Hydrological section automatic flow measurement control method based on multi-mode adaptive switching

Abstract

The application discloses a kind of based on multi-mode adaptive switching hydrological section automatic flow measurement control method, it is related to water conservancy and hydrology automatic measurement control technical field;Single threshold value determination is inaccurate, mode switching leads to data breakpoint, scheme fixed is not applicable to dynamic working condition Problem.The application fuses multi-source hydrological perception data calculation section comprehensive working condition abnormality parameter, determines adaptive threshold value based on historical contemporaneous statistical characteristics to execute three-zone switching decision, constructs precision benefit function, safety risk function and time efficiency function to solve pareto optimal scheme set and select optimal scheme to drive truss automatic flow measurement car execution, in double mode parallel transition mode, according to the relative position of the section comprehensive working condition abnormality parameter in transition interval to determine weighting coefficient and to flow data Weighted fusion.The application realizes the adaptive smooth switching of flow measurement mode under all working conditions and the dynamic optimization of flow measurement scheme.

Description

Technical Field

[0001] This invention belongs to the field of automatic measurement and control technology of water conservancy and hydrology, specifically a method for automatic flow measurement and control of hydrological sections based on multi-mode adaptive switching. Background Technology

[0002] Hydrological cross-sectional flow is one of the fundamental hydrological elements in water resource management, flood control and drought relief, water allocation, and water environment protection. In various hydrological flow measurement sections, such as natural rivers, irrigation canals, and reservoir spillways, a mobile flow measurement method is typically used to obtain flow data. The traditional mobile flow measurement method relies on a hydrological bridge or cableway carrying a current meter that moves laterally along the cross-section. At each velocity measurement vertical, the current meter is lowered to a specified relative water depth to collect velocity data point by point. Combined with water level observations and cross-sectional geometric parameters, the flow rate is calculated using the velocity-area method. This method, with its measurement point layout and flow calculation process conforming to the technical requirements of relevant hydrological surveying industry standards, is currently the most widely used flow measurement method in hydrological station networks.

[0003] With the increasing demands for real-time, continuous, and safe flow monitoring in the water conservancy and hydrology industry, both contact and non-contact flow velocity measuring instruments have been gradually introduced into the field of automatic flow measurement. Contact rotor current meters directly sense water flow motion through propellers, offering high measurement accuracy in clean water bodies and under low to medium flow velocity conditions. Non-contact radar current meters emit electromagnetic waves towards the water surface and receive the reflected echoes, measuring surface velocity through the Doppler frequency shift effect. In harsh hydrological conditions such as high flood levels, abundant floating debris, or high sediment content, this type of instrument avoids submersion, ensuring equipment safety.

[0004] The following problems exist in the existing technology: Existing methods generally use the cross-sectional water level as the basis for mode switching. When the water level exceeds a certain preset value, the switching is triggered. However, in actual hydrological environments, there are situations where the water level is not high but the sediment content has increased significantly, or the water level is still within the normal range but a large amount of floating debris has appeared. A single water level threshold cannot identify such working conditions where the water level dimension is normal but other dimensions have deteriorated, leading to inaccurate mode switching judgments. In such working conditions, contact instruments are still operating in the water, which poses a risk of equipment damage. Existing methods directly switch from contact measurement to non-contact measurement at the threshold, lacking a transition between the two modes. This rigid jump causes a step break in the flow data before and after the switching moment due to the different measurement principles, disrupting the continuity of the flow time series and making it difficult to meet the requirements of data continuity for hydrological data compilation. In existing methods, the number of verticals and the number of measurement points used in different flow measurement modes are usually preset, and one or a few fixed schemes are used throughout the entire operating range. When the hydrological conditions of the cross section change significantly, the fixed scheme cannot dynamically balance the flow measurement accuracy, operational safety and flow measurement timeliness. Either during the high flood season, the scheme is too detailed and takes too long, increasing the risk to the equipment, or during the normal water season, the scheme is too coarse and loses data accuracy. Summary of the Invention

[0005] The present invention aims to solve at least one of the technical problems existing in the prior art; to this end, the present invention proposes an automatic flow measurement and control method for hydrological cross sections based on multi-mode adaptive switching, which is used to solve the above-mentioned technical problem.

[0006] The first aspect of the present invention provides an automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching, comprising the following steps: S1: Collect multi-source hydrological sensing data of the cross section using the rotor current meter, radar current meter and auxiliary sensors mounted on the truss automatic flow measurement vehicle, and fuse the multi-source hydrological sensing data to calculate the cross section comprehensive working condition anomaly parameter. S2: Based on the statistical characteristics of the historical hydrological data of the cross section, determine the rotor mode holding threshold and the radar wave mode triggering threshold, compare the cross section comprehensive working condition anomaly parameter with the rotor mode holding threshold and the radar wave mode triggering threshold, and execute a three-zone switching decision including rotor current meter mode, dual-mode parallel transition mode and radar wave current meter mode according to the comparison result. S3: Under the flow measurement mode determined by the three-zone switching decision, construct the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function. Solve the Pareto optimal solution set including the number of vertical lines and the number of measurement point layers in the three-dimensional target space composed of the three functions. Based on the comparison results of the cross-section comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, select the optimal flow measurement scheme composed of the optimal number of vertical lines and the optimal number of measurement point layers from the Pareto optimal solution set, and drive the truss automatic flow measurement vehicle to execute it. S4: After the truss automatic flow measurement vehicle executes the optimal flow measurement scheme, when the three-zone switching decision is the dual-mode parallel transition mode, the weighting coefficient is determined according to the relative position of the cross-sectional comprehensive working condition anomaly parameter within the transition interval defined by the rotor mode holding threshold and the radar wave mode triggering threshold. The flow data measured by the rotor flow meter and the radar wave flow meter are weighted and fused to output the cross-sectional flow.

[0007] Preferably, the multi-source hydrological sensing data in S1 includes at least: the real-time cross-sectional water level value H(t) collected by the radar water level gauge, the water level variability dH / dt obtained by linear fitting of the radar water level gauge time series through a sliding window, and the vertical velocity distribution skewness coefficient S calculated by the rotor velocity meter after performing stratified velocity measurements at multiple specified relative water depths along the current vertical. vel (t), and the instantaneous value I of the electric winch drive current collected by the electric winch drive current sensor in the auxiliary sensor. wch (t).

[0008] Preferably, in step S1, the calculation formula for the cross-sectional comprehensive working condition anomaly parameter CSI(t) is as follows: , Among them, H base (t) represents the multi-year average water level of the cross-section during the current calendar period, and α is the nonlinear amplification exponent of the water level deviation term. I is the historical standard deviation of the current seasonal water level variability of the cross-section. base β is the reference current value when the electric winch is operating normally, β is the current anomaly sensitivity index, and k is the velocity distortion contribution coefficient.

[0009] Preferably, in step S2, determining the rotor mode hold-up threshold and the radar wave mode trigger threshold based on the statistical characteristics of historical hydrological data of the cross section includes the following steps: During the current calendar period of the cross section, the historical data of water level process line, water level variability process line, electric winch drive current process line and flow velocity distribution skewness coefficient of the same period in previous years are read from the historical database of the hydrological center station to which the cross section belongs. The historical data of the same period in previous years are back-calculated according to the calculation method of the cross-section comprehensive working condition anomaly parameter to obtain the historical cross-section comprehensive working condition anomaly parameter sequence of the cross-section in the current calendar period. The anomaly parameter sequence of the comprehensive working conditions of the historical cross-sections is statistically sorted, and the Pth value is taken. rotor Percentile is used as the rotor mode preservation threshold, and is taken as the Pth percentile. radar Percentile is used as the radar wave mode trigger threshold, where P rotor <P radar ;P rotor and P radar The value of P is set according to the season, with the value used during the main flood season. rotor and P radar The P values ​​were lower than those used during the dry season. rotor and P radar .

[0010] Preferably, in step S2, the three-zone switching decision, including rotor current meter mode, dual-mode parallel transition mode, and radar wave current meter mode, is performed based on the comparison result, including the following steps: When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the rotor flow meter mode is adopted for decision-making. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the decision is made to adopt the radar wave current meter mode and send a lifting command to the electric winch to retrieve the underwater lead fish and rotor current meter to a safe height. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the decision adopts a dual-mode parallel transition mode, and the rotor flow meter and the radar wave flow meter are started simultaneously for parallel acquisition, forming a continuous fuzzy transition interval between the rotor mode holding threshold and the radar wave mode triggering threshold.

[0011] Preferably, in step S3, constructing the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function includes the following steps: The flow measurement accuracy benefit function is constructed as follows: , Where, n v Let n be the number of perpendicular lines. p N represents the number of measurement point layers. max λ1 is the maximum number of perpendicular lines required by the cross-section specification, γ is the accuracy saturation coefficient, η is the vertical resolution sensitivity index, and η is the radar wave mode accuracy loss coefficient. The value is 1 when the three-zone switching decision is in radar wave current meter mode, and 0 otherwise. The operational safety risk function is constructed as follows: , Wherein, CSI(t) is the anomaly parameter of the comprehensive working condition of the cross section, and τ is the vehicle movement risk coefficient. The underwater retention risk coefficient, The value is 0 when the three-zone switching decision is in radar wave current meter mode, and 1 otherwise. The flow measurement time-effect function is constructed as follows: , Among them, t move The average travel time between adjacent perpendicular lines is t. point The average sampling time for flow velocity at a single measuring point is given by L, where L is the current water surface width at the cross-section, and v is the average sampling time at a single measuring point. critical The characteristic flow velocity for the propagation of the flood peak at the cross section is given.

[0012] Preferably, in step S3, the Pareto optimal solution set is solved in the three-dimensional target space composed of three functions; and the optimal flow measurement scheme is selected from the Pareto optimal solution set based on the comparison results of the cross-sectional comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, including the following steps: Iterate through the number of perpendicular lines n v With the number of measurement point layers n p For all possible combinations of values, the three constructed functions are used to calculate the flow measurement accuracy benefit function value, operational safety risk function value, and flow measurement timeliness function value corresponding to each value combination. By comparing each pair, all solutions that satisfy the Pareto optimal condition are selected from all value combinations as the Pareto optimal solution set. The Pareto optimal condition means that for any value combination in the Pareto optimal solution set, there is no other value combination that satisfies that its flow measurement accuracy benefit function value is not lower than that value combination, its operational safety risk function value is not higher than that value combination, its flow measurement timeliness function value is not higher than that value combination, and it is better than that value combination in at least one dimension. When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the scheme that maximizes the flow measurement accuracy benefit function value is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the scheme that maximizes S(x,t)·(1-R(x,t)) is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the number of measuring point layers n is set. p =1, given the total number of perpendicular lines n v From the Pareto optimal solution set composed of the given values, select the solution that satisfies R(x,t)≤R max The schemes are selected, and the scheme that minimizes the value of the flow measurement time-effect function is chosen, where R max This is the preset maximum acceptable risk threshold.

[0013] Preferably, in step S4, the weighting coefficient includes the following steps: Calculate the absolute value of the difference between the radar wave mode trigger threshold and the cross-section comprehensive working condition anomaly parameter, calculate the absolute value of the difference between the radar wave mode trigger threshold and the rotor mode holding threshold, and use the ratio of the former to the latter as a weighting coefficient. When the cross-sectional comprehensive working condition anomaly parameter continuously changes from the rotor mode holding threshold to the radar wave mode triggering threshold within the transition range, the weighting coefficient continuously changes from 1 to 0.

[0014] Compared with the prior art, the beneficial effects of the present invention are: This invention collects real-time water level values, water level variability, vertical velocity distribution skewness coefficient, and instantaneous values ​​of electric winch drive current at the cross-section, and integrates them into a comprehensive cross-section working condition anomaly parameter. It comprehensively and quantitatively judges the cross-section working condition from four dimensions: water level amplitude, water level rise and fall rate, vertical distortion degree of velocity field, and stress state of underwater equipment, thus avoiding the inaccuracy of single-dimensional judgment. This invention establishes a dual-mode parallel transition mode between the rotor mode holding threshold and the radar wave mode trigger threshold, forming a continuous transition interval. This allows the flow measurement mode to smoothly switch from the rotor current meter mode to the radar wave current meter mode via the transition interval, rather than a rigid jump at a single threshold. Simultaneously, within the transition interval, weighting coefficients are determined based on the relative positions of the cross-sectional comprehensive working condition anomaly parameters to perform weighted fusion of the flow data from the rotor current meter and the radar wave current meter, ensuring the continuity of flow output during the switching process and meeting the data continuity requirements of hydrological data compilation. This invention constructs a flow measurement accuracy benefit function, an operational safety risk function, and a flow measurement timeliness function, solves the Pareto optimal solution set in a three-dimensional target space, and selects the optimal flow measurement scheme. This allows the number of vertical lines and the number of measurement point layers to be adaptively adjusted according to the real-time hydrological conditions of the cross section. Under high flood risk conditions, safety and timeliness are prioritized, while under stable conditions, accuracy is prioritized, thus achieving dynamic matching between the flow measurement scheme and the operating conditions. Attached Figure Description

[0015] Figure 1 This is a schematic diagram of the method flow of the present invention. Detailed Implementation

[0016] The technical solution of the present invention will be clearly and completely described below with reference to the embodiments. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments.

[0017] Please see Figure 1 This invention relates to an automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching, comprising the following steps: S1: Collect multi-source hydrological sensing data of the cross section using the rotor current meter, radar current meter and auxiliary sensors mounted on the truss automatic flow measurement vehicle, and fuse the multi-source hydrological sensing data to calculate the cross section comprehensive working condition anomaly parameter. S2: Based on the statistical characteristics of the historical hydrological data of the cross section, determine the rotor mode holding threshold and the radar wave mode triggering threshold, compare the cross section comprehensive working condition anomaly parameter with the rotor mode holding threshold and the radar wave mode triggering threshold, and execute a three-zone switching decision including rotor current meter mode, dual-mode parallel transition mode and radar wave current meter mode according to the comparison result. S3: Under the flow measurement mode determined by the three-zone switching decision, construct the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function. Solve the Pareto optimal solution set including the number of vertical lines and the number of measurement point layers in the three-dimensional target space composed of the three functions. Based on the comparison results of the cross-section comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, select the optimal flow measurement scheme composed of the optimal number of vertical lines and the optimal number of measurement point layers from the Pareto optimal solution set, and drive the truss automatic flow measurement vehicle to execute it. S4: After the truss automatic flow measurement vehicle executes the optimal flow measurement scheme, when the three-zone switching decision is the dual-mode parallel transition mode, the weighting coefficient is determined according to the relative position of the cross-sectional comprehensive working condition anomaly parameter within the transition interval defined by the rotor mode holding threshold and the radar wave mode triggering threshold. The flow data measured by the rotor flow meter and the radar wave flow meter are weighted and fused to output the cross-sectional flow.

[0018] Specifically, the rotor current meter, radar current meter, and auxiliary sensors mounted on the truss-mounted automatic flow measurement vehicle simultaneously collect multi-source hydrological sensing data of the cross-section. This multi-source hydrological sensing data is then fused and calculated into a comprehensive anomaly parameter for the cross-section, quantifying the degree of transition from a stable to a dangerous state. Subsequently, hydrological data from the same period in previous years are retrieved from the hydrological center station to which the cross-section belongs. The historical sequence of comprehensive anomaly parameters is calculated retrospectively, and after statistical sorting of this sequence, the corresponding percentiles are determined as the rotor mode holding threshold and the radar current meter trigger threshold, respectively. The percentiles are set according to the season, with values ​​lower during the main flood season than during the dry season. The real-time calculated comprehensive anomaly parameter is compared with the two thresholds, and a three-zone switching decision is executed based on the comparison results: if the value does not exceed the rotor mode holding threshold, the rotor current meter mode is used; if it is not less than the radar current meter trigger threshold, the radar current meter mode is used, and a lifting command is issued to the electric winch to retrieve the underwater equipment; if it is between these two thresholds, a dual-mode parallel transition mode is adopted, and two sets of instruments are simultaneously activated for parallel data collection. After determining the flow measurement mode, a flow measurement accuracy benefit function, an operational safety risk function, and a flow measurement timeliness function are constructed. The Pareto optimal solution set is solved in the three-dimensional target space, and the optimal flow measurement scheme is selected based on the interval of the cross-sectional comprehensive working condition anomaly parameter. Specifically, accuracy is prioritized in the low-value zone, a balance between accuracy and safety is achieved in the transition zone, and timeliness is prioritized under safety constraints in the high-value zone. The scheme includes the number of vertical lines and the number of measurement point layers, driving the truss automatic flow measurement vehicle to execute. After execution, if it is a dual-modal parallel transition mode, the weighting coefficients are determined according to the relative position of the cross-sectional comprehensive working condition anomaly parameter within the transition interval. The flow data measured by the rotor current meter and the radar current meter are weighted and fused before outputting the cross-sectional flow rate; otherwise, the flow data measured by a single mode is directly output.

[0019] In one embodiment of the present invention, the multi-source hydrological sensing data in S1 includes at least: the real-time cross-sectional water level value H(t) collected by a radar water level gauge, the water level variability dH / dt obtained by linear fitting of the radar water level gauge time series through a sliding window, and the vertical velocity distribution skewness coefficient S calculated by a rotor velocity meter after performing stratified velocity measurements at multiple specified relative water depths along the current vertical line. vel (t), and the instantaneous value I of the electric winch drive current collected by the electric winch drive current sensor in the auxiliary sensor. wch (t).

[0020] Specifically, multi-source hydrological sensing data includes at least the real-time water level value of the cross section, water level variability, vertical velocity distribution skewness coefficient, and instantaneous value of electric winch drive current. These data describe the cross section conditions from four dimensions: cross section water level amplitude, water level rise and fall rate, vertical distortion degree of velocity field, and stress state of underwater equipment.

[0021] The real-time water level H(t) of the cross-section is collected by a radar level gauge installed on the bank or bridge of the flow measurement section. The radar level gauge is installed perpendicular to the water surface, and its installation elevation is precisely calibrated by leveling. The radar level gauge emits electromagnetic wave pulses to the water surface at a fixed sampling frequency of 1Hz and receives the reflected echoes. The distance from the sensor to the water surface is obtained by measuring the round-trip time of the electromagnetic wave. The real-time water level value of the cross-section is then obtained by subtracting this distance from the installation elevation. For cross-sections with more drastic water level changes, the sampling frequency will be increased to 2Hz to 5Hz. H(t) is continuously recorded in chronological order of acquisition and cached in the flow measurement control terminal, thus forming a time series of real-time water level values ​​of the cross-section.

[0022] The water level variability dH / dt is obtained by linearly fitting a sliding window to the time series of real-time water level values ​​at the cross-section. The sliding window width ΔT w The selection of ΔT needs to consider both the smoothness of the water level variability calculation and the response speed to sudden changes in water level; for cross-sections with relatively stable flow conditions, ΔT w 120s can be taken; for mountainous sections or sections during the flood season where water levels rise and fall rapidly, ΔT w The timeframe can be from 30s to 60s. In one specific embodiment, the current time t is taken as the right endpoint of the window, and t-ΔT is taken forward. w For all cross-sectional real-time water level sampling points within time period t, a univariate linear regression is performed with time as the independent variable and the cross-sectional real-time water level as the dependent variable. The slope of the regression line is the water level variability at the current time t. The sliding window advances forward once every 1 second, thus obtaining a water level variability sequence that is updated synchronously with the water level data (i.e., the cross-sectional real-time water level value and the cross-sectional real-time water level value time series).

[0023] Vertical velocity distribution skewness coefficient S vel(t) is calculated by a rotor current meter after performing stratified velocity measurements at multiple specified relative depths along the current vertical line. The rotor current meter is installed below the electric winch of the truss automatic flow measurement vehicle and connected to the lead weight suspended by the electric winch, which is lowered to the predetermined water depth position along with the lead weight. On a vertical line, following the conventional multi-point flow measurement method, three specified relative depth positions are selected for velocity measurement: relative depths of 0.2, 0.6, and 0.8, defined as 0 for the water surface and 1 for the riverbed. At each measurement point, the rotor current meter remains stationary and continuously samples. The sampling duration at a single point must meet the requirement that the velocity pulsation error is within an acceptable range. Based on conventional experience, the sampling duration at a single point is generally 30s to 60s, which can be appropriately extended to 60s for sections with strong turbulence. The arithmetic mean of the sampling periods at different relative depths is taken as the representative velocity v at that measurement point. 0.2 v 0.6 and v 0.8 Based on three representative velocity values, the skewness coefficient of the velocity distribution along this vertical line is calculated according to the definition of the skewness coefficient in statistics. , in, For v 0.2 v 0.6 v 0.8 The arithmetic mean of three measurements, v i Take v in sequence 0.2 v 0.6 v 0.8 When the vertical velocity distribution is approximately symmetrical, the skewness coefficient approaches 0; for example, under uniform flow conditions in a standard open channel, S vel (t) within ±0.05 is considered normal measurement fluctuation.

[0024] Instantaneous value of electric winch drive current I wch (t) is collected by the electric winch drive current sensor in the auxiliary sensor; the current sensor is connected in series in the power supply circuit of the electric winch motor, and adopts the non-contact Hall current detection principle to detect the current flowing through the electric winch motor in real time at a sampling frequency of 10Hz. For working conditions with drastic current changes, the sampling frequency can be appropriately increased to 20Hz.

[0025] When the lead weight and rotor velocity meter are freely suspended in still water and the winch operates at a constant speed, the motor current remains at a low steady-state value. This steady-state value is the reference current value I during normal operation of the electric winch. base During equipment installation and commissioning or routine maintenance, the lead weight is suspended in still water, and the electric winch is run unloaded at a normal lifting speed. Motor current data is continuously recorded. After eliminating abnormal values ​​at startup and braking, the arithmetic mean of the current data over the recording period is taken as I. baseIn actual operation, when underwater equipment is subjected to increased drag from the water flow, impact from floating objects, or siltation in the riverbed, the electric winch motor needs to output a larger current to maintain its position or perform lifting operations. Consequently, the instantaneous value of the drive current increases. base The increase is positively correlated with the degree of increase in mechanical load.

[0026] In one embodiment of the present invention, the calculation formula for the cross-sectional comprehensive working condition anomaly parameter CSI(t) in step S1 is as follows: , Among them, H base (t) represents the multi-year average water level of the cross-section during the current calendar period, and α is the nonlinear amplification exponent of the water level deviation term. I is the historical standard deviation of the current seasonal water level variability of the cross-section. base β is the reference current value when the electric winch is operating normally, β is the current anomaly sensitivity index, and k is the velocity distortion contribution coefficient.

[0027] Specifically, water level records for the same calendar period over the past 5 to 15 years are retrieved from the historical database of the hydrological station to which the cross-section belongs. The daily or monthly average water levels for each year are statistically analyzed, and the multi-point moving average for that period is calculated as H. base (t). Taking a wide and shallow plain river section as an example, if the current date is mid-July and the historical average water level for the same period is 2.35m, then H base (t) = 2.35 m.

[0028] The nonlinear amplification exponent α of the water level deviation term is used to ensure that the contribution of hydrological conditions to CSI(t) increases nonlinearly and rapidly when the water level deviation is large, while requiring α > 1. The calibration of α is based on historical high-flood hydrological data for the section, selecting several representative flood events, and using manually compiled condition labels (such as clearly defined high-flood periods) as the target. α is adjusted to ensure a significant jump in CSI(t) within the corresponding period. Typical calibration methods use least squares or grid search to maximize the correlation coefficient between the calculated CSI(t) and the labeled condition level. Based on calibration experience across multiple sections, the value of α generally ranges from 1.2 to 2.0; for example, for mountain stream sections with large annual water level variations, α is taken as 1.8 to 2.0.

[0029] Utilizing H base (t) Using the same historical hydrological data, take the time series of water level variability dH / dt for the same period in previous years, and calculate the standard deviation of the series after removing obviously unreasonable jump points. For example, statistically analyze the water level variability of a certain cross section in mid-July of previous years to obtain... =0.12m / h.

[0030] The current anomaly sensitivity index β is used to control the overall sensitivity of the relative change in the electric winch drive current to CSI(t). Its calibration method is similar to α, also based on historical high-flood data from the cross-section; and β > 0. When determining I... base Then, I was extracted during the periods of significant current increase in each flood event. wch (t) data, using manually judged periods of abnormal equipment stress as a reference, adjust β to make the mechanical abnormality term (i.e., During normal operation, β remains stable around 1, but increases significantly when encountering obstruction or siltation. Based on multi-section experience, the value of β is generally taken from 0.5 to 1.5; when the water in the section has a high sediment content and a lot of floating debris, β can be appropriately taken as a larger value (such as 1.2 to 1.5) to enhance the response to abnormal current.

[0031] The velocity distortion contribution coefficient k is used to adjust the contribution ratio of the vertical velocity distribution skewness coefficient to CSI(t), and k>0. The calibration of k utilizes historical anomalous flow records of this cross-section, selecting typical periods of significant velocity distribution distortion, and calculating St during these periods. vel Based on (t), k is adjusted so that the CSI(t) corresponding to this time period reflects the corresponding increase in danger, while ensuring that the velocity distortion term (i.e., (1+k·S) is maintained under normal flow conditions. vel The overall value of (t) is close to 1, which will not cause substantial disturbance to CSI(t). The common range of k is 0.5 to 1.5; for cross-sections with complex cross-sectional shapes and strong natural asymmetry in velocity distribution, k can be appropriately reduced (e.g., 0.5 to 0.8) to avoid normal flow distortion being misjudged as deterioration of operating conditions.

[0032] α, β, k and standard deviation The calibration process uses historical hydrological data from at least 3 to 5 typical flood events at the cross section. The working condition level (e.g., stable, alert, and high flood) is manually labeled for different time periods of each flood event. The parameter combinations of α, β, and k within their respective value ranges are traversed, and the CSI(t) value for the corresponding time period is calculated for each parameter combination. The consistency between the CSI(t) value and the manually labeled working condition level is compared. The parameter combination that achieves the optimal consistency is selected as the calibration result.

[0033] The magnitude of CSI(t) directly reflects the degree to which the current operating conditions of the cross section deviate from historical norms. The closer CSI(t) is to 1, the closer the real-time water level, water level variability, instantaneous value of electric winch drive current, and skewness coefficient of vertical velocity distribution are to historical norms, indicating a stable hydrological environment. The larger the CSI(t), the more significantly one or more of the four dimensions of cross section water level amplitude, water level rise and fall rate, vertical distortion of velocity field, and stress state of underwater equipment have deviated from the historical baseline. The higher the overall risk level of the cross section, the more it tends to be a dangerous operating condition that requires the use of non-contact radar current meters to measure and lift underwater equipment.

[0034] In one embodiment of the present invention, step S2, based on the statistical characteristics of historical hydrological data of the cross section, determines the rotor mode maintenance threshold and the radar wave mode triggering threshold, including the following steps: During the current calendar period of the cross section, the historical data of water level process line, water level variability process line, electric winch drive current process line and flow velocity distribution skewness coefficient of the same period in previous years are read from the historical database of the hydrological center station to which the cross section belongs. The historical data of the same period in previous years are back-calculated according to the calculation method of the cross-section comprehensive working condition anomaly parameter to obtain the historical cross-section comprehensive working condition anomaly parameter sequence of the cross-section in the current calendar period. The anomaly parameter sequence of the comprehensive working conditions of the historical cross-sections is statistically sorted, and the Pth value is taken. rotor Percentile is used as the rotor mode preservation threshold, and is taken as the Pth percentile. radar Percentile is used as the radar wave mode trigger threshold, where P rotor <P radar ;P rotor and P radar The value of P is set according to the season, with the value used during the main flood season. rotor and P radar The P values ​​were lower than those used during the dry season. rotor and P radar .

[0035] Specifically, hydrological monitoring data for the same period in previous years are retrieved from the historical database of the hydrological center station to which the cross-section belongs. The retrieved data includes historical water level process lines, water level variability process lines, electric winch drive current process lines, and historical records of flow velocity distribution skewness coefficients. The time span of the historical data is generally 5 to 15 years.

[0036] For the hydrological data from the same period in previous years, the Comprehensive Anomaly Index (CSI) of the cross-section is calculated retrospectively point by point according to the calculation method of the CSI, to obtain the historical CSI sequence of the cross-section in the current calendar period. For each hydrological data sample point from the same period in previous years, a historical CSI value is calculated, and all historical CSI values ​​are arranged in chronological order to form the historical CSI sequence of the cross-section.

[0037] The sequence of anomaly parameters of the comprehensive working conditions of the historical cross-section is sorted in ascending order from smallest to largest, and its Pth value is taken. rotor Percentile as rotor mode preservation threshold θ rotor Take the Pth radar Percentile as radar wave mode trigger threshold θ radar , where P rotor <P radar For example, if P rotor =30%, then the CSI value does not exceed θ at 30% of the times in the historical cross-section comprehensive working condition anomaly parameter sequence. rotor These moments generally correspond to stable or relatively normal hydrological conditions; similarly, if P radar =85%, which means that historically, the CSI value did not exceed θ 85% of the time. radar The remaining 15% of the times that exceed this threshold correspond to historically rare high flood conditions or dangerous working conditions.

[0038] P rotor and P radar The value of P varies with the season; the seasons are divided into three categories based on the hydrological characteristics of the cross-section: the main flood season, the dry season, and the normal water season. The main flood season generally corresponds to the months when torrential rains and floods occur frequently. For example, for a cross-section of a tributary in the middle reaches of the Yangtze River, the main flood season is set from June to September; the dry season is set from December to February of the following year; and the normal water season is from March to May and from October to November. Different percentile values ​​are used in different seasons; the P value used in the main flood season is... rotor and P radar All were lower than the P used during the dry season. rotor and P radar A typical example of P's value is during the main flood season. rotor Take 20% to 30%, P radar Take 70% to 80%; during the dry season, P rotor Take 40% to 50%, P radar Take 85% to 95%; during the normal water period, the two percentiles are taken as the midpoint between the two seasons mentioned above, for example, P. rotor Take 30% to 40%, P radar Take 80% to 85%.

[0039] In one embodiment of the present invention, step S2 involves performing a three-zone switching decision based on the comparison result, including the rotor current meter mode, the dual-mode parallel transition mode, and the radar wave current meter mode, comprising the following steps: When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the rotor flow meter mode is adopted for decision-making. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the decision is made to adopt the radar wave current meter mode and send a lifting command to the electric winch to retrieve the underwater lead fish and rotor current meter to a safe height. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the decision adopts a dual-mode parallel transition mode, and the rotor flow meter and the radar wave flow meter are started simultaneously for parallel acquisition, forming a continuous fuzzy transition interval between the rotor mode holding threshold and the radar wave mode triggering threshold.

[0040] Specifically, when CSI(t) ≤ θ rotor At that time, the cross-sectional comprehensive working condition anomaly parameter CSI(t) did not exceed the rotor mode hold threshold θ. rotor This indicates that the real-time water level, water level variability, electric winch drive current, and vertical velocity distribution at the cross-section are all within the normal range for the same period in history, indicating a stable hydrological environment. In this case, the rotor current meter mode is used. The rotor current meter mode refers to the use of the rotor current meter to perform contact velocity measurements. Utilizing its ability to deploy measuring points in layers and directly contact the water body, it obtains high-precision vertical velocity distribution data. In this mode, the radar current meter is in standby mode and does not participate in the data acquisition for this measurement.

[0041] When CSI(t) ≥ θ radar At that time, the anomaly parameter of the cross-section's comprehensive working condition has reached or exceeded the radar wave mode trigger threshold θ. radar This indicates that the cross-section has entered a high-water hydrological condition or a dangerous operating condition, at which point the radar current meter mode is adopted. The radar current meter mode refers to the use of a non-contact radar current meter to measure flow velocity. This instrument does not contact the water body; it measures surface flow velocity by emitting electromagnetic waves towards the water surface and receiving the reflected echoes. Simultaneously, after the decision is executed, a lifting command is immediately issued to the electric winch to retrieve the underwater lead weight and rotor current meter to a safe height, thus removing the underwater equipment from areas where floating debris may impact or where high-sediment-laden water flow may scour.

[0042] When θ rotor <CSI(t)<θ radarAt this point, the cross-sectional comprehensive working condition anomaly parameter CSI(t) is between two thresholds, indicating that the cross-section is in a critical transition state from stable to dangerous. In this case, a dual-modal parallel transition mode is adopted. The dual-modal parallel transition mode refers to the simultaneous activation of the rotor current meter and radar current meter, which simultaneously measure the velocity of the cross-section within the same measurement cycle. The rotor current meter still collects vertical stratified velocity data in a contact manner, while the radar current meter collects surface velocity data in a non-contact manner. The two instruments independently complete their respective measurement processes. The rotor current meter performs multi-point contact measurements based on the number of verticals and measurement layers, and calculates the flow rate using the velocity-area method. The radar current meter completes cross-sectional scanning and calculates the flow rate using surface velocity coefficient conversion and the velocity-area method. Both instruments produce their respective complete cross-sectional flow rate calculation results at the end of the measurement cycle. In this mode, the underwater lead weight and rotor current meter remain operational in the water. The dual-modal parallel transition mode at θ... rotor With θ radar A continuous transition interval is formed between them; when CSI(t) changes within this transition interval, the three-zone switching decision will not jump directly between the rotor flow meter mode and the radar wave flow meter mode, thus avoiding the flow sequence breakpoint caused by rigid mode switching at a single threshold; when CSI(t) continuously rises from the low value zone that does not exceed the rotor mode holding threshold to the high value zone that is not less than the radar wave mode trigger threshold, the three-zone switching decision successively goes through the rotor flow meter mode, the dual-mode parallel transition mode, and the radar wave flow meter mode, and finally the output cross-sectional flow gradually transitions from the flow data measured by the rotor flow meter to the flow data measured by the radar wave flow meter.

[0043] Taking a specific cross-section as an example, assuming that after historical data statistics and calibration, the rotor mode of this cross-section maintains the threshold θ during the main flood season. rotor =1.8, radar wave mode trigger threshold θ radar =5.2; During a flood, the CSI(t) calculated before the rise in water level is 1.2, not exceeding 1.8, so the rotor current meter mode is executed, and only the rotor current meter performs contact measurement. As the water level rises, CSI(t) gradually rises to 3.5, falling between 1.8 and 5.2, and switches to the dual-mode parallel transition mode, simultaneously starting the rotor current meter and radar current meter to collect data; When the flood peak approaches, CSI(t) rises to 6.0, reaching above 5.2, and switches to the radar current meter mode. The electric winch executes the lifting command to retrieve the underwater lead weight and rotor current meter to a safe height, while the radar current meter completes the flow measurement alone. During the flood receding phase, CSI(t) falls back to 4.0, and the dual-mode parallel transition mode is entered again; When CSI(t) continues to fall back to 1.5, the rotor current meter mode is restored.

[0044] In one embodiment of the present invention, step S3, which involves constructing the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function, includes the following steps: The flow measurement accuracy benefit function is constructed as follows:

[0045] Where, n v Let n be the number of perpendicular lines. p N represents the number of measurement point layers. max λ1 is the maximum number of perpendicular lines required by the cross-section specification, γ is the accuracy saturation coefficient, η is the vertical resolution sensitivity index, and η is the radar wave mode accuracy loss coefficient. The value is 1 when the three-zone switching decision is in radar wave current meter mode, and 0 otherwise. The operational safety risk function is constructed as follows: , Wherein, CSI(t) is the anomaly parameter of the comprehensive working condition of the cross section, and τ is the vehicle movement risk coefficient. The underwater retention risk coefficient, The value is 0 when the three-zone switching decision is in radar wave current meter mode, and 1 otherwise. The flow measurement time-effect function is constructed as follows: , Among them, t move The average travel time between adjacent perpendicular lines is t. point The average sampling time for flow velocity at a single measuring point is given by L, where L is the current water surface width at the cross-section, and v is the average sampling time at a single measuring point. critical The characteristic flow velocity for the propagation of the flood peak at the cross section is given.

[0046] Specifically, n v n represents the number of vertical lines included in the selected optimal flow measurement scheme, i.e., the number of velocity measurement vertical lines deployed by the truss automatic flow measurement vehicle at the flow measurement break section. p N represents the number of measurement point layers in the selected optimal flow measurement scheme, i.e., the number of velocity measurement points laid out on each vertical line. The value can be 1, 2, or 3, corresponding to the one-point method, two-point method, or three-point method, respectively. max The maximum number of perpendiculars required by the cross-section specification is determined by current hydrological surveying industry standards based on the cross-section's water surface width and depth; for example, for a medium-sized river cross-section with a water surface width of 50m to 100m, N max It is usually taken as 10 to 15.

[0047] The current measurement accuracy benefit function S(x,t) is constructed as a three-term product structure, i.e. , The accuracy saturation coefficient λ1 is used to control the rate of decline of marginal accuracy gains as the number of verticals increases, and λ1 > 0. The calibration of λ1 is based on historical flow measurement data for the cross-section. Flow comparison data under several different vertical layout schemes are selected, and the flow value obtained from denser verticals is used as an approximate true value. The relative error of flow measurement under different vertical number schemes is calculated, and the decline curve of the relative error with the number of verticals is fitted with an exponential function to obtain λ1. According to multi-section calibration experience, the value range of λ1 is generally 3.0 to 5.0; for example, when the lateral distribution of flow velocity in the cross-section is relatively uniform, λ1 can take a smaller value (such as 3.0 to 3.5), indicating that the rate at which accuracy tends to saturate with the increase of the number of verticals is slower. The vertical resolution sensitivity index γ reflects the nonlinear contribution of increasing the number of measuring point layers to the accuracy of vertical average velocity estimation, and γ > 0. γ is also calibrated using historical comparative data. Vertical average velocities obtained from the one-point, two-point, and three-point methods at the same vertical line are selected. Using the three-point method result as a benchmark, the relative deviation between the one-point and two-point methods is calculated, and a power function is fitted to the decrease in deviation with the number of measuring point layers. The typical value range of γ is 0.6 to 1.2, with a default value of 1.0. When γ = 1.0, this term has a linear relationship with the number of measuring point layers. When γ > 1.0, the accuracy advantage of the three-point method over the one-point method is amplified, making it suitable for deep-water sections with large vertical velocity gradients. When γ < 1.0, the accuracy tends to level off with the increase in the number of measuring point layers, making it suitable for shallow-water sections with relatively uniform vertical velocity distribution. The accuracy loss factor η of the radar wave mode is used to reflect the inherent accuracy difference between the non-contact measurement of the radar wave current meter and the contact measurement of the rotor current meter, and 0 < η < 1. η is determined through field comparative experiments. At different water level and flow velocity levels, the rotor current meter and the radar wave current meter are used simultaneously to measure the flow velocity along the same vertical line. The average vertical flow velocity obtained from multiple measuring points of the rotor current meter is used as the benchmark. The relative deviation between the radar wave current meter surface flow velocity converted to the average vertical flow velocity and the benchmark value is calculated. The average value of the relative deviations from multiple comparative measurements is taken as η. The typical range of η is 0.10 to 0.25, indicating that the accuracy loss of the radar wave mode is about 10% to 25%. This is an indicator function. It takes the value 1 when the three-zone switching decision is radar wave current meter mode, and 0 when it is rotor current meter mode or dual-mode parallel transition mode. This indicator function ensures that the modal loss only takes effect when the radar wave current meter is used alone to perform flow measurement. In dual-mode parallel transition mode, the rotor current meter still operates normally and collects data. This item takes the value 1, and the accuracy gain is not affected by the modal loss.

[0048] The operational safety risk function R(x,t) is constructed as a three-term product structure, i.e. ; in, This is used to quantify the physical exposure of the truss automatic flow measurement vehicle as it travels back and forth on the track; τ is the vehicle movement risk coefficient, τ>0. The more vertical lines there are, the greater the range and time of the vehicle's movement on the cross section, and the higher the probability of direct impact from floating objects; the calibration of τ is based on the statistical relationship between the historical floating object impact records of the cross section and the frequency of vehicle movement; the number of times the vehicle encounters floating object impacts during the flow measurement process over the years is statistically analyzed and correlated with the number of vertical lines, the width of the cross section water surface, and the frequency of floating object occurrence, and τ is fitted to obtain the result; the typical value range of τ is 0.3 to 0.8; for example, for cross sections of mountain streams or densely vegetated areas where floating objects are frequent, τ takes a larger value (such as 0.6 to 0.8). Used to quantify the physical exposure of underwater equipment; The underwater retention risk coefficient, >0, the more layers of measuring points there are during contact measurement, the longer the underwater stay time of a single vertical line, and the higher the risk of the lead fish and rotor current meter being entangled by floating objects or buried by silt. The calibration is based on the statistical records of historical underwater equipment encounters or damage events at this cross section, and the correlation with the number of measurement point layers and the sampling time at a single point is analyzed. The typical value range is 0.2 to 0.6. The underwater equipment has an indication function. When the three-zone switching decision is the rotor current meter mode or the dual-mode parallel transition mode, it is set to 1. At this time, the lead weight and rotor current meter are suspended below the electric winch, and the underwater equipment is in the water. When the three-zone switching decision is the radar wave current meter mode, it is set to 0. At this time, the lead weight and rotor current meter have been raised to a safe height, and no underwater equipment is exposed.

[0049] The flow measurement time-effect function T(x,t) is constructed as the ratio of the total time to the flood allowable time window, i.e. ; The numerator represents the estimated total operation time required to complete all vertical measurements according to the current plan, and the denominator represents the characteristic time scale of flood evolution. When the T(x,t) value is close to or exceeds 1, it indicates that the estimated flow measurement time has approached or even exceeded the safe time window allowed by the flood, and the scale of flow measurement needs to be reduced. move t represents the measured statistical value of the time required for the truss-mounted automatic flow measurement vehicle to travel between adjacent verticals. Specifically, it is the average time taken for the truss-mounted automatic flow measurement vehicle to travel horizontally along the track to the next vertical after the electric winch has completed sampling at all measurement points on one vertical. This value is obtained by statistically averaging the actual travel times from multiple flow measurements at this cross-section. pointsThis is the measured statistical value of the time required for velocity sampling at a single measuring point, i.e., the average time required for a rotor current meter or radar current meter to stabilize, sample, and record data at a single measuring point. It is also obtained by statistically analyzing the actual sampling time of multiple flow measurements at this cross-section. L is the current water surface width of the cross-section, obtained by interpolation from pre-stored large-section data based on the real-time water level value collected by the radar level gauge. critical The characteristic velocity of the flood peak propagation at the cross-section is determined based on historical flood data analysis of the cross-section; several typical flood events at the cross-section are selected to calculate the average velocity of the flood peak propagation from relevant upstream stations to this cross-section, or the characteristic values ​​of the high-water portion of the curve relating the flow rate and the average velocity at this cross-section are analyzed; v critical The typical range of values ​​for v is related to the gradient and cross-sectional morphology of the river section where the cross-section is located, for example, the v value of a plain river channel. critical Typically between 1.0 m / s and 2.5 m / s.

[0050] In one embodiment of the present invention, in step S3, the Pareto optimal solution set is solved in the three-dimensional target space composed of three functions; and the optimal flow measurement scheme is selected from the Pareto optimal solution set based on the comparison results of the cross-sectional comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, including the following steps: Iterate through the number of perpendicular lines n v With the number of measurement point layers n p For all possible combinations of values, the three constructed functions are used to calculate the flow measurement accuracy benefit function value, operational safety risk function value, and flow measurement timeliness function value corresponding to each value combination. By comparing each pair, all solutions that satisfy the Pareto optimal condition are selected from all value combinations as the Pareto optimal solution set. The Pareto optimal condition means that for any value combination in the Pareto optimal solution set, there is no other value combination that satisfies that its flow measurement accuracy benefit function value is not lower than that value combination, its operational safety risk function value is not higher than that value combination, its flow measurement timeliness function value is not higher than that value combination, and it is better than that value combination in at least one dimension. When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the scheme that maximizes the flow measurement accuracy benefit function value is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the scheme that maximizes S(x,t)·(1-R(x,t)) is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the number of measuring point layers n is set. p =1, given the total number of perpendicular lines n vFrom the Pareto optimal solution set composed of the given values, select the solution that satisfies R(x,t)≤R max The schemes are selected, and the scheme that minimizes the value of the flow measurement time-effect function is chosen, where R max This is the preset maximum acceptable risk threshold.

[0051] Specifically, determine all possible combinations of the number of perpendicular lines and the number of measuring point layers; the number of perpendicular lines n v The lower limit of the value is the minimum number of perpendicular lines allowed by the cross-section specification, and the upper limit of the value is the maximum number of perpendicular lines N required by the cross-section specification. max Number of measurement points n p The value can be 1, 2, or 3. Each value of the vertical line number within its range is paired with each value of the measuring point layer number to form all possible combinations. For example, if the N value of a certain cross-section... max =10, the minimum number of perpendicular lines is 5, so the number of perpendicular lines can be 5, 6, 7, 8, 9, 10, and the number of measuring point layers can be 1, 2, 3, resulting in a total of 18 possible combinations. Each combination corresponds to a candidate flow measurement scheme. The flow measurement accuracy benefit function, operational safety risk function, and flow measurement timeliness function are used to calculate the corresponding flow measurement accuracy benefit function value S, operational safety risk function value R, and flow measurement timeliness function value T for each combination. During calculation, n v Take the value of the perpendicular line in this combination, n p Take the measurement point layer values ​​in the combination; after calculating each combination of values, the corresponding three target values ​​(S, R, T) are obtained for each combination of values.

[0052] For any two combinations of values, compare their corresponding flow measurement accuracy benefit function, operational safety risk function, and flow measurement timeliness function. If one combination of values ​​is no less than the other combination in terms of flow measurement accuracy benefit function, no more than the other combination in terms of operational safety risk function, and no more than the other combination in terms of flow measurement timeliness function, and is strictly superior to the other combination in at least one dimension, then this combination of values ​​is Pareto optimal, and the dominant combination of values ​​is removed from the candidate set. After iterating through all pairwise comparisons between all combinations of values, the remaining combinations of values ​​that are not dominant are the Pareto optimal solutions, and all Pareto optimal solutions constitute the Pareto optimal solution set.

[0053] To illustrate with a specific calculation example, suppose the decision for a certain cross-section is based on the rotor velocity meter mode, CSI(t) = 1.2, N max =10, there are a total of 18 possible combinations. Let's take three combinations for comparison as an example; combination A is n. v =8、n p =3, calculated to be S A =0.88, R A =0.20、TA =0.35; Combination B is n v =6、n p =2, therefore S B =0.72, R B =0.15, T B =0.28; Combination C is n v =10, n p =2, therefore S C =0.85, R C =0.22、T C =0.42. Compare combination A and combination B, S A >S B R A >R B T A >T B Neither is superior to the other; comparing combination A and combination C, S A >S C R A <R C T A <T C Combination A is superior to combination C in all three dimensions, so combination C is eliminated. After all pairwise comparisons, the remaining combinations constitute the Pareto optimal solution set; in this example, combination A may be retained, combination B may be retained, and combination C is eliminated.

[0054] After obtaining the Pareto optimal solution set, based on the cross-sectional comprehensive working condition anomaly parameter CSI(t) and the rotor mode retention threshold θ rotor Radar wave mode trigger threshold θ radar Based on the comparison results, the optimal flow measurement scheme is selected from the Pareto optimal scheme set. When CSI(t) does not exceed θ... rotor At this time, the hydrological environment at the cross-section is stable, the safety risk is low, and there is ample time. Therefore, the optimal flow measurement scheme is selected from the Pareto optimal scheme set, maximizing the benefit function value of the flow measurement accuracy, to obtain the most complete and high-quality hydrological data. When θ rotor <CSI(t)<θ radar At this point, the cross-section is in a critical transition state, requiring a balance between accuracy and safety. The optimal flow measurement scheme is selected from the Pareto optimal scheme set, maximizing S(x,t)·(1-R(x,t)). This product index considers both accuracy gains and safety margins, seeking the optimal balance between the two. When CSI(t) is not less than θ... radar At this time, the cross-section is in a dangerous working condition, and safety and timeliness take priority; at this point, the electric winch has already retrieved the underwater lead weight and rotor current meter to a safe height, and current measurement is performed only by a non-contact radar current meter, with n set. p =1, the alternative solutions consist of all n vThe values ​​of each option form a Pareto optimal solution set; from this set, solutions that satisfy the condition that the operational safety risk function value does not exceed the preset maximum acceptable risk threshold R are selected. max The solution; R max The value of is determined based on the equipment protection level and operation management requirements of the cross section, and is generally between 0.50 and 0.80, with a default value of 0.60. Among the schemes that meet this safety constraint, the scheme that minimizes the flow measurement time effect function value is selected as the optimal flow measurement scheme. In this case, a scheme with fewer vertical lines and a single-point method is usually selected to complete the flow measurement as quickly as possible and ensure that effective measurement data are obtained before significant changes in the flood.

[0055] In one embodiment of the present invention, step S4, the weighting coefficient, includes the following steps: Calculate the absolute value of the difference between the radar wave mode trigger threshold and the cross-section comprehensive working condition anomaly parameter, calculate the absolute value of the difference between the radar wave mode trigger threshold and the rotor mode holding threshold, and use the ratio of the former to the latter as a weighting coefficient. When the cross-sectional comprehensive working condition anomaly parameter continuously changes from the rotor mode holding threshold to the radar wave mode triggering threshold within the transition range, the weighting coefficient continuously changes from 1 to 0.

[0056] Specifically, based on the number of measuring point layers in the optimal flow measurement scheme, the average velocity of each measuring point along a vertical line is calculated using the commonly used vertical average velocity calculation method in hydrological measurements. When the number of measuring point layers is 1, a single-point method is used, with the measured velocity v at a relative water depth of 0.6 being the average velocity. 0.6 The average velocity of the vertical line directly used as the vertical line ,Right now =v 0.6 When the number of measuring points is 2, the two-point method is used, and the measured flow velocity v at a relative water depth of 0.2 is taken. 0.2 The measured flow velocity v at a relative water depth of 0.8 0.8 The arithmetic mean of the values ​​is used as the vertical average velocity, i.e. =0.5(v 0.2 +v 0.8 When the number of measuring points is 3, the three-point method is used, and the measured flow velocity v is taken at relative water depths of 0.2, 0.6, and 0.8. 0.2 v 0.6 v 0.8 The vertical average velocity was calculated using weighting factors of 0.25, 0.50, and 0.25. =0.25v 0.2 +0.50v 0.6 +0.25v 0.8 .

[0057] After calculating the average flow velocity along each vertical line, the cross-sectional area is calculated based on the real-time water level H(t) collected by the radar level gauge and pre-stored large-section data. The large-section data includes riverbed elevation data at each starting point along the cross-sectional direction. During calculation, the water depth h at each location is obtained from the difference between the current H(t) and the riverbed elevation at each starting point. i Where i represents the sequence number of the vertical lines laid out transversely from one bank to the other along the cross-section, i = 1, 2, ..., m, and m is the total number of vertical lines in this flow measurement scheme. The water surface width b between adjacent vertical lines... i The depth is determined by the difference in distance from the starting point. For any two adjacent perpendicular positions i and i+1, the corresponding water depths are h, respectively. i and h i+1 The sub-section is approximated as a trapezoid, and its partial area A i =0.5(h i +h i+1 )·b i The total cross-sectional area of ​​the water passage is obtained by summing the areas between all adjacent perpendicular lines and the areas along both banks. After obtaining the average flow velocity and cross-sectional area of ​​each vertical line, calculate the partial flow rate between adjacent vertical lines i and i+1. The total flow rate of the cross section .

[0058] If the three-zone switching decision is set to rotor velocity meter mode, the flow rate data Q measured by the rotor velocity meter is directly used as the cross-sectional flow rate output for this flow measurement, without performing the subsequent weighted fusion step. If the three-zone switching decision is set to radar velocity meter mode, the flow rate data measured by the radar velocity meter is directly used as the cross-sectional flow rate output for this flow measurement, without performing the subsequent weighted fusion step. The radar velocity meter measures surface velocity in a non-contact manner, and its surface velocity needs to be converted into vertical average velocity according to the surface velocity coefficient determined by the historical comparative measurement data of the cross-section. After conversion, the flow rate data measured by the radar velocity meter is obtained through the same process of superimposing the area and partial flow rate as described above. If the three-zone switching decision is set to dual-mode parallel transition mode, the weighted fusion process is entered. In dual-mode parallel transition mode, the rotor velocity meter and the radar velocity meter are started simultaneously for parallel acquisition, and the flow rate data Q measured by the rotor velocity meter is generated simultaneously in this flow measurement. rotor The flow rate data Q measured by the radar current meter radar Both sets of flow data are scalar values ​​produced after each independently completing the entire cross-sectional flow measurement process. The fusion is directly performed on the Q... rotor and Q radar The process involves two secondary flow scalar measurements without aligning the flow velocity time series data at the sampling point level between the two instruments. The two sets of flow data are then fused using a weighting coefficient to obtain the final output cross-sectional flow rate.

[0059] λ represents the weight of the flow rate data measured by the rotor velocity meter in the fused output, and its value is a continuous value between 0 and 1. The radar wave mode trigger threshold θ is calculated. radar The absolute value of the difference between the cross-sectional comprehensive working condition anomaly parameter CSI(t) and the radar wave mode triggering threshold θ is used to calculate the radar wave mode triggering threshold θ. radar Maintain threshold θ with rotor mode rotor The absolute value of the difference, with the ratio of the former to the latter as the weighting factor, i.e. , Where θ radar -θ rotor This refers to the width of the transition interval. The calculation method of this weighting coefficient determines the relative proportion of the flow rate data measured by the rotor velocity meter and the flow rate data measured by the radar velocity meter in the fused output when CSI(t) is at different positions within the transition interval; when CSI(t) approaches θ... rotor When, the molecule |θ radar -CSI(t) approaches |θ radar -θ rotor When |λ approaches 1, the fused output is dominated by the flow rate data measured by the rotor velocity meter; when CSI(t) approaches θ radar When the molecule approaches 0 and λ approaches 0, the fused output is dominated by the flow rate data measured by the radar wave velocity meter. When CSI(t) is at the midpoint of the transition interval, i.e. When λ=0.5, the flow data measured by the two modes are fused with equal weight. The final cross-sectional flow rate Q is then obtained from this flow measurement. out =λ·Q rotor +(1-λ)·Q radar .

[0060] The above embodiments are only used to illustrate the technical methods of the present invention and are not intended to limit it. Although the present invention has been described in detail with reference to preferred embodiments, those skilled in the art should understand that modifications or equivalent substitutions can be made to the technical methods of the present invention without departing from the spirit and scope of the technical methods of the present invention.

Claims

1. An automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching, characterized in that, Includes the following steps: S1: Collect multi-source hydrological sensing data of the cross section using the rotor current meter, radar current meter and auxiliary sensors mounted on the truss automatic flow measurement vehicle, and fuse the multi-source hydrological sensing data to calculate the cross section comprehensive working condition anomaly parameter. S2: Based on the statistical characteristics of the historical hydrological data of the cross section, determine the rotor mode holding threshold and the radar wave mode triggering threshold, compare the cross section comprehensive working condition anomaly parameter with the rotor mode holding threshold and the radar wave mode triggering threshold, and execute a three-zone switching decision including rotor current meter mode, dual-mode parallel transition mode and radar wave current meter mode according to the comparison result. S3: Under the flow measurement mode determined by the three-zone switching decision, construct the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function. Solve the Pareto optimal solution set including the number of vertical lines and the number of measurement point layers in the three-dimensional target space composed of the three functions. Based on the comparison results of the cross-section comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, select the optimal flow measurement scheme composed of the optimal number of vertical lines and the optimal number of measurement point layers from the Pareto optimal solution set, and drive the truss automatic flow measurement vehicle to execute it. S4: After the truss automatic flow measurement vehicle executes the optimal flow measurement scheme, when the three-zone switching decision is the dual-mode parallel transition mode, the weighting coefficient is determined according to the relative position of the cross-sectional comprehensive working condition anomaly parameter within the transition interval defined by the rotor mode holding threshold and the radar wave mode triggering threshold. The flow data measured by the rotor flow meter and the radar wave flow meter are weighted and fused to output the cross-sectional flow.

2. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 1, characterized in that, The multi-source hydrological sensing data in S1 includes at least: the real-time cross-sectional water level value H(t) collected by the radar water level gauge, the water level variability dH / dt obtained by linear fitting of the radar water level gauge time series through a sliding window, and the vertical velocity distribution skewness coefficient S calculated by the rotor velocity meter after performing stratified velocity measurements at multiple specified relative water depths along the current vertical. vel (t), and the instantaneous value I of the electric winch drive current collected by the electric winch drive current sensor in the auxiliary sensor. wch (t).

3. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 2, characterized in that, In S1, the calculation formula for the cross-sectional comprehensive working condition anomaly parameter CSI(t) is as follows: , Among them, H base (t) represents the multi-year average water level of the cross-section during the current calendar period, and α is the nonlinear amplification exponent of the water level deviation term. I is the historical standard deviation of the current seasonal water level variability of the cross-section. base β is the reference current value when the electric winch is operating normally, β is the current anomaly sensitivity index, and k is the velocity distortion contribution coefficient.

4. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 1, characterized in that, In step S2, based on the statistical characteristics of historical hydrological data from the same period of the cross section, the rotor mode maintenance threshold and the radar wave mode trigger threshold are determined, including the following steps: During the current calendar period of the cross section, the historical data of water level process line, water level variability process line, electric winch drive current process line and flow velocity distribution skewness coefficient of the same period in previous years are read from the historical database of the hydrological center station to which the cross section belongs. The historical data of the same period in previous years are back-calculated according to the calculation method of the cross-section comprehensive working condition anomaly parameter to obtain the historical cross-section comprehensive working condition anomaly parameter sequence of the cross-section in the current calendar period. The anomaly parameter sequence of the comprehensive working conditions of the historical cross-sections is statistically sorted, and the Pth value is taken. rotor Percentile is used as the rotor mode preservation threshold, and is taken as the Pth percentile. radar Percentile is used as the radar wave mode trigger threshold, where P rotor <P radar ;P rotor and P radar The value of P is set according to the season, with the value used during the main flood season. rotor and P radar The P values ​​were lower than those used during the dry season. rotor and P radar .

5. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 4, characterized in that, In step S2, a three-zone switching decision is made based on the comparison results, including the rotor current meter mode, the dual-mode parallel transition mode, and the radar wave current meter mode. This includes the following steps: When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the rotor flow meter mode is adopted for decision-making. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the decision is made to adopt the radar wave current meter mode and send a lifting command to the electric winch to retrieve the underwater lead fish and rotor current meter to a safe height. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the decision adopts a dual-mode parallel transition mode, and the rotor flow meter and the radar wave flow meter are started simultaneously for parallel acquisition, forming a continuous fuzzy transition interval between the rotor mode holding threshold and the radar wave mode triggering threshold.

6. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 1, characterized in that, In step S3, the construction of the flow measurement accuracy benefit function, the operation safety risk function, and the flow measurement timeliness function includes the following steps: The flow measurement accuracy benefit function is constructed as follows: , Where, n v Let n be the number of perpendicular lines. p N represents the number of measurement point layers. max λ1 is the maximum number of perpendicular lines required by the cross-section specification, γ is the accuracy saturation coefficient, η is the vertical resolution sensitivity index, and η is the radar wave mode accuracy loss coefficient. The value is 1 when the three-zone switching decision is in radar wave current meter mode, and 0 otherwise. The operational safety risk function is constructed as follows: , Wherein, CSI(t) is the anomaly parameter of the comprehensive working condition of the cross section, and τ is the vehicle movement risk coefficient. The underwater retention risk coefficient, The value is 0 when the three-zone switching decision is in radar wave current meter mode, and 1 otherwise. The flow measurement time-effect function is constructed as follows: , Among them, t move The average travel time between adjacent perpendicular lines is t. point The average sampling time for flow velocity at a single measuring point is given by L, where L is the current water surface width at the cross-section, and v is the average sampling time at a single measuring point. critical The characteristic flow velocity for the propagation of the flood peak at the cross section is given.

7. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 6, characterized in that, In step S3, a Pareto optimal solution set, including the number of perpendicular lines and the number of measuring point layers, is solved in the three-dimensional target space composed of three functions; and based on the comparison results of the cross-sectional comprehensive working condition anomaly degree parameter, the rotor mode holding threshold, and the radar wave mode triggering threshold, the optimal flow measurement scheme composed of the optimal number of perpendicular lines and the optimal number of measuring point layers is selected from the Pareto optimal solution set, including the following steps: Iterate through the number of perpendicular lines n v With the number of measurement point layers n p For all possible combinations of values, the three constructed functions are used to calculate the flow measurement accuracy benefit function value, operational safety risk function value, and flow measurement timeliness function value corresponding to each value combination. By comparing each pair, all solutions that satisfy the Pareto optimal condition are selected from all value combinations as the Pareto optimal solution set. The Pareto optimal condition means that for any value combination in the Pareto optimal solution set, there is no other value combination that satisfies that its flow measurement accuracy benefit function value is not lower than that value combination, its operational safety risk function value is not higher than that value combination, its flow measurement timeliness function value is not higher than that value combination, and it is better than that value combination in at least one dimension. When the cross-sectional comprehensive working condition anomaly parameter does not exceed the rotor mode holding threshold, the scheme that maximizes the flow measurement accuracy benefit function value is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is between the rotor mode holding threshold and the radar wave mode triggering threshold, the scheme that maximizes S(x,t)·(1-R(x,t)) is selected from the Pareto optimal scheme set. When the cross-sectional comprehensive working condition anomaly parameter is not less than the radar wave mode trigger threshold, the number of measuring point layers n is set. p =1, given the total number of perpendicular lines n v From the Pareto optimal solution set composed of the given values, select the solution that satisfies R(x,t)≤R max The schemes are selected, and the scheme that minimizes the value of the flow measurement time-effect function is chosen, where R max This is the preset maximum acceptable risk threshold.

8. The automatic flow measurement and control method for hydrological cross-sections based on multi-mode adaptive switching according to claim 1, characterized in that, In step S4, the weighting coefficient includes the following steps: Calculate the absolute value of the difference between the radar wave mode trigger threshold and the cross-section comprehensive working condition anomaly parameter, calculate the absolute value of the difference between the radar wave mode trigger threshold and the rotor mode holding threshold, and use the ratio of the former to the latter as a weighting coefficient. When the cross-sectional comprehensive working condition anomaly parameter continuously changes from the rotor mode holding threshold to the radar wave mode triggering threshold within the transition range, the weighting coefficient continuously changes from 1 to 0.

Images