along the vertical direction on each vertical line to form a two-dimensional flow velocity measurement matrix. The layered flow velocity sensor is any one of a propeller-type flow velocity sensor, an electromagnetic flow velocity sensor, and an ultrasonic flow velocity sensor. All layered velocity sensors in the two-dimensional velocity measurement matrix are sampled synchronously via LoRa wireless to obtain distributed velocity data including lateral position, vertical depth, and velocity.

5. The flood control channel flow monitoring method based on the velocity-area method according to claim 3, characterized in that, The method for obtaining the flow velocity of the small unit is as follows: For each small unit, all layered flow velocity sensors whose straight-line distance from the small unit is less than a preset distance threshold are taken as the layered flow velocity sensor set of the small unit. The spatial straight-line distance between the position of the small unit and each layered flow velocity sensor in the layered flow velocity sensor set is calculated. The reciprocal of the spatial straight-line distance is substituted into the inverse distance weighted interpolation formula to calculate the flow velocity of the small unit. If a small unit does not have a set of hierarchical flow velocity sensors, the flow velocity of that small unit is calculated by linear interpolation based on the flow velocities of its adjacent small units.

6. The flood control channel flow monitoring method based on the velocity-area method according to claim 5, characterized in that, The aggregation uses the DBSCAN density clustering algorithm, with the flow rate tolerance threshold as the cluster radius, to aggregate n small units into m large units, where m≤n; The total area of ​​each large unit is the sum of the areas of its constituent small units, and the average flow velocity of each large unit is the area-weighted average of the flow velocities of its constituent small units.

7. The flood control channel flow monitoring method based on the velocity-area method according to claim 5, characterized in that, The method for calculating the velocity deviation is as follows: the area-weighted velocity, i.e. the reference velocity, is calculated by substituting the velocity and area of ​​all small units into the area-weighted average velocity formula. Based on this, the reference velocity and the average velocity of the large units are substituted into the velocity deviation formula to calculate the velocity deviation of the large units.

8. The flood control channel flow monitoring method based on the velocity-area method according to claim 7, characterized in that, The method for identifying the vortex region is as follows: set a velocity anomaly deviation threshold, traverse all large units, and if the velocity deviation of a large unit is greater than the velocity anomaly deviation threshold, then mark the large unit as a velocity anomaly unit. If there are three or more spatially adjacent large units of a velocity anomaly unit and the ratio of the smallest circumscribed circle of the aggregation region to the area of ​​the aggregation region is greater than 0.3, then the aggregation region is determined to be a vortex region. The aggregation region is a continuous region jointly formed by the velocity anomaly unit and all velocity anomaly units spatially adjacent to the velocity anomaly unit.

9. The flood control channel flow monitoring method based on the velocity-area method according to claim 7, characterized in that, The velocity distribution coefficient is obtained by inputting the vortex feature vector into a pre-trained gradient boosting regression tree. The vortex feature vector includes the number of vortex regions, the area of ​​the vortex regions, and the channel flow characteristics.

Images