Root stone non-contact underwater detection method for flood control engineering safety

Abstract

The invention discloses a root stone non-contact underwater detection method for flood control engineering safety, and relates to the field of non-contact underwater detection, and the method comprises the steps: building a three-dimensional dynamic flow field prediction model based on fluid mechanics characteristics, predicting the speed fluctuation component of turbulence in a detection region through the flow field prediction model, and evaluating the expected interference degree of the turbulence; performing iterative optimization on the candidate position by using a genetic algorithm based on an expected interference degree evaluation result of turbulence, and selecting an optimal operation area of the detection equipment; based on the optimal operation area selection result of the detection equipment, planning the advancing route of the detection equipment, and adaptively adjusting the counter weight of the detector and the rotating speed of a propeller according to the real-time flow field information of the advancing route; for false echoes caused by a multipath effect in an underwater detection process, a multipath signal denoising algorithm is used for processing after a signal is received, and dynamic compensation is carried out on an echo signal.

Description

Technical Field

[0001] The invention relates to the field of non-contact underwater detection, and in particular to a non-contact underwater root stone detection method for flood control engineering safety. Background Art

[0002] In flood control projects, root stones are exposed to the water flow of rivers or floods for a long time, and are easily displaced or sunk by the scouring and erosion of water flow. Traditional root stone detection methods rely on sonar or laser scanning technology, but in practical applications, the turbulent effect of water flow in the water flow environment often has a significant impact on the detection accuracy. In a complex water flow environment, traditional sonar or laser detection methods are often affected by water turbulence, resulting in signal reflection angle deviation, multipath interference, and signal frequency distortion.

[0003] This solution uses a dynamic water flow prediction model in conjunction with an automatic detection position adjustment mechanism to optimize the layout of the detection equipment in real time. By adjusting the center of gravity of the detector and the propeller speed to balance the impact of the water flow, the route can be controlled, and the signal propagation path can avoid high turbulence areas as much as possible, reducing the direct interference of the water flow on the signal. At the same time, the intelligent signal compensation mechanism ensures that the intensity, frequency and phase of the signal in different environments can be effectively adjusted through intelligent compensation of the detection signal, further improving the anti-interference ability and data stability of the root stone detection system. This technical solution not only improves the accuracy of underwater detection, but also reduces the risk of misjudgment and missed detection, which is of great significance to the safety of flood control projects. Summary of the invention

[0004] In order to solve the above technical problems, a non-contact underwater detection method for root stones used for flood control project safety is provided. This technical solution solves the problems raised in the above background technology.

[0005] In order to achieve the above purpose, the technical solution adopted by the present invention is:

[0006] A non-contact underwater detection method for root stones used for flood control engineering safety, comprising:

[0007] A three-dimensional dynamic flow field prediction model is established based on fluid mechanics characteristics. The velocity fluctuation component of turbulence in the detection area is predicted through the flow field prediction model to evaluate the expected interference degree of turbulence.

[0008] Based on the evaluation results of the expected interference degree of turbulence, the candidate positions are iteratively optimized using genetic algorithms to select the optimal operating area of ​​the detection equipment;

[0009] Based on the optimal operating area selection results of the detection equipment, the detection equipment's route is planned, and the detector's counterweight and propeller speed are adaptively adjusted according to the real-time flow field information of the route;

[0010] In order to deal with the false echoes caused by the multipath effect during underwater detection, a multipath signal denoising algorithm is used to process the received signal and dynamically compensate the echo signal.

[0011] Preferably, the three-dimensional dynamic flow field prediction model is established based on the fluid mechanics characteristics, and the velocity fluctuation component of the turbulence in the detection area is predicted by the flow field prediction model, and the expected interference degree of the turbulence is evaluated, which specifically includes:

[0012] Array sensor nodes are set up in the detection area. After the sensor nodes are calibrated for timing and data consistency through a distributed synchronization mechanism based on a data aggregation algorithm, different water flow data are independently collected and transmitted in real time.

[0013] Based on the multi-level Kalman filter structure, the water flow fluctuation pattern is refined and classified to capture the water flow micro-changes and complex fluctuation patterns;

[0014] The flow field prediction model is established by combining the velocity field and pressure field data of the water flow in the detection area with the computational fluid dynamics algorithm to perform three-dimensional modeling and numerical simulation.

[0015] Input the real-time collected water flow data into the flow field prediction model to obtain the three-dimensional distribution information of turbulent flow direction and intensity;

[0016] The detection area is divided into three-dimensional grid space, and the flow field prediction model is used to predict the velocity fluctuation component of the turbulence in the detection area grid in real time, and the expected turbulence interference coefficient of the grid area is calculated. The calculation formula of the expected turbulence interference coefficient is:

[0017]

[0018] Where, T c is the expected turbulence interference coefficient, is the turbulent kinetic energy, The gradient of turbulent kinetic energy in the x, y, and z directions, λ, μ, and γ are the directional weight coefficients in the x, y, and z directions respectively.

[0019] Preferably, the method of iteratively optimizing candidate positions using a genetic algorithm based on the expected interference degree evaluation result of turbulence to select the optimal operating area of ​​the detection equipment specifically includes:

[0020] A fitness function is constructed in the three-dimensional grid space of the detection area based on the expected turbulence interference coefficient to evaluate the detection fitness of each position;

[0021] A set number of candidate detection positions are randomly and uniformly generated in the three-dimensional space of the detection area as initial detection positions.

[0022] Calculate the fitness value of each candidate detection position in the initial detection position and sort them, take the first n candidate positions with the highest fitness as the parent candidates, and mark the i-th candidate position as P i , P i =(x i ,y i ,z i );

[0023] Select the coordinate pairs from the parent candidate and generate new positions using the dynamic turbulence offset intersection method, specifically:

[0024] Based on the prediction results of the flow field prediction model, the turbulence intensity gradient in the area where the parent candidate position is located is calculated According to the opposite direction of the turbulence gradient Perform offset, set offset distance Δd, and calculate the offset position P i ′, in Represent the components of the turbulence gradient in the x, y, and z directions respectively;

[0025] On the parent position adjusted by turbulence offset, a crossover operation is performed to generate a new position by randomly exchanging coordinates;

[0026] Genetic algorithms are used to continuously iterate and optimize candidate positions, and the candidate position with the highest fitness is selected as the optimal operating point for the detection equipment.

[0027] Preferably, the method of planning the travel route of the detection device based on the optimal operation area selection result of the detection device, and adaptively adjusting the detector counterweight and propeller speed according to the real-time flow field information of the travel route specifically includes:

[0028] The three-dimensional grid of the detection area is used as the path point, and the Dijkstra algorithm is used to calculate the fitness point by point to plan the optimal path;

[0029] The detection device monitors the turbulence intensity in real time during the movement. When the detection device moves to the next path point, the current center of gravity height is automatically adjusted through the counterweight steel ball in the device according to the directional component of the turbulence. The calculation formula for the center of gravity height adjustment is:

[0030]

[0031] In the formula, Δh cg is the height adjustment of the center of gravity, α is the center of gravity adjustment coefficient, T x , T y , T z are the components of turbulence in the x, y, and z directions, respectively;

[0032] The real-time flow field data is analyzed by fast Fourier transform, the main frequency component ω of turbulence is extracted, the energy density of turbulence is calculated based on the turbulence intensity and frequency data, and the propeller speed is adaptively adjusted according to the turbulence energy density. The propeller speed adjustment amount is calculated as follows:

[0033]

[0034] Where ΔR is the propeller speed adjustment, δ is the speed adjustment coefficient, and T is the turbulence intensity.

[0035] Preferably, the false echo caused by the multipath effect in the underwater detection process is processed by using a multipath signal denoising algorithm after receiving the signal, and the dynamic compensation of the echo signal specifically includes:

[0036] Receive and analyze underwater detection signals, calculate the time difference of each signal path through the time difference arrival algorithm, mark the echo signal as a false signal or a main signal according to the time delay and signal strength, and identify multi-path false echoes;

[0037] De-noise the multipath signal marked as a false echo, retain the main signal path, and use an adaptive coherent denoising algorithm to suppress false signals by analyzing phase and intensity characteristics, so that the intensity fluctuation of the denoised signal is within the set range;

[0038] According to the denoised main signal, the Kalman filter compensation algorithm is used to dynamically compensate the phase and intensity of the signal. The ideal signal state is predicted based on historical data, and the actual signal is adjusted to the compensated signal so that the signal strength and phase are stable within the set range.

[0039] Compared with the prior art, the present invention has the following beneficial effects:

[0040] By establishing a three-dimensional dynamic flow field prediction model, the velocity fluctuation component of the turbulence in the detection area is predicted, the degree of turbulence interference is evaluated, and the operating position of the detection equipment is optimized by intelligent means, so as to make the best configuration in the high interference area. At the same time, the adaptive control strategy based on real-time flow field information can dynamically adjust the counterweight and propeller speed of the detector to adapt it to the complex water flow environment, thereby improving the stability and detection accuracy of the detector. In order to solve the problem of false echoes caused by underwater multipath effects, the present invention introduces a multipath signal denoising algorithm to dynamically compensate for the echo signal, thereby enhancing the authenticity and effectiveness of the detection data. Through the above means, the present invention can more accurately and effectively complete the underwater detection task of root stones, ensure the safety of flood control projects, and provide technical support for complex underwater operations. BRIEF DESCRIPTION OF THE DRAWINGS

[0041] Figure 1A flow chart of a non-contact underwater detection method for root stones used for flood control engineering safety according to the present invention;

[0042] Figure 2 A flow chart of establishing a three-dimensional dynamic flow field prediction model based on fluid mechanics characteristics, predicting the velocity fluctuation component of turbulence in the detection area through the flow field prediction model, and evaluating the expected interference degree of turbulence;

[0043] Figure 3 The flowchart of the present invention is to use a genetic algorithm to iteratively optimize the candidate positions and select the optimal operating area of ​​the detection equipment based on the expected interference degree evaluation result of turbulence;

[0044] Figure 4 The flow chart of the present invention for selecting the optimal operating area of ​​the detection device, planning the route of the detection device, and adaptively adjusting the detector counterweight and propeller speed according to the real-time flow field information of the route;

[0045] Figure 5 This is a flow chart of the present invention for dynamically compensating the echo signal by using a multipath signal denoising algorithm to process the false echo caused by the multipath effect during underwater detection. DETAILED DESCRIPTION

[0046] The following description is used to disclose the present invention so that those skilled in the art can implement the present invention. The preferred embodiments described below are only examples, and those skilled in the art may think of other obvious variations.

[0047] Reference Figure 1 As shown, a non-contact underwater detection method for root stones used for flood control engineering safety includes:

[0048] A three-dimensional dynamic flow field prediction model is established based on fluid mechanics characteristics. The velocity fluctuation component of turbulence in the detection area is predicted through the flow field prediction model to evaluate the expected interference degree of turbulence.

[0049] Based on the evaluation results of the expected interference degree of turbulence, the candidate positions are iteratively optimized using genetic algorithms to select the optimal operating area of ​​the detection equipment;

[0050] Based on the optimal operating area selection results of the detection equipment, the detection equipment's route is planned, and the detector's counterweight and propeller speed are adaptively adjusted according to the real-time flow field information of the route;

[0051] In order to deal with the false echoes caused by the multipath effect during underwater detection, a multipath signal denoising algorithm is used to process the received signal and dynamically compensate the echo signal.

[0052] Reference Figure 2As shown in the figure, a three-dimensional dynamic flow field prediction model is established based on the fluid mechanics characteristics. The velocity fluctuation component of the turbulence in the detection area is predicted by the flow field prediction model, and the expected interference degree of the turbulence is evaluated, which specifically includes:

[0053] Array sensor nodes are set up in the detection area. After the sensor nodes are calibrated for timing and data consistency based on the data aggregation algorithm through a distributed synchronization mechanism, different water flow data are independently collected and transmitted in real time. Through the distributed synchronization mechanism, each node synchronously collects data and aligns the timing to eliminate data delays between sensor nodes. The network time protocol (NTP) or distributed clock synchronization algorithm is used to keep the acquisition timing error between nodes within microseconds, thereby ensuring data consistency.

[0054] Based on the multi-level Kalman filter structure, the water flow fluctuation patterns are refined and classified to capture the micro-changes in water flow and complex fluctuation patterns. The primary Kalman filter is mainly used to eliminate noise and large fluctuations; the intermediate Kalman filter is used to capture the regular fluctuation characteristics of water flow; the advanced Kalman filter is used to detect small changes in water flow and further distinguish complex fluctuation patterns.

[0055] After Kalman filtering, characteristic parameters such as fluctuation frequency, phase, amplitude, etc. are extracted and counted to establish a detailed water flow fluctuation characteristic database. This process can be achieved through spectrum analysis tools (such as fast Fourier transform) to identify the unique characteristics of various fluctuation patterns.

[0056] Three-dimensional modeling and numerical simulation are performed by combining the velocity field and pressure field data of the water flow in the detection area with the computational fluid dynamics algorithm. A flow field prediction model is established by jointly processing the velocity field and the pressure field using the computational fluid dynamics (CFD) algorithm.

[0057] The real-time collected water flow data is input into the flow field prediction model to obtain the three-dimensional distribution information of turbulent flow direction and intensity. Specifically, based on the CFD modeling results, the flow field in the detection area is numerically simulated to simulate the complex fluctuations and turbulence phenomena of the water flow. The spatial distribution characteristics of the flow field at different time steps are obtained by solving the velocity field and pressure field equations through multiple iterations.

[0058] The entire area is divided into several subdivided units (such as cubic units), and the data inside each unit can be approximately considered uniform. The accuracy of the grid division is adjusted according to the complexity of the turbulence changes, and a finer grid division is used in dense turbulence areas. The flow field prediction model is used to predict the velocity fluctuation component of the turbulence in the detection area grid in real time, and the expected turbulence interference coefficient of the grid area is calculated. The expected turbulence interference coefficient calculation formula is:

[0059]

[0060] Where, T c is the expected turbulence interference coefficient, is the turbulent kinetic energy, The gradient of turbulent kinetic energy in the x, y, and z directions, λ, μ, and γ are the directional weight coefficients in the x, y, and z directions respectively.

[0061] The calculation results of the flow field prediction model are mapped to each grid unit to obtain the velocity fluctuation and pressure change inside each unit. The model predicts the turbulent velocity fluctuation component in each grid unit in real time for detailed analysis of turbulent characteristics.

[0062] Reference Figure 3 As shown in the figure, based on the evaluation results of the expected interference degree of turbulence, the candidate positions are iteratively optimized using a genetic algorithm, and the optimal operating area of ​​the detection equipment is selected, including:

[0063] Based on the expected turbulence interference coefficient, a fitness function is constructed in the three-dimensional grid space of the detection area to evaluate the detection fitness of each position. The fitness function is used to evaluate the detection adaptability of each grid unit, specifically:

[0064]

[0065] A set number of candidate detection positions are randomly and uniformly generated in the three-dimensional space of the detection area as initial detection positions. The number of randomly generated positions is set according to the size of the area to ensure that each potential detection point is evenly covered in the area.

[0066] For example, 100 candidate positions are randomly generated in the three-dimensional space of the detection area as the initial population. The distribution of these positions is set according to the area size, turbulence distribution density and water flow fluctuation characteristics, and their number meets the population size requirements of the genetic algorithm to provide sufficient search space in the initial iteration.

[0067] Calculate the fitness value of each candidate detection position in the initial detection position and sort them, take the first n candidate positions with the highest fitness as the parent candidates, and mark the i-th candidate position as P i , P i =(x i ,y i ,z i );

[0068] The system screens based on fitness values ​​and retains candidate positions with higher fitness values ​​as parent candidates, usually selecting the top 20 to ensure the transmission of excellent genes.

[0069] Select the coordinate pairs from the parent candidate and generate new positions using the dynamic turbulence offset intersection method, specifically:

[0070] Based on the prediction results of the flow field prediction model, the turbulence intensity gradient in the area where the parent candidate position is located is calculated According to the opposite direction of the turbulence gradient Perform offset, set offset distance Δd, and calculate the offset position P i ′, in Represent the components of the turbulence gradient in the x, y, and z directions respectively;

[0071] On the parent position adjusted by turbulence offset, a crossover operation is performed to generate a new position by randomly exchanging coordinates;

[0072] Genetic algorithms are used to continuously iterate and optimize candidate locations, and the candidate locations with the highest fitness are selected as the optimal operating points for the detection equipment. After each round of crossover and mutation operations, the system recalculates the fitness values ​​of the newly generated candidate locations and gradually updates the population until the fitness values ​​converge or reach the preset upper limit of 100 iterations.

[0073] After completing multiple rounds of iterative optimization, the system selects the candidate position with the highest fitness value as the optimal layout point for the detection device, which has the highest adaptability in terms of signal strength and turbulence interference. The system moves the detection device to this position and remeasures the signal strength and turbulence intensity at this position. If the remeasured signal strength and turbulence interference meet the expected standards of the fitness function, the system will fix the device at this position; otherwise, restart the iterative optimization until the final position that meets the signal stability and quality requirements is found.

[0074] Reference Figure 4 As shown, based on the optimal operating area selection result of the detection equipment, the detection equipment's route is planned, and according to the real-time flow field information of the route, the detector's counterweight and propeller speed are adaptively adjusted. Specifically, the following are included:

[0075] After the detection device selects the optimal operating area, the system needs to plan the optimal route from the current device position to the target area. Due to the complex changes in turbulence in the underwater environment, the planned route must not only ensure that it reaches the target area, but also avoid the high-influence area of ​​turbulence to ensure the stability of the equipment and the quality of signal transmission. During the path planning process, the system inputs the flow field data into the path calculation module in real time, gradually optimizes the path points, and ensures that the adaptability of each path point meets the requirements of the detection task.

[0076] The three-dimensional grid of the detection area is used as the path point. Each path point records the turbulence intensity, signal intensity, pressure field change and other information in the flow field. The Dijkstra algorithm is used to calculate the fitness planning optimal path point by point. Without the need for heuristic estimation, the minimum cost path from the starting point to the target area is calculated. The algorithm traverses all the path points in the three-dimensional grid point by point, calculates the cumulative cost from the starting point to each path point in turn, and selects the path with the minimum cost for expansion. Each time the path is expanded, the system collects the real-time turbulence data of the current path point through the flow field sensor, calculates the updated turbulence interference coefficient, and adjusts the fitness value of the path point in real time to ensure that the path planning can reflect the latest flow field changes.

[0077] The detection device monitors the turbulence intensity in real time during the movement. When the detection device moves to the next path point, the current center of gravity height is automatically adjusted through the counterweight steel ball in the device according to the directional component of the turbulence. The calculation formula for the center of gravity height adjustment is:

[0078]

[0079] In the formula, Δh cg is the height adjustment of the center of gravity, α is the center of gravity adjustment coefficient, T x , T y , T z are the components of turbulence in the x, y, and z directions, respectively;

[0080] After the counterweight steel ball is adjusted, the attitude sensor in the device (such as accelerometer, gyroscope) monitors the pitch angle and roll angle of the device in real time to ensure the effectiveness of the center of gravity adjustment and prevent the device from becoming unstable or tilting. For example: if the pitch angle or roll angle deviation of the device exceeds the preset threshold (such as ±3 degrees), the system regards the deviation as a signal of insufficient stability and triggers further adjustment.

[0081] As the device continues to move, the system periodically re-evaluates the direction and intensity of the turbulence. If the turbulence intensity changes significantly (such as an increase of more than 10%), the system will re-trigger the adjustment process of the counterweight steel balls and optimize the center of gravity position in real time.

[0082] The real-time flow field data is analyzed by fast Fourier transform, the main frequency component ω of turbulence is extracted, the energy density of turbulence is calculated based on the turbulence intensity and frequency data, and the propeller speed is adaptively adjusted according to the turbulence energy density. The propeller speed adjustment amount is calculated as follows:

[0083]

[0084] Where ΔR is the propeller speed adjustment, δ is the speed adjustment coefficient, and T is the turbulence intensity.

[0085] Fast Fourier transform is used to extract the main frequency of turbulence, and the energy density is calculated in combination with the turbulence intensity. The propeller speed is adaptively adjusted in real time based on the energy density to ensure the stability and efficiency of the equipment in turbulent environments. This dynamic adjustment mechanism improves the adaptability and attitude control capabilities of the equipment in complex underwater environments.

[0086] Reference Figure 5 As shown in the figure, for the false echo caused by the multipath effect in the underwater detection process, a multipath signal denoising algorithm is used to process the received signal, and the dynamic compensation of the echo signal specifically includes:

[0087] During underwater detection, the receiver collects underwater reflected signals. These signals are reflected through different paths in the water body and reach the receiving device, forming multipath echoes, including the main signal path and the reflected signals of multiple false paths. Due to the different propagation distances of each path, the multipath echoes have time delays and different intensities.

[0088] Receive and analyze underwater detection signals, calculate the time difference of each signal path through the time difference arrival algorithm, mark the echo signal as a false signal or a main signal according to the time delay and signal strength, and identify multi-path false echoes. For example: the system sets a time difference threshold, and if the time difference of a path signal exceeds the threshold (such as 5 milliseconds), the path signal is marked as a potential false signal.

[0089] The multipath signal marked as a false echo is denoised, the main signal path is retained, and the adaptive coherent denoising algorithm is used to suppress false signals by analyzing the phase and intensity characteristics. The system detects the signal strength fluctuation after denoising. If the intensity fluctuation of the denoised signal is lower than the set threshold (such as 5%), the denoising is considered to be completed; otherwise, the system repeats the denoising process based on the feedback, and ultimately ensures that the intensity fluctuation of the denoised signal is within the set range.

[0090] According to the main signal after denoising, the Kalman filter compensation algorithm is used to dynamically compensate the phase and strength of the signal. The ideal signal state is predicted based on historical data, and the actual signal is adjusted to the compensated signal to stabilize the signal strength within the set range. When the frequency fluctuation causes the phase change, the system adjusts the signal phase according to the predicted phase deviation so that the compensated signal remains synchronized, avoiding signal distortion caused by phase deviation.

[0091] The TDOA algorithm is used to identify multipath false echoes, and the adaptive coherent denoising algorithm is used to suppress false signals. The Kalman filter algorithm is then used to dynamically compensate the denoised main signal, thereby maintaining the signal strength and phase stability in complex underwater environments.

[0092] Furthermore, the present solution also proposes a storage medium for a non-contact underwater detection method for root stones for the safety of flood control projects, on which a computer-readable program is stored. When the computer-readable program is called, it executes the above-mentioned non-contact underwater detection method for root stones for the safety of flood control projects.

[0093] It is understandable that the storage medium may be a magnetic medium, such as a floppy disk, a hard disk, or a magnetic tape; an optical medium, such as a DVD; or a semiconductor medium, such as a solid state drive (SSD).

[0094] The above shows and describes the basic principles, main features and advantages of the present invention. Those skilled in the art should understand that the present invention is not limited to the above embodiments. The above embodiments and descriptions only describe the principles of the present invention. The present invention may be subject to various changes and improvements without departing from the spirit and scope of the present invention. These changes and improvements fall within the scope of the present invention. The scope of protection claimed by the present invention is defined by the attached claims and their equivalents.

Claims

1. A non-contact underwater detection method for root stones used for flood control engineering safety, characterized in that: include: A three-dimensional dynamic flow field prediction model is established based on fluid mechanics characteristics. The velocity fluctuation component of turbulence in the detection area is predicted through the flow field prediction model to evaluate the expected interference degree of turbulence. Based on the evaluation results of the expected interference degree of turbulence, the candidate positions are iteratively optimized using genetic algorithms to select the optimal operating area of ​​the detection equipment; Based on the optimal operating area selection results of the detection equipment, the detection equipment's route is planned, and the detector's counterweight and propeller speed are adaptively adjusted according to the real-time flow field information of the route; In order to deal with the false echoes caused by the multipath effect during underwater detection, a multipath signal denoising algorithm is used to process the received signal and dynamically compensate the echo signal.

2. A non-contact underwater detection method for root stones used for flood control engineering safety according to claim 1, characterized in that: The three-dimensional dynamic flow field prediction model is established based on the fluid mechanics characteristics, and the velocity fluctuation component of the turbulence in the detection area is predicted by the flow field prediction model, and the expected interference degree of the turbulence is evaluated, which specifically includes: Array sensor nodes are set up in the detection area. After the sensor nodes are calibrated for timing and data consistency through a distributed synchronization mechanism based on a data aggregation algorithm, different water flow data are independently collected and transmitted in real time. Based on the multi-level Kalman filter structure, the water flow fluctuation pattern is refined and classified to capture the micro-changes and complex fluctuation patterns of the water flow; The flow field prediction model is established by combining the velocity field and pressure field data of the water flow in the detection area with the computational fluid dynamics algorithm to perform three-dimensional modeling and numerical simulation. Input the real-time collected water flow data into the flow field prediction model to obtain the three-dimensional distribution information of turbulent flow direction and intensity; The detection area is divided into three-dimensional grid space, and the flow field prediction model is used to predict the velocity fluctuation component of the turbulence in the detection area grid in real time, and the expected turbulence interference coefficient of the grid area is calculated. The calculation formula of the expected turbulence interference coefficient is: Where, T c is the expected turbulence interference coefficient, is the turbulent kinetic energy, The gradient of turbulent kinetic energy in the x, y, and z directions, λ, μ, and γ are the directional weight coefficients in the x, y, and z directions respectively.

3. A non-contact underwater detection method for root stones used for flood control engineering safety according to claim 2, characterized in that: The method of iteratively optimizing candidate positions using a genetic algorithm based on the expected interference degree evaluation result of turbulence to select the optimal operating area of ​​the detection equipment specifically includes: A fitness function is constructed in the three-dimensional grid space of the detection area based on the expected turbulence interference coefficient to evaluate the detection fitness of each position; A set number of candidate detection positions are randomly and uniformly generated in the three-dimensional space of the detection area as initial detection positions. Calculate the fitness value of each candidate detection position in the initial detection position and sort them, take the first n candidate positions with the highest fitness as the parent candidates, and mark the i-th candidate position as P i , P i =(x i ,y i ,z i ); Select the coordinate pairs from the parent candidate and generate new positions using the dynamic turbulence offset intersection method, specifically: Based on the prediction results of the flow field prediction model, the turbulence intensity gradient in the area where the parent candidate position is located is calculated According to the opposite direction of the turbulence gradient Perform offset, set offset distance Δd, and calculate the offset position P i ′, in Represent the components of the turbulence gradient in the x, y, and z directions respectively; On the parent position adjusted by turbulence offset, a crossover operation is performed to generate a new position by randomly exchanging coordinates; Genetic algorithms are used to continuously iterate and optimize candidate positions, and the candidate position with the highest fitness is selected as the optimal operating point for the detection equipment.

4. The non-contact underwater detection method for root stones used for flood control engineering safety according to claim 3 is characterized in that: The method of selecting the optimal operation area of ​​the detection device based on the detection device, planning the detection device's route, and adaptively adjusting the detector's counterweight and propeller speed according to the real-time flow field information of the route specifically includes: The three-dimensional grid of the detection area is used as the path point, and the Dijkstra algorithm is used to calculate the fitness point by point to plan the optimal path; The detection device monitors the turbulence intensity in real time during the movement. When the detection device moves to the next path point, the current center of gravity height is automatically adjusted through the counterweight steel ball in the device according to the directional component of the turbulence. The calculation formula for the center of gravity height adjustment is: In the formula, Δh cg is the height adjustment of the center of gravity, α is the center of gravity adjustment coefficient, T x 、T y 、T z are the components of turbulence in the x, y, and z directions, respectively; The real-time flow field data is analyzed by fast Fourier transform, the main frequency component ω of turbulence is extracted, the energy density of turbulence is calculated based on the turbulence intensity and frequency data, and the propeller speed is adaptively adjusted according to the turbulence energy density. The propeller speed adjustment amount is calculated as follows: Where ΔR is the propeller speed adjustment, δ is the speed adjustment coefficient, and T is the turbulence intensity.

5. A non-contact underwater detection method for root stones used for flood control engineering safety according to claim 4, characterized in that: The method of dynamically compensating the false echo caused by the multipath effect during underwater detection by using a multipath signal denoising algorithm to process the received signal specifically includes: Receive and analyze underwater detection signals, calculate the time difference of each signal path through the time difference arrival algorithm, mark the echo signal as a false signal or a main signal according to the time delay and signal strength, and identify multi-path false echoes; De-noise the multipath signal marked as a false echo, retain the main signal path, and use an adaptive coherent denoising algorithm to suppress false signals by analyzing phase and intensity characteristics, so that the intensity fluctuation of the denoised signal is within the set range; According to the denoised main signal, the Kalman filter compensation algorithm is used to dynamically compensate the phase and intensity of the signal. The ideal signal state is predicted based on historical data, and the actual signal is adjusted to the compensated signal so that the signal strength and phase are stable within the set range.

Images