A reservoir underwater siltation dynamic monitoring system and method

Abstract

The application relates to the technical field of reservoir siltation monitoring, and discloses a reservoir underwater siltation dynamic monitoring system and method. The method comprises a layered intelligent sensor array of a pressure sensor, an acoustic Doppler current profile sensor and a fiber bragg grating strain sensor, and a point-line-surface three-dimensional monitoring network is formed; a plurality of source monitoring data such as pressure data, flow velocity data and strain data output by the array are collected in real time through an edge computing node; the plurality of source monitoring data are subjected to real-time fusion processing to generate fusion data; based on a water flow dynamics model and a sediment settlement model, dynamic inversion calculation is carried out on the fusion data to obtain siltation evolution trend data; an adaptive calibration algorithm is used to calibrate the siltation evolution trend data, sensor drift and water body interference are eliminated, and finally, calibrated siltation state monitoring results are generated. The method can realize comprehensive and real-time monitoring of reservoir underwater siltation, and provide accurate state information for reservoir siltation management.

Description

A dynamic monitoring system and method for underwater sedimentation in reservoirs Technical Field

[0001] This invention relates to the field of reservoir siltation monitoring technology, specifically to a dynamic monitoring system and method for underwater siltation in reservoirs. Background Technology

[0002] As a crucial component of water conservancy systems, reservoirs perform a wide range of functions, including flood control and water storage, agricultural irrigation, hydropower generation, and urban and rural water supply. Their operational status directly impacts regional water resource allocation, ecological stability, and socio-economic development. Over long-term operation, silt carried by upstream rivers enters the reservoir with the flow. Due to factors such as slower water flow and gravity, the silt gradually settles at the bottom of the reservoir, forming underwater sediment. As sediment accumulates, the effective storage capacity of the reservoir gradually decreases, leading to a decline in its flood control, water storage, and power generation efficiency. Simultaneously, accumulated silt may cover critical structures at the bottom of the reservoir, such as intakes and spillways, affecting their normal operation and even increasing equipment wear and maintenance costs. Furthermore, uneven silt distribution can alter the water flow pattern within the reservoir, causing localized turbulence and potentially impacting the aquatic ecosystem of the reservoir area.

[0003] Currently, underwater sedimentation monitoring in reservoirs mainly relies on traditional methods, which have significant limitations in practical applications. Manual sampling and analysis is a relatively basic monitoring method. Workers need to use boats equipped with sampling devices to collect sediment samples at different points in the reservoir area, and then analyze them in the laboratory to determine the thickness and composition of the sediment. This method not only requires a large investment of manpower and time, but also has limited sampling points, making it difficult to comprehensively reflect the overall distribution of sedimentation in the reservoir area. Furthermore, the sampling process is greatly affected by natural conditions such as weather and water flow, making continuous monitoring impossible. While acoustic monitoring can obtain information about the sedimentation interface through sound wave reflection, its accuracy is easily affected by factors such as suspended particles and air bubbles in the water, especially in areas with greater water depth or complex sediment composition, making data reliability difficult to guarantee.

[0004] In some monitoring scenarios, a single type of sensor is used for data acquisition. For example, pressure sensors may be used to monitor water level changes to indirectly infer siltation, or flow velocity sensors may be used to acquire water flow data. These methods cannot simultaneously acquire multi-dimensional information related to siltation. The data is limited in scope, making it difficult to establish a correlation between siltation evolution and water flow and sediment movement, and thus failing to accurately reflect the dynamic changes in siltation. While remote sensing technology can achieve large-scale monitoring, its resolution limitations result in insufficient accuracy in monitoring near-shore areas of reservoirs and areas with complex reservoir bottom topography. Furthermore, remote sensing data acquisition involves time intervals, which cannot meet the need for real-time monitoring of siltation changes. Subsequent data processing also relies on complex algorithms, placing high demands on both hardware and software.

[0005] With the increasing demand for refined and dynamic monitoring in reservoir management, the shortcomings of traditional monitoring methods in terms of coverage, real-time performance, data integrity, and accuracy have become increasingly apparent. These methods are insufficient to support accurate judgment of the evolution trend of underwater sedimentation in reservoirs. A technical solution that can achieve comprehensive, real-time, and accurate monitoring is needed to meet the needs of long-term scientific management of reservoirs. Summary of the Invention

[0006] The purpose of this invention is to provide a dynamic monitoring system and method for underwater sedimentation in reservoirs to solve the problems mentioned in the background art.

[0007] To achieve the above objectives, the present invention provides a method for dynamic monitoring of underwater sedimentation in reservoirs, the method comprising:

[0008] A hierarchical intelligent sensor array is deployed, which includes a pressure sensor, an acoustic Doppler flow profile sensor, and a fiber optic strain sensor to form a three-dimensional monitoring network of points, lines, and surfaces.

[0009] The multi-source monitoring data of the hierarchical intelligent sensor array is collected in real time through edge computing nodes. The multi-source monitoring data includes pressure data, flow velocity data, and strain data.

[0010] The multi-source monitoring data is fused in real time to generate fused data;

[0011] Based on the hydrodynamic model and sediment deposition model, the fused data is dynamically inverted to obtain sedimentation evolution trend data.

[0012] An adaptive calibration algorithm is used to calibrate the sedimentation evolution trend data, eliminating the influence of sensor drift and water body interference, and generating calibrated sedimentation status monitoring results.

[0013] Preferably, the deployment of the hierarchical smart sensor array includes:

[0014] Based on the topographic features of the reservoir, the monitoring area is divided into a surface monitoring area, a middle monitoring area, and a bottom monitoring area.

[0015] The pressure sensor is deployed in the surface monitoring area to collect surface sediment pressure data;

[0016] The acoustic Doppler velocity profile sensor is deployed in the middle layer monitoring area to collect middle layer velocity distribution data;

[0017] The fiber optic strain sensor is deployed in the underlying monitoring area to collect underlying sediment strain data.

[0018] Preferably, the real-time fusion processing of the multi-source monitoring data includes:

[0019] The surface sedimentation pressure data is normalized to generate standardized pressure data;

[0020] The mid-layer velocity distribution data is spatiotemporally aligned to generate aligned velocity data;

[0021] The underlying sedimentation strain data is filtered and denoised to generate denoised strain data;

[0022] The standardized pressure data, the aligned flow velocity data, and the denoised strain data are fused at the feature level to generate the fused data.

[0023] Preferably, the dynamic inversion calculation of the fused data based on the hydrodynamic model and sediment deposition model includes:

[0024] Based on the fused data, flow velocity features, pressure features, and strain features are extracted;

[0025] The flow velocity characteristics are input into the hydrodynamic model to calculate the hydrodynamic distribution data.

[0026] The pressure characteristics and strain characteristics are input into the sediment settling model to calculate the sediment settling rate data.

[0027] The sedimentation evolution trend data is generated by combining the hydrodynamic distribution data and the sediment settling rate data.

[0028] Preferably, the calibration of the siltation evolution trend data using an adaptive calibration algorithm includes:

[0029] Acquire historical siltation monitoring data and establish a sensor drift compensation parameter library;

[0030] Based on the sensor drift compensation parameter library, drift compensation is performed on the siltation evolution trend data to generate preliminary calibration data;

[0031] The preliminary calibration data is corrected for interference based on the water turbidity data to generate the calibrated sedimentation status monitoring results.

[0032] Preferably, establishing the sensor drift compensation parameter library includes:

[0033] The reference data of the layered intelligent sensor array in a non-accumulation state is collected, the deviation value between the reference data and the real-time monitoring data is calculated, drift compensation parameters are generated, and the drift compensation parameters are stored in the sensor drift compensation parameter library in time series.

[0034] Preferably, the interference correction of the preliminary calibration data based on water turbidity data includes:

[0035] Real-time water turbidity data is collected using an optical sensor. An interference correction coefficient is determined based on the real-time water turbidity data. The interference correction coefficient is then used to perform weighted correction on the preliminary calibration data.

[0036] Preferably, after generating the calibrated siltation status monitoring results, the method further includes:

[0037] The calibrated siltation status monitoring results are stored in a siltation database in time series, and a siltation change trend report is generated based on the siltation database.

[0038] Preferably, generating a sedimentation change trend report based on the sedimentation database includes:

[0039] Extract the siltation status monitoring results from the siltation database for different time periods;

[0040] Calculate the rate of change of siltation thickness and the rate of expansion of siltation range in adjacent time periods;

[0041] A report on the trend of siltation change is generated based on the rate of change of siltation thickness and the rate of expansion of siltation range.

[0042] Preferably, the present invention also includes a dynamic monitoring system for underwater siltation in a reservoir, the system including a memory, a processor, and a computer program stored in the memory and running on the processor, wherein when the processor executes the computer program, it implements the steps of the dynamic monitoring method for underwater siltation in a reservoir as described above.

[0043] Compared with the prior art, the beneficial effects of the present invention are:

[0044] By deploying a hierarchical intelligent sensor array, combining pressure sensors, acoustic Doppler velocity profile sensors, and fiber optic strain sensors, a three-dimensional monitoring network is constructed along the starting line and surface. This multi-sensor collaborative monitoring architecture can simultaneously acquire underwater pressure, flow velocity, and strain data in the reservoir, breaking through the limitations of traditional single-sensor or localized monitoring. It achieves comprehensive capture of sedimentation-related parameters at different depths and in different areas of the reservoir, spatially covering various scenarios where sedimentation may occur in the reservoir. Whether in shallow or deep water areas near the shore, it can form a continuous and complete monitoring data chain, fully revealing the distribution characteristics and details of sedimentation changes.

[0045] The application of edge computing nodes provides crucial support for the real-time acquisition of multi-source monitoring data. Compared to the traditional monitoring model where data needs to be transmitted to a remote server for processing, edge computing nodes can directly acquire data at the monitoring site, significantly reducing data transmission distance and time loss, and avoiding monitoring delays caused by data transmission latency. Simultaneously, the real-time acquisition mode can capture instantaneous changes during the sedimentation process, such as fluctuations in sediment deposition caused by sudden changes in water flow velocity within a short period, or changes in sedimentation rate due to increased upstream sediment load after heavy rainfall, providing timely data support for understanding the dynamic evolution of sedimentation.

[0046] Real-time fusion processing of multi-source monitoring data can effectively integrate the monitoring advantages of different types of sensors and reduce the limitations and errors of single data. Pressure data can reflect changes in water level and sediment load, flow velocity data can reflect the driving effect of water flow on sediment movement, and strain data can help determine the stability of sediment layer structure. Through fusion processing, these data are no longer isolated information fragments, but form mutually corroborating and complementary fused data, which can more objectively reflect the actual state of sedimentation and reduce the impact of single data deviations on monitoring results.

[0047] Dynamic inversion calculations based on hydrodynamic and sediment deposition models provide a scientific technical approach for analyzing sedimentation evolution trends. Hydrodynamic models can simulate the flow patterns in reservoir areas, clarifying the correlation between flow velocity, direction, and sediment transport in different regions. Sediment deposition models, combined with sediment particle characteristics and aquatic environmental parameters, analyze sediment deposition rates and distribution characteristics. Combining these two models to perform inversion calculations on fused data allows for a mechanistic analysis of the sedimentation formation and development process, yielding sedimentation evolution trend data that reflects actual conditions, rather than relying solely on historical data for speculation. This makes judgments about future sedimentation changes more scientific and reasonable.

[0048] The application of adaptive calibration algorithms can effectively eliminate interference factors during the monitoring process. In long-term monitoring, sensors may drift due to factors such as water corrosion and temperature changes, leading to data deviations. Simultaneously, suspended particulate matter and water flow disturbances in the water can also interfere with the monitoring signal. Adaptive calibration algorithms can identify these interference factors in real time and dynamically adjust according to the changing patterns of the monitoring data, correcting deviations and ensuring that the final calibrated siltation monitoring results more closely reflect the actual situation. This improves the reliability of the monitoring data and provides a more accurate reference for reservoir siltation management decisions. Attached Figure Description

[0049] Figure 1 is a schematic diagram illustrating the working principle of the underwater sedimentation dynamic monitoring method for reservoirs described in this invention.

[0050] Figure 2 is a flowchart of deploying a hierarchical smart sensor array;

[0051] Figure 3 is a flowchart of the dynamic inversion calculation;

[0052] Figure 4 is a comprehensive map of reservoir siltation monitoring. Detailed Implementation

[0053] The technical solutions of the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0054] Referring to Figure 1, this invention provides a method for dynamic monitoring of underwater sedimentation in reservoirs. The method includes: deploying a layered intelligent sensor array in a predetermined area of ​​the reservoir. This array consists of pressure sensors, acoustic Doppler velocity profile sensors, and fiber optic strain sensors. These sensors are scientifically arranged to form a three-dimensional monitoring network covering the surface, middle, and bottom layers of the water body, ensuring the spatial comprehensiveness of the monitoring data. After deployment, multi-source monitoring data from the sensor array are collected in real time at a set sampling frequency through pre-set edge computing nodes. This data specifically includes pressure data reflecting sediment load, flow velocity data describing water flow, and strain data sensing changes in the bottom sediment. The collected raw multi-source monitoring data is transmitted to edge computing nodes or a central processing unit for real-time fusion processing. This process involves data alignment, standardization, and denoising, aiming to eliminate differences in spatiotemporal references and dimensions between different sensor data, generating a time-synchronized, format-uniform fused dataset. A mathematical model based on mature hydrodynamic principles and sediment deposition theory was employed, using fused data as input for dynamic inversion calculations. This process deduces the thickness, distribution, and evolution trend of sedimentation at the reservoir bottom from indirect observation data, generating sedimentation evolution trend data. To further enhance data reliability, the system uses an adaptive calibration algorithm to process the sedimentation evolution trend data obtained from the inversion calculations. This algorithm can identify and compensate for long-term drift errors of sensors and eliminate interference introduced by environmental factors such as changes in water turbidity. Ultimately, it outputs high-precision, calibrated sedimentation status monitoring results, providing direct evidence for reservoir scheduling and dredging decisions.

[0055] Example 1: Referring to Figure 2, the deployment of the layered intelligent sensor array is the physical foundation for the entire monitoring system. Its design concept stems from a deep understanding of the layered characteristics of reservoir sedimentation. Reservoir sedimentation is not a uniform overall process, but a complex phenomenon that evolves in layers and stages from the water surface to the bottom under different physicochemical conditions. Before the deployment work begins, a thorough on-site survey and scheme design are required. The survey includes the underwater topography of the reservoir, historical hydrological data, seasonal variation patterns of sediment concentration entering the reservoir, and the general distribution of existing sediment. Based on this preliminary survey data, geographic information system technology is used to scientifically divide the monitoring area of ​​the reservoir. The core principle of the division is to divide the vertical water column into three monitoring zones with distinct functions according to hydrodynamic conditions and sediment deposition patterns: the surface monitoring zone, the middle layer monitoring zone, and the bottom layer monitoring zone. Each zone corresponds to a specific link in the sedimentation formation chain. Surface monitoring zones are typically located at a certain depth above the water surface. This area is directly affected by wind and wave disturbances, convection caused by temperature differences, and the impact of inflow water. It is the first interface between suspended sediment and water. Pressure sensors deployed in this area need to have high sensitivity and good dynamic response characteristics. These pressure sensors are not installed in isolation, but rather through support cables pre-embedded in the reservoir area or bracket structures fixed to underwater platforms, ensuring that their sensing diaphragms are precisely oriented towards the reservoir bottom. The installation depth needs to be optimized based on the reservoir's annual water level fluctuation range to avoid direct impact from surface waves while sensitively capturing the small increments in hydrostatic pressure caused by the initial settling of suspended particles in the surface water. The data output from the pressure sensors reflects the total weight of water and suspended matter per unit area of ​​water column. Its trend over time can indirectly reveal changes in surface sediment flux, providing first-hand data for determining the sediment load entering the reservoir.

[0056] The mid-layer monitoring zone occupies most of the vertical space of the water body and is a crucial transition zone connecting the surface and bottom layers. It is also the area where sediment transport is most active. The water flow structure and velocity shear force in this zone directly determine the transport path and sediment deposition timing of sediment particles. The acoustic Doppler current profiler deployed in this area is an advanced underwater acoustic device. It emits sound pulses of a specific frequency into the water and receives the echo signals scattered from suspended particles in the water. Using the Doppler frequency shift principle, it calculates the water flow velocity at different depth points along the direction of the sound beam. The sensor is typically installed using bottom-mounted, anchored buoy-suspended, or fixed to the sidewall of an underwater structure. The core requirement is to ensure that the transducer's emitting surface has an unobstructed acoustic field of view, covering the target water layer from near the surface to near the bottom. A single measurement by this sensor can obtain a velocity distribution cloud map on a vertical profile, i.e., mid-layer velocity distribution data. This data reveals the kinetic energy gradient of the water flow, turbulence intensity, and possible density current phenomena, providing crucial information for analyzing whether sediment will continue to be transported to the bottom or remain suspended in the water. The bottom monitoring zone, directly in contact with the reservoir bed, is the primary area for sediment deposition, silt formation, and evolution. Monitoring this zone is crucial for accurately assessing silt thickness and density. Fiber Bragg grating strain sensors deployed in this area leverage the extreme sensitivity of fiber Bragg gratings to minute strains. These sensors are typically pre-encapsulated in an array within a corrosion-resistant, pressure-resistant flexible substrate, forming sensing optical cables. Deployment can be achieved in two main ways: one is to lay the sensing optical cables directly on the reservoir bed matrix and fix them between pre-set grid-like foundation piles, ensuring close contact with the sediment surface; the other is to vertically drive inclinometer tubes or dedicated monitoring piles, integrating fiber Bragg grating sensors, into the reservoir bed sediment layer, with the sensors distributed along the pile shaft to detect compressive deformation at different depths. When new sediment particles are deposited above the sensor or when the underlying sediment undergoes consolidation and deformation, they cause minute stretching or compression of the substrate on which the sensor is located. This results in a shift in the center wavelength of the fiber optic grating. By monitoring this wavelength shift using demodulation equipment, the strain experienced by the underlying sediment can be calculated, i.e., the underlying sediment strain data. This data is crucial for assessing the loading effect of new sediment, the compaction process of the sediment layer, and the risk of local slippage.

[0057] The division of these three monitoring zones and the deployment of sensors are not independent but form an organic whole. A network of point-based pressure sensors, linear profile flow sensors, and area-covering strain sensors weaves together a three-dimensional monitoring network. This network achieves full spatial coverage of key sedimentation processes in the reservoir and provides complementary verification of pressure (load), flow velocity (dynamic force), and strain (response) in terms of monitored physical quantities. This deployment strategy, combining points, lines, and areas and coordinating multiple parameters, maximizes the spatial representativeness and physical integrity of the original collected data, providing a rich and reliable source of field observation data for subsequent data fusion and model inversion. The entire deployment process requires comprehensive consideration of the feasibility of underwater operations, the long-term stability of sensor operation, and the convenience of subsequent data acquisition and transmission, ensuring that this physical infrastructure can function stably and reliably in the complex actual environment of the reservoir over the long term.

[0058] Example 2: Referring to Figure 3, real-time fusion processing of multi-source monitoring data serves as a bridge connecting raw data acquisition and advanced model inversion. Its core objective is to transform heterogeneous monitoring data from different types, locations, and timestamps into a high-quality dataset that is unified in time, space, and dimensions and can be directly used by the model. This process begins with the preprocessing of various types of raw data. The primary object of preprocessing is the surface sedimentation pressure data transmitted back from pressure sensors in the surface monitoring area. These raw voltage or frequency signals are converted into engineering physical quantities, i.e., pressure values. However, due to inherent zero-point offsets, sensitivity differences, and slight drifts that may occur during long-term operation between different pressure sensors, direct use will introduce systematic errors. Therefore, normalization is required. This process sets a reference point for each sensor, typically the average of a set of data collected when the system is initially deployed and the reservoir conditions are relatively stable. The real-time pressure data is subtracted from the sensor's reference value and then divided by its standard range. This maps all pressure data to a dimensionless, zero-reference relative numerical range, generating standardized pressure data that eliminates individual sensor differences and the influence of initial conditions. This process makes pressure sensor data from different locations comparable. The next step is to process the mid-layer velocity distribution data obtained from acoustic Doppler velocity profile sensors in the mid-layer monitoring area. This type of data has a complex structure, containing velocity magnitude and direction information at a series of discrete depth layers along the vertical water depth direction at the same time point. Furthermore, the sensors perform profile measurements continuously or intermittently, resulting in a series of time-series profile data. These data exhibit spatiotemporal mismatches; for example, the measurement times of different sensors may vary slightly, or the depth reference level may change due to water level fluctuations when the same sensor measures at different times. Therefore, spatiotemporal alignment processing is necessary. Spatiotemporal alignment processing utilizes a high-precision clock to synchronize the data timestamps of all sensors, unifying all flow velocity data under an absolute elevation coordinate system to eliminate the influence of water level changes. For data points in the time series, linear interpolation or spline interpolation methods are used to resample non-uniformly sampled data to a unified time node set by the system. For the spatial profile, flow velocity data measured by different sensors at different depths are used to generate a continuous flow velocity field covering the entire monitoring area at a standard depth layer through spatial interpolation algorithms. Finally, aligned flow velocity data that is synchronized in time, continuous in space, and has unified coordinates are generated, providing a consistent spatiotemporal framework for subsequent analysis.

[0059] The signal characteristics of the sedimentation strain data transmitted from the fiber Bragg grating strain sensor in the bottom monitoring area are significantly different from the previous two. Although the center wavelength offset output by the fiber Bragg grating sensor can accurately reflect the strain, the original signal is extremely weak and easily interfered with by various environmental noises. These noises may originate from mechanical vibrations of the water body, thermal noise caused by temperature fluctuations, electromagnetic interference, and electronic noise from the demodulation equipment itself. Therefore, filtering and denoising processing must be performed to extract the effective strain signal that truly reflects the sedimentation changes. Filtering and denoising processing usually employs digital signal processing techniques. Initial outlier removal is performed on the original wavelength data, and abrupt changes that clearly exceed the physical reasonable range are considered interference and removed. A threshold denoising method based on wavelet transform is used because wavelet analysis can analyze signals in both the time and frequency domains simultaneously, making it very suitable for processing non-stationary strain signals. By selecting appropriate wavelet basis functions and decomposition levels, the signal is decomposed into different frequency subbands. Soft or hard thresholding is applied to the high-frequency detail coefficients containing the main noise to suppress noise energy. Finally, the signal is reconstructed to generate smooth, reliable, and denoised strain data that retains the true strain change trend.

[0060] After preprocessing the pressure, velocity, and strain data separately, the core stage of data fusion, feature-level fusion, begins. The goal of this stage is not simply to stack the three sets of data together, but rather to extract characteristic information from each type of data that is fundamentally related to the sedimentation process, starting from the physical mechanisms, and then organically combining them. In the feature extraction stage, for standardized pressure data, the moving average or slope of the trend relative to the long-term background value is calculated as the pressure feature to characterize the cumulative rate of net sedimentation load. For aligned velocity data, the vertical velocity gradient, near-bottom boundary layer velocity magnitude, and turbulent kinetic energy are calculated as velocity features; these features are directly related to the initiation, suspension, and transport capacity of sediment. For denoised strain data, the time integral or the cumulative strain change within a certain time window is calculated as the strain feature, reflecting the degree of compaction or overall deformation of the bottom sediments. The feature-level fusion process assigns corresponding weights or establishes correlation functions to these different types of feature vectors based on a pre-established sedimentation physics model. For example, it uses a fusion algorithm based on Kalman filtering or a simple weighted average model to integrate the pressure features representing the load, the flow velocity features representing the dynamics, and the strain features representing the response into a multi-dimensional, comprehensive feature vector. This fused feature vector integrates information from different sensors that reflects different aspects of sedimentation, ultimately generating a set of fused data that can comprehensively describe the movement and deposition of sediment in the water body and eliminates redundancy and contradictions in the original data. This set of data serves as the standardized input, providing a solid and reliable data foundation for the subsequent dynamic inversion calculations of hydrodynamic models and sediment deposition models.

[0061] Dynamic inversion calculations based on hydrodynamic and sediment deposition models using fused data represent a significant advancement from "observation" to "cognition." Essentially, this involves using field observation data to drive physical and mathematical models, thereby quantitatively estimating key sedimentation parameters that cannot be directly measured, such as sediment thickness and evolution trends. Dynamic inversion calculations require in-depth mining of the fused data to extract the driving features needed for the model. From the velocity field information contained in the fused data, key hydrodynamic features can be extracted, such as the vertical average velocity, the depth of maximum velocity, and the distribution of shear stress in the riverbed. These feature parameters are core variables describing the sediment-carrying capacity and energy dissipation of the flow. From the pressure and strain information contained in the fused data, sediment-related features can be extracted, such as the spatial gradient of pressure changes, the time rate of strain accumulation, and the correlation of strain responses at different locations. These features reflect the distribution of sediment loads and the mechanical state of the underlying medium. The extracted flow velocity features are imported as the main input parameters into a pre-constructed hydrodynamic model, which is usually built based on shallow water equations or three-dimensional Reynolds-averaged Navier-Stokes equations and has been customized with parameters calibrated for reservoir boundary conditions. After receiving the real-time flow velocity features, the model can simulate and calculate more detailed hydrodynamic distribution data within the monitoring area by solving the governing equations, including streamline distribution, vortex structure, and spatial variation of bottom shear stress. These calculation results exceed the limitations of direct sensor measurement points, providing continuous dynamic field information for the entire area, clearly revealing where the energy of the flow decays, thus facilitating sediment deposition, and where strong dynamics may cause resuspension of sediment.

[0062] The pressure and strain characteristics extracted from the fused data were input into a sediment settling model. This model, based on sediment motion mechanics, considers sediment particle settling velocity, flocculation, floc settling, and the initial consolidation process after deposition. Pressure characteristics provide information on the total vertical stress acting on the bed, helping to distinguish between water gravity and sediment gravity. Strain characteristics directly reflect the compressive deformation behavior of the sedimentary layer under load, and its rate of change over time is closely related to the consolidation rate of the sediment and the newly added sediment thickness. By integrating these characteristics, the model can calculate the net sediment flux per unit area, the settling rate of sediments of different particle sizes, and the resulting instantaneous increase in sediment thickness, thus generating sediment settling rate data. This data quantifies the process of sediment separating from the water body and depositing onto the bed surface. The final step involves coupling and dynamically inverting the hydrodynamic distribution data calculated by the hydrodynamic model with the sediment deposition rate data calculated by the sediment deposition model. This coupling is not a simple superposition but considers the interaction and feedback between the two. For example, in high shear stress areas, sediment deposition may be immediately resuspended even if it occurs, while in low-energy environments, deposition dominates. The inversion algorithm iteratively adjusts the internal state variables of the model to achieve the best match between the model's output (such as the predicted velocity distribution and deposition rate) and the observation features extracted from the fused data. In this process, hydrodynamic conditions constrain sediment transport and redistribution, while the deposition rate reflects the net deposition effect. Through this data assimilation technique, the system can dynamically deduce the spatial distribution of sediment thickness, the morphological characteristics of sediment bodies, and the extrapolated evolution trend based on the current hydraulic and sediment conditions of the reservoir bed. This generates sediment evolution trend data containing quantitative information and uncertainty assessment, elevating discrete observation point information into a continuous, quantitative, and predictive understanding of the sedimentation status of the entire monitoring area.

[0063] Example 3: Adaptive calibration algorithms for correcting sedimentation evolution trend data are a core element in improving the long-term robustness and reliability of the monitoring system. This process aims to systematically identify and compensate for the combined effects of slow performance changes in the sensor system itself and instantaneous fluctuations in the external aquatic environment on the final monitoring results. Adaptive calibration is not a simple linear process, but a closed-loop feedback loop involving historical data learning, real-time parameter updates, and dynamic correction. It relies on continuously accumulated monitoring data to optimize the current output. Its primary task is to construct a sensor drift compensation parameter library that can characterize the long-term performance degradation patterns of the sensor. The foundation for constructing this parameter library is acquiring a large amount of historical sedimentation monitoring data. This data comes from all raw observations, intermediate processing results, and final sedimentation trend data stored in the database since the system was put into operation. It also needs to be combined with baseline true values ​​obtained at specific time points through high-precision, traceable independent measurement methods (such as periodic underwater topographic mapping). These independent measurement data provide an objective reference system for evaluating the long-term accuracy of the sensor system. The process of establishing a sensor drift compensation parameter library is a continuous data mining and model fitting process. The system periodically analyzes the output sequence of each sensor to identify its systematic deviation relative to the initial stable operating period of the system or relative to recent independent benchmark measurements. For pressure sensors, drift may manifest as zero-point drift coefficient and sensitivity variation coefficient; for flow profilers, it may manifest as time-varying adjustments to the acoustic attenuation compensation coefficient or flow velocity calibration coefficient; for fiber optic strain gauges, it may manifest as a slow shift in the reference center wavelength. Specifically, pressure sensor drift is mainly manifested as zero-point drift coefficient and sensitivity variation coefficient. The zero-point drift coefficient refers to the benchmark shift of the sensor's output value under no-load conditions, while the sensitivity variation coefficient reflects the change in the ratio between the sensor output and the actual pressure value over time. The specific compensation process begins with benchmark data acquisition under non-siltation conditions, such as during the hydrological stability period in the reservoir area, recording the initial output value of the pressure sensor as a benchmark. The system periodically compares the deviation values ​​of real-time monitoring data and benchmark data, and identifies systematic shifts by calculating the moving average or trend slope of the deviation. These offsets are quantified as zero-point drift coefficients and sensitivity variation coefficients, and stored in the sensor drift compensation parameter library. During compensation, the system retrieves the corresponding drift coefficient based on the current timestamp and performs reverse correction on the real-time pressure data. For example, it subtracts the zero-point drift from the reading and adjusts the sensitivity ratio to generate calibrated data. The drift of an acoustic Doppler velocity profile sensor may manifest as a time-varying adjustment of the acoustic attenuation compensation coefficient or the velocity calibration coefficient. The acoustic attenuation compensation coefficient is used to correct for the absorption effect of suspended particles in the water on the acoustic signal, while the velocity calibration coefficient reflects the change in the proportional relationship between the measured velocity value and the true value.In the specific compensation process, the system initially collects the sensor's baseline output under low flow velocity conditions during deployment and periodically compares the deviation between real-time data and the baseline value. Deviation analysis focuses on the attenuation law of acoustic signal intensity and the consistency of flow velocity output. Through modeling, the deviation is converted into adjustment amounts for acoustic attenuation compensation coefficients or flow velocity calibration coefficients. These coefficients are stored in time series and applied to real-time flow velocity data before data fusion. For example, the signal intensity is corrected by adjusting the acoustic attenuation compensation coefficient, or the measured value is rescaled using the flow velocity calibration coefficient to eliminate the effects of long-term drift. The drift of the fiber Bragg grating strain sensor is mainly manifested as a slow shift in the reference center wavelength, that is, a small change in the baseline value of the center wavelength of the sensor under strain-free conditions over time. The specific compensation process includes collecting the baseline value of the center wavelength of the fiber Bragg grating sensor during the initial stable operation of the system and continuously monitoring the real-time wavelength data. The system calculates the deviation between the real-time wavelength and the baseline value, identifies the shift trend, and quantifies the shift as a compensation parameter for the reference center wavelength. This parameter is correlated with the sensor's cumulative operating time or ambient temperature and stored in the drift compensation parameter library. In actual monitoring, the system retrieves the current compensation parameters and corrects the real-time wavelength data. For example, it restores the wavelength value to the reference state by adjusting the offset, thereby ensuring the accuracy of the strain data and eliminating drift caused by material aging or temperature fluctuations. These parameters are modeled as functions related to the sensor's cumulative operating time, the average temperature of the working environment, etc., and are stored in an orderly manner according to the time series, forming a sensor health status profile that evolves over time.

[0064] After acquiring the current sedimentation evolution trend data, the calibration algorithm immediately queries the sensor drift compensation parameter library to retrieve the latest drift compensation parameter estimates for all sensors involved in generating this trend data within the current time period. Based on these parameters, it performs reverse compensation on the original inversion calculation process or directly corrects the calculation results. For example, if the drift parameters of a pressure sensor indicate a positive zero-point drift, a corresponding estimated drift amount will be subtracted from the data involving that sensor before data fusion or after model inversion. This targeted compensation operation aims to eliminate systematic errors introduced by slow changes in sensor performance, generating a set of preliminary calibration data that is theoretically closer to the ideal sensor output state. The preliminary calibration data is still exposed to complex external aquatic environmental interference, among which changes in water turbidity are a significant factor with wide-ranging and rapid effects. High turbidity water not only attenuates the acoustic signal intensity of the acoustic Doppler current profiler but may also indirectly affect the overall fluid density sensed by the pressure sensor due to changes in suspended particulate matter concentration. Therefore, further interference correction must be performed based on real-time water turbidity data. This system is usually integrated with an auxiliary optical backscattering turbidimeter to continuously provide high temporal resolution water turbidity data.

[0065] The core of interference correction based on turbidity data is to establish a dynamic, turbidity-related correction model. This model learns the influence of turbidity changes on different sensor readings and the final inversion results through historical data. This correction relationship can be expressed as a function mapping, where turbidity is the independent variable and the correction amount is the dependent variable. A mathematical expression representing this relationship is as follows:

[0066]

[0067] in: This represents the compensation value for the pressure sensor reading calculated from the current turbidity state. Its dimension is the same as pressure, in Pascals (Pa) or kilopascals (kPa). Coefficient It is a scale factor with pressure dimensions, measured in Pascals, which determines the maximum compensation magnitude for the effect of turbidity changes on pressure readings. Variable This represents the real-time turbidity value of the water body, measured in NTU. (Variable) This is the selected background reference turbidity, also in NTU. Parameter This is a scaling factor with turbidity dimensions, measured in NTU, which controls the sensitivity of the compensation value to changes as the turbidity difference increases. The use of a logarithmic function causes the compensation value to increase with increasing turbidity difference, but the rate of increase gradually slows down, which aligns with physical intuition that the marginal effect weakens at extremely high turbidity levels. The entire logarithmic parameter... It is dimensionless, therefore the dimensions on both sides of the equation are pressure dimensions, maintaining consistency.

[0068] Based on the real-time turbidity readings, the system dynamically calculates a specific pressure reading compensation value using the calibrated formula described above. This compensation value is applied to the pressure-related portion of the preliminary calibration data. For example, before using pressure data to invert sediment thickness, the original pressure readings or sediment loads extrapolated from pressure are corrected using this compensation value, thereby partially offsetting the measurement bias caused by changes in fluid density due to variations in suspended sediment concentration at the physical level. This secondary correction based on real-time environmental parameters effectively suppresses short-term, non-sedimentary signal interference introduced by drastic fluctuations in water particulate matter concentration. Ultimately, a set of calibrated sedimentation monitoring results is generated that corrects for both the sensor's inherent long-term drift and compensates for transient external environmental interference. This significantly improves the monitoring system's adaptability to complex hydrological environments, enabling its output of key indicators such as sediment thickness, distribution range, and evolution trends to more accurately and reliably reflect the actual dynamics of sediments at the bottom of the reservoir.

[0069] Referring to Figure 4, the complete workflow and data analysis results of the system are illustrated. The upper figure clearly shows the entire processing of pressure sensor data from raw data to final calibration data, with lines of different gray levels clearly displaying the gradual optimization effects of sensor drift compensation and environmental impact correction. The lower figure reveals the intrinsic relationship between the evolution trend of sedimentation thickness and environmental parameters, comparing the long-term sedimentation process with the dynamic changes in water turbidity, intuitively reflecting the degree of influence of environmental factors on the monitoring results. Through professional visualization design, the entire chart fully demonstrates the effective role of the adaptive calibration algorithm in eliminating sensor performance degradation and environmental interference, reflecting the robustness and reliability of the monitoring system in maintaining long-term stable operation in complex hydrological environments.

[0070] Example 4: Establishing a sensor drift compensation parameter library is a dynamic process relying on long-term data accumulation and self-learning. This process quantifies performance changes by comparing the sensor system's output under known steady-state conditions with real-time measurements. This process is typically implemented during periods of relatively stable reservoir hydrological conditions, such as the dry season when flow velocities are low, inflow sediment content is low, and water stratification is stable. At this time, it can be approximated that the reservoir bottom sedimentation has not changed significantly, and the theoretical changes in sensor readings should mainly originate from environmental noise and the sensor's own performance drift. The system initiates a baseline data acquisition mode, continuously recording the output data of all relevant sensors over a period of time. Pressure sensors record their hydrostatic pressure values, flow profilers record their low-velocity background noise, and fiber optic grating sensors record the long-term fluctuations of their center wavelength. These data are averaged over a certain time window, and the calculated average value is established as the baseline data for the sensor under current operating conditions. After the baseline data is established, the system enters a routine drift monitoring phase. The real-time acquired monitoring data is compared with the baseline data of the corresponding sensor to calculate the deviation value. This deviation value not only includes numerical differences but also relates to environmental parameters at the time of calculation, such as water temperature and depth, because the drift characteristics of some sensors are closely related to ambient temperature. For example, a pressure sensor installed at a water depth of ten meters might have an initial reference pressure value of 101 kPa. In a subsequent measurement, if the water depth and temperature are confirmed to be completely consistent with the reference state, but the sensor reading becomes 101.5 kPa, then this 0.5 kPa offset will be initially identified as a potential drift. To confirm that this is a systematic drift rather than random fluctuation, the system will continuously record the deviation values ​​for multiple measurement periods and perform trend analysis. Only those deviations that show persistence and directionality will be ultimately confirmed as valid drift.

[0071] Once the confirmed drift is calculated and verified, a drift compensation parameter record is generated. This record contains several key fields: a unique sensor identifier, a timestamp of the parameter calculation, the calculated drift compensation value, reference values ​​of the environmental conditions used in the calculation, and a credibility score for the record. The credibility score depends on the duration and stability of the data used for the calculation; drift parameters calculated based on a week of stable data have higher credibility than those calculated based on a single day's data. All these records are stored chronologically to form a sensor drift compensation parameter library, which effectively becomes a historical archive recording the evolution of each sensor's "health status."

[0072] As data accumulates in the parameter library, the system can perform more in-depth analysis of the drift patterns of each sensor. For example, it can fit a functional relationship between the drift amount and the sensor's cumulative operating time or ambient temperature. Table 1 shows a simplified fragment of a pressure sensor drift compensation parameter library, recording the sensor's drift over a period of time.

[0073] Table 1: Record of Drift Compensation Parameters for Pressure Sensor P-001

[0074] Recording timestamp, sensor ID, reference water temperature (°C), reference water depth (m), calculated drift (kPa), reliability score. 2023-06-01 P-00 115.2 10.0 0.00 1.00 2023-09-15 P-00 118.5 9.8 0.05 0.85 2023-12-20 P-00 18.3 10.2 0.12 0.95 2024-03-10 P-00 111.7 9.9 0.18 0.90 surface

[0075] Correcting interference in preliminary calibration data based on water turbidity data is a post-processing step to address the impact of short-term environmental changes. The core of this step lies in establishing a correlation model between turbidity changes and observation system errors. The correction process begins with the acquisition of real-time water turbidity data, typically performed by optical turbidity sensors deployed near monitoring points. These sensors measure turbidity values ​​in the water at a frequency of minutes or seconds and provide the data to the processing system. The system maintains a background turbidity value, usually taken as the moving average of recent turbidity readings during periods of low flow velocity and no rainfall disturbance, representing the background turbidity level of the water body under calm conditions. After a real-time measurement is completed, the system reads the current turbidity value and compares it with the background turbidity, calculating the turbidity deviation value. Based on a correction rule base trained beforehand using experimental or historical data, this turbidity deviation value is mapped to a specific interference correction coefficient or correction amount. For example, historical data analysis might show that when turbidity instantaneously increases by more than 20 NTUs above the background value, the acoustic Doppler current profiler's measurements in the near-bottom region may be systematically lower by 3% to 5% due to sound wave attenuation. When a similar situation occurs again, the system will use this empirical relationship to make a positive compensation correction to the bottom shear stress calculated from the flow velocity in the initial calibration data. For pressure sensors, high turbidity means an increase in the concentration of suspended particulate matter in the water, and a slight increase in the average density of the entire water column. This results in a higher pressure sensor reading at the same water depth compared to clear water. Therefore, it is necessary to estimate the increase in water density based on the real-time turbidity value and subtract the pressure change caused by the density change from the pressure reading, retaining only the pressure signal generated by the weight of sediment.

[0076] The interference correction coefficient is typically a function related to the turbidity deviation value. This function may be a piecewise linear relationship or a simple lookup table, and it is stored in the system's configuration file. There are several ways to apply the correction coefficient. It can be used to directly scale the final sediment thickness result after initial calibration, or it can be applied back to the pre-fusion stage to adjust certain characteristic parameters significantly affected by turbidity. For example, during periods of abnormally high turbidity, the system may choose to reduce the weight of acoustic velocity data in data fusion, while relying more on pressure and strain data for inversion calculations. This adaptive weight adjustment weakens the influence of unreliable data sources. The calibrated sediment status monitoring results generated after all corrections are completed have significantly enhanced anti-interference capabilities and can more realistically reflect the actual dynamics of underwater sedimentation. The entire correction process is a dynamic feedback loop; the latest correction effects are recorded and used to optimize future correction rules, continuously enhancing the system's adaptability to changes in the external environment.

[0077] Example 5: After generating the calibrated sedimentation monitoring results, the final step in realizing the value of the entire monitoring method is to systematically manage, deeply analyze, and intuitively present them. This process transforms discrete, specialized data into a sustainable, traceable historical archive and cognitive information that can assist in decision-making. The system automatically stores the calibrated sedimentation monitoring results generated at the end of each processing step, along with their corresponding data acquisition timestamps, spatial coordinates of the monitoring area, data quality identifiers, and other metadata, in a time-series ordered manner into a specially designed sedimentation database. This database is constructed using a time-series database or a relational database with a time index to accommodate the efficient writing, storage, and rapid querying of massive amounts of monitoring data by time range. Each record represents a snapshot of the sedimentation status at a specific time and location. Key data fields may include monitoring point number, latitude and longitude coordinates, elevation, calculated sediment thickness, estimated sediment dry density, data uncertainty range, and the algorithm version number used to generate the data. This structured storage method not only ensures the integrity and traceability of the data but also provides a complete data foundation for any subsequent historical data backtracking analysis or algorithm re-verification.

[0078] As data accumulates, the siltation database gradually forms a valuable long-sequence historical dataset on the evolution of the reservoir's bottom topography. Based on this ever-expanding database, the system can perform preset data mining and trend analysis tasks, automatically generating siltation change trend reports periodically. The report generation cycle can be configured according to management needs, such as monthly, quarterly, or annually. The report generation process is fully automated, requiring no manual intervention. The system triggers an analysis script at a preset time point, extracting all relevant data records for the specified time period from the siltation database. The report aims to go beyond a single point in time, revealing the dynamic process, speed, and spatial patterns of siltation development. It transforms dry data sequences into indicators and charts with clear physical and management significance. Generating a siltation change trend report based on the siltation database is a multi-step analysis process. The system extracts all siltation status monitoring results for the specified analysis period from the database. These data are then interpolated or statistically analyzed in a gridded manner according to their spatial coordinates, forming a spatial distribution map of siltation thickness at each time point. By comparing the siltation thickness grid data of adjacent time points (e.g., the end of this month and the end of last month), the system calculates the change in siltation thickness grid by grid, that is, the end-of-period thickness minus the beginning-of-period thickness. This change is divided by the time interval (in months or years) to obtain the siltation thickness change rate of each grid point in that time period. This change rate intuitively reflects the speed of siltation in different areas of the reservoir. A positive value indicates increased siltation, while a negative value indicates scouring or dredging.

[0079] Simultaneously, the system calculates another important indicator: the siltation range expansion rate. This indicator focuses on the spatial expansion of siltation. The analysis requires defining a critical siltation thickness threshold (e.g., 0.1 meters). Areas exceeding this threshold are considered effective siltation zones. The system calculates the total area of ​​the effective siltation zone at each time point, calculates the change in this area between adjacent time points, and divides this change by the time interval to obtain the siltation range expansion rate. This indicator reveals the speed of horizontal development of siltation. In addition to these quantitative indicators, the system also performs spatial feature analysis, such as identifying hotspots with the highest siltation thickness change rate, the most significant outward expansion of the siltation range, and stable or scoured areas that may continue to be eroded. After calculating these quantitative indicators and spatial features, the system calls the report generation module to integrate the numerical results with visualization elements, automatically generating a structured siltation change trend report, typically output in PDF or web page format. The report typically includes an abstract outlining the overall siltation status of the monitoring area during the reporting period; the main body contains detailed data tables listing the rate of change and extent of siltation thickness in each zone; and a series of charts, such as time-series curves of siltation thickness at different monitoring points, spatial overlay maps of siltation thickness contour lines at different periods, and spatial distribution cloud maps of siltation velocity. These charts make the dynamic process of siltation readily apparent. The report may also include technical support information such as data sources, algorithm descriptions, and uncertainty analysis. This type of siltation trend report transforms raw, professional monitoring data into decision support information that reservoir operation and management personnel can directly understand and use. It clearly demonstrates severely affected areas of siltation, their development speed, and potential future risks, providing objective and quantitative evidence for optimized reservoir scheduling, flood control and disaster reduction, and scientific dredging planning.

[0080] It should be noted that, in this document, relational terms such as "first" and "second" are used only to distinguish one entity or operation from another, and do not necessarily require or imply any such actual relationship or order between these entities or operations. Furthermore, the terms "comprising," "including," or any other variations thereof are intended to cover non-exclusive inclusion, such that a process, method, article, or apparatus that comprises a list of elements includes not only those elements but also other elements not expressly listed, or elements inherent to such process, method, article, or apparatus.

[0081] Although embodiments of the invention have been shown and described, it will be understood by those skilled in the art that various changes, modifications, substitutions and alterations can be made to these embodiments without departing from the principles and spirit of the invention, the scope of which is defined by the appended claims and their equivalents.

Claims

1. A method for dynamic monitoring of underwater sedimentation in reservoirs, characterized in that, Includes the following steps: A hierarchical intelligent sensor array is deployed, comprising a pressure sensor, an acoustic Doppler flow profile sensor, and a fiber optic strain sensor, forming a three-dimensional monitoring network of points, lines, and surfaces. Multi-source monitoring data from the hierarchical intelligent sensor array, including pressure data, flow velocity data, and strain data, are collected in real time via edge computing nodes. The multi-source monitoring data is then fused in real time to generate fused data. Based on the hydrodynamic model and sediment deposition model, the fused data is dynamically inverted to obtain sedimentation evolution trend data. An adaptive calibration algorithm is used to calibrate the siltation evolution trend data to eliminate the influence of sensor drift and water body interference, and generate calibrated siltation status monitoring results. The deployment of the layered intelligent sensor array includes: dividing the monitoring area according to the reservoir topography into a surface monitoring area, a middle monitoring area, and a bottom monitoring area; deploying the pressure sensor in the surface monitoring area to collect surface sedimentation pressure data; deploying the acoustic Doppler velocity profile sensor in the middle monitoring area to collect middle-layer velocity distribution data; and deploying the fiber optic strain sensor in the bottom monitoring area to collect bottom-layer sedimentation strain data. The real-time fusion processing of the multi-source monitoring data includes: normalizing the surface sedimentation pressure data to generate standardized pressure data; and performing spatiotemporal alignment processing on the middle-layer velocity distribution data to generate aligned velocity data. The underlying sedimentation strain data is filtered and denoised to generate denoised strain data; the standardized pressure data, aligned flow velocity data, and denoised strain data are fused at the feature level to generate fused data; the dynamic inversion calculation of the fused data based on the hydrodynamic model and sediment deposition model includes: extracting flow velocity features, pressure features, and strain features from the fused data; inputting the flow velocity features into the hydrodynamic model to calculate hydrodynamic distribution data; inputting the pressure features and strain features into the sediment deposition model to calculate sediment deposition rate data; and combining the hydrodynamic distribution data and sediment deposition rate data to generate sedimentation evolution trend data.

2. The method for dynamic monitoring of underwater sedimentation in reservoirs according to claim 1, characterized in that, The calibration of the sedimentation evolution trend data using an adaptive calibration algorithm includes: acquiring historical sedimentation monitoring data and establishing a sensor drift compensation parameter library; performing drift compensation on the sedimentation evolution trend data based on the sensor drift compensation parameter library to generate preliminary calibration data; and performing interference correction on the preliminary calibration data based on water turbidity data to generate the calibrated sedimentation status monitoring results.

3. The method for dynamic monitoring of underwater sedimentation in reservoirs according to claim 2, characterized in that, The establishment of the sensor drift compensation parameter library includes: collecting reference data of the hierarchical intelligent sensor array in a non-accumulation state, calculating the deviation value between the reference data and the real-time monitoring data, generating drift compensation parameters, and storing the drift compensation parameters in the sensor drift compensation parameter library in a time series.

4. The method for dynamic monitoring of underwater sedimentation in reservoirs according to claim 3, characterized in that, The interference correction of the preliminary calibration data based on water turbidity data includes: collecting real-time water turbidity data through an optical sensor, determining an interference correction coefficient based on the real-time water turbidity data, and using the interference correction coefficient to perform weighted correction on the preliminary calibration data.

5. The method for dynamic monitoring of underwater sedimentation in reservoirs according to claim 4, characterized in that, After generating the calibrated siltation status monitoring results, the method further includes: storing the calibrated siltation status monitoring results into a siltation database in a time series manner, and generating a siltation change trend report based on the siltation database.

6. The method for dynamic monitoring of underwater sedimentation in reservoirs according to claim 5, characterized in that, The process of generating a siltation change trend report based on the siltation database includes: extracting siltation status monitoring results from the siltation database for different time periods; calculating the siltation thickness change rate and siltation range expansion rate for adjacent time periods; and generating the siltation change trend report based on the siltation thickness change rate and the siltation range expansion rate.

7. A dynamic monitoring system for underwater sedimentation in a reservoir, comprising a memory, a processor, and a computer program stored in the memory and running on the processor, characterized in that, When the processor executes the computer program, it implements the steps of the dynamic monitoring method for underwater siltation in reservoirs as described in any one of claims 1 to 6.

Images