Method and system for monitoring sediment deposition of underwater structures based on multi-source perception

Abstract

The application provides a kind of underwater building siltation monitoring method and system based on multi-source perception, it is related to underwater engineering technical field.The system includes: sensor module, the sensor of sensor module is arranged around or below the building, sensor module is used to collect siltation data around the building;Data transmission module, data transmission module is connected with sensor module, for transmitting siltation data;Data processing module is used to receive siltation data transmitted by data transmission module, siltation trend prediction is carried out based on pre-trained siltation model and siltation data, and early warning is carried out according to the prediction result.The application embodiment can monitor siltation condition by using various sensors, effectively improve the accuracy and comprehensiveness of siltation monitoring, and can automatically predict siltation trend according to the monitoring result, with high accuracy and fast speed, meet the current underwater building safety operation needs.

Description

Technical Field

[0001] This application relates to the field of underwater engineering technology, and more specifically, to a method and system for monitoring sediment deposition in underwater structures based on multi-source sensing. Background Technology

[0002] Underwater structures refer to buildings that exist underwater. They are structures made of steel or reinforced concrete and are submerged in water. They can be standalone buildings or the underwater portion of a building.

[0003] During the construction of underwater structures, silt carried by water currents gradually accumulates around and at the base of the structure. This siltation not only alters the flow pattern of the water around the underwater structure, increasing the erosion of the foundation by localized water currents, but also changes the stress conditions of the structure itself. When the siltation exceeds the design capacity, it will further increase the structural load, causing structural deformation, displacement, or even instability and failure, seriously threatening the operational safety of the underwater structure. Therefore, continuous monitoring of siltation around underwater structures, timely understanding of siltation dynamics, and early warning of abnormal siltation are of great significance for ensuring the safe operation of underwater structures. However, current traditional underwater siltation monitoring mostly uses a single sensor for monitoring, with manual warnings based on the monitoring results. This method is easily affected by the complex underwater environment and the limitations of single-sensory monitoring equipment, resulting in insufficient accuracy and comprehensiveness of the monitoring data. It is difficult to accurately predict siltation trends and cannot meet the current monitoring needs for the safe operation and maintenance of underwater structures. Summary of the Invention

[0004] This application provides a method and system for monitoring sediment deposition in underwater structures based on multi-source sensing. This method addresses the shortcomings of existing sediment deposition monitoring systems, such as insufficient accuracy and comprehensiveness, difficulty in accurately predicting deposition trends, and inability to meet monitoring requirements. To achieve this objective, this application provides the following solutions.

[0005] According to one aspect of the embodiments of this application, a multi-source sensing-based underwater structure sediment deposition monitoring system is provided, comprising: The sensor module has sensors deployed around or below the building. The sensor module is used to collect data on siltation around the building. The sensors include at least two of the following: sonar siltation sensor, laser scanner, flow velocity and direction sensor, shear force sensor, silt concentration sensor, and pressure sensor. A data transmission module, which is connected to the sensor module, is used to transmit the siltation data; The data processing module is used to receive the sedimentation data transmitted by the data transmission module, predict the sedimentation trend based on the pre-trained sedimentation model and the sedimentation data, and issue an early warning based on the prediction results.

[0006] In one possible implementation, an installation module is included, to which the sensor module and the data transmission module are connected and fixed to the building.

[0007] In one possible implementation, the fixing method of the installation module includes at least one of negative pressure, magnetic attraction, bolts, and clips, and the fixing method of the installation module corresponds to the installation environment and the surface material of the building.

[0008] In one possible implementation, the mounting module includes a suction cup and a vacuum pump, the vacuum pump being connected to the suction cup. The vacuum pump is used to evacuate the suction cup after it is attached to the building to fix the mounting module.

[0009] In one possible implementation, a leveling module is included, on which the sensor is fixed and one side of the leveling module is connected to the mounting module. The leveling module is used to adjust the orientation of the sensor.

[0010] In one possible implementation, the leveling module includes a housing and a telescopic unit, with the sensor disposed on the side of the housing away from the mounting module; The telescopic unit is located on the side of the housing away from the sensor, and includes a telescopic column and a telescopic rod. The telescopic column is used to adjust the height of the housing, and the telescopic rod is used to adjust the housing to a horizontal state.

[0011] In one possible implementation, a power module and a positioning module are included, wherein the power module and the positioning module are mounted on the housing; The positioning module is used to obtain current positioning information; The power module is used to move the installation module to the predetermined installation position according to the positioning information during installation.

[0012] According to one aspect of the embodiments of this application, a method for monitoring sediment deposition in underwater structures based on multi-source sensing is provided, characterized in that it is used in the underwater structure sediment deposition monitoring system based on multi-source sensing as described above, the method comprising: The sensor module is used to acquire sediment deposition data, and the sediment deposition data is preprocessed. The sediment deposition data includes water flow velocity, sediment concentration, sediment deposition thickness, water level, and flow direction. Extract the feature parameters from the sediment deposition data, and input the feature parameters and the preprocessed sediment deposition data into the pre-trained sediment deposition model to obtain sediment deposition trend prediction information; Early warnings are issued based on the predicted siltation trend information and actual siltation data.

[0013] In one possible implementation, training the sediment deposition model includes: A training sample set is constructed based on historical data, including historical monitoring data, engineering survey data, and siltation disaster case data. The model to be trained is determined, the model is trained using the training sample set, and the hyperparameters of the model are adjusted according to the type of the model. The sediment deposition model is obtained based on the adjusted hyperparameters and model training results.

[0014] In one possible implementation, the early warning based on the siltation trend prediction information and actual siltation data includes: Based on the siltation trend prediction information, the predicted siltation height and the actual siltation height corresponding to the actual siltation data are obtained; Early warnings are issued based on the actual siltation height, the predicted siltation height, and the distance between the building's predetermined location.

[0015] The beneficial effects of the technical solutions provided in this application are: The underwater structure sediment deposition monitoring system based on multi-source sensing provided in this application includes: a sensor module, whose sensors are deployed around or below the structure, for collecting sediment deposition data around the structure; the sensors include at least two of the following: sonar sedimentation sensors, laser scanners, flow velocity and direction sensors, shear force sensors, sediment concentration sensors, and pressure sensors; a data transmission module, connected to the sensor module, for transmitting sediment deposition data; and a data processing module, for receiving the sediment deposition data transmitted by the data transmission module, predicting sediment deposition trends based on a pre-trained sediment deposition model and the sediment deposition data, and issuing early warnings based on the prediction results. This application's embodiments can utilize multiple sensors to monitor sediment deposition, effectively improving the accuracy and comprehensiveness of sediment deposition monitoring, and can automatically predict sediment deposition trends based on monitoring results, with high accuracy and speed, meeting the current needs of safe operation and maintenance of underwater structures. Attached Figure Description

[0016] To more clearly illustrate the technical solutions in the embodiments of this application, the accompanying drawings used in the description of the embodiments of this application will be briefly introduced below.

[0017] Figure 1A structural diagram of the underwater structure sediment deposition monitoring system based on multi-source sensing provided in this application embodiment; Figure 2 A top view schematic diagram of a portion of the underwater structure sediment deposition monitoring system provided in this application embodiment; Figure 3 for Figure 2 Schematic sectional view along section AA; Figure 4 A flowchart illustrating the workflow of the underwater structure sediment deposition monitoring system provided in this application embodiment; Figure 5 A flowchart of a method for monitoring sediment deposition in underwater structures based on multi-source sensing, provided in an embodiment of this application; In the picture: 1. Sensor module; 2. Mounting module; 3. Data processing module; 4. Power module; 51. Housing; 52. Telescopic unit. Detailed Implementation

[0018] The embodiments of this application are described below with reference to the accompanying drawings. It should be understood that the embodiments described below with reference to the accompanying drawings are exemplary descriptions for explaining the technical solutions of the embodiments of this application, and do not constitute a limitation on the technical solutions of the embodiments of this application.

[0019] Those skilled in the art will understand that, unless specifically stated otherwise, the singular forms “a,” “an,” “the,” and “the” used herein may also include the plural forms. It should be further understood that the terms “comprising” and “including” as used in embodiments of this application mean that the corresponding feature can be implemented as the presented feature, information, data, step, operation, element, and / or component, but do not exclude implementation as other features, information, data, step, operation, element, component, and / or combinations thereof supported by the art. It should be understood that when we say that an element is “connected” or “coupled” to another element, the one element can be directly connected or coupled to the other element, or it can mean that the one element and the other element establish a connection relationship through an intermediate element. Furthermore, “connected” or “coupled” as used herein can include wireless connection or wireless coupling. The term “and / or” as used herein indicates at least one of the items defined by the term; for example, “A and / or B” indicates implementation as “A,” or implementation as “A,” or implementation as “A and B.”

[0020] To make the objectives, technical solutions, and advantages of this application clearer, the embodiments of this application will be described in further detail below with reference to the accompanying drawings.

[0021] The technical solutions of this application and their effects are described below through several exemplary embodiments. It should be noted that the following embodiments can be referenced, borrowed from, or combined with each other. Identical terms, similar features, and similar implementation steps in different embodiments will not be repeated.

[0022] The method and system for monitoring sediment deposition in underwater structures based on multi-source sensing provided in this application aim to solve at least one technical problem existing in the prior art.

[0023] Optionally, such as Figures 1-4 As shown, the underwater structure sediment deposition monitoring system based on multi-source sensing of this application includes: a sensor module 1, whose sensors are deployed around or below the structure, and is used to collect sediment deposition data around the structure. The sensors include at least two of the following: a sonar sedimentation sensor, a laser scanner, a flow velocity and direction sensor, a shear force sensor, a sediment concentration sensor, and a pressure sensor; a data transmission module, which is connected to the sensor module 1 and is used to transmit sediment deposition data; and a data processing module 3, which is used to receive the sediment deposition data transmitted by the data transmission module, predict the sediment deposition trend based on a pre-trained sediment deposition model and the sediment deposition data, and issue an early warning based on the prediction results.

[0024] Optionally, the underwater structure can be a immersed tube, pile foundation, or other underwater structures. Specifically, the underwater structure can be an immersed tube used for submarine tunnels. This immersed tube is prefabricated in a specialized prefabrication yard and transported to the designated installation area by transport barges or pusher vessels. However, due to the influence of construction schedule and environmental conditions, it is often necessary to temporarily or for extended periods store the prefabricated immersed tube. The storage of the immersed tube alters the water flow pattern in the nearby waters, thereby affecting the bed sedimentation balance and the safety of the stored immersed tube. Therefore, it is necessary to utilize the monitoring system of this application for sedimentation monitoring and early warning.

[0025] Optionally, sedimentation data includes multi-dimensional data related to sedimentation monitoring, such as water flow velocity, sediment concentration, sedimentation thickness, water level, and flow direction.

[0026] Optionally, the sonar sedimentation sensor can be used to scan the underwater bed surface and transmit underwater topographic maps at different times to monitor sediment deposition. A laser scanner can be installed at the bottom of the underwater structure. When the space at the bottom of the underwater structure is large, the scanning data from the laser scanner can supplement and verify the data from the sonar sedimentation sensor; when the space at the bottom of the structure is small and the sonar sedimentation sensor cannot function properly, the laser scanner serves as the primary device for monitoring sedimentation data.

[0027] Flow velocity and direction sensors are used to monitor water flow dynamics, while shear force sensors can be used to detect the shear force generated by sediment. Sediment concentration sensors are used to detect the suspended sediment content in water bodies; pressure sensors are used to monitor stress changes in underwater structures.

[0028] Optionally, the flow velocity and direction sensor can be installed on the upstream side of the underwater structure, the sediment concentration sensor can be installed at the bottom of the underwater structure, and the pressure sensor can also be installed at the bottom of the underwater structure to calculate the water depth through pressure.

[0029] Optionally, the various modules of the monitoring system can adopt a modular design (e.g., sensor module 1 and data transmission module are different modules). Each module can be pre-installed, debugged, and calibrated before leaving the factory, significantly reducing the difficulty of on-site construction. Furthermore, each module can be equipped with interfaces, connectors, and slots for auxiliary fixing. On-site installation only requires module docking, interface plugging / unplugging, and simple fixing operations, eliminating the need for complex wiring and debugging procedures. This significantly shortens the installation cycle, improves construction efficiency, and facilitates subsequent module replacement, maintenance, and upgrades.

[0030] Optionally, the data transmission module can be connected to the sensor in sensor module 1 via a wired or wireless connection. Furthermore, this data transmission module can use wireless transmission to upload sensor data to the control center (which could be data processing module 3, or two different objects) in real time or at regular intervals. The data transmission module can also be equipped with a wired port for emergency data transmission in case of wireless transmission failure.

[0031] Optionally, the data processing module 3 can be located in the same place as the data transmission module (e.g., both can be located around or below the underwater structure), or they can be located separately.

[0032] In one embodiment, the data processing unit supports dual-mode deployment, including both cloud and local server deployments. The deployment method can be flexibly selected based on project needs, data security requirements, and network conditions. Cloud deployment eliminates the need for local hardware resources, facilitating centralized management and data sharing across multiple projects. Local server deployment is suitable for scenarios with unstable networks and high data confidentiality requirements, ensuring real-time data processing and security. Simultaneously, the data processing unit can remotely connect to other devices via wireless communication modules (such as WiFi or Bluetooth) or networking modules (such as network management systems), enabling remote parameter configuration, firmware upgrades, fault diagnosis, and status monitoring without on-site operation, reducing maintenance costs and improving the timeliness and convenience of system maintenance.

[0033] Optionally, the system also includes an installation module 2, with the sensor module 1 and data transmission module connected to the installation module 2 and fixed to the structure via the installation module 2. The installation module 2 secures the sensor module 1 and the data transmission module. The installation module 2 can be fixed to the underwater structure manually or automatically.

[0034] Optionally, the installation module 2 can be fixed to or around the building before it is submerged in water, or it can be fixed after the building is submerged in water.

[0035] Optionally, the mounting module 2 can be fixed by at least one of negative pressure, magnetic attraction, bolts, and clips, and the fixing method of the mounting module 2 corresponds to the installation environment and the surface material of the building.

[0036] In one embodiment, the monitoring system offers a variety of fixing methods, including negative pressure, magnetic attraction, bolts, and clips, to suit different building surface materials (concrete, steel structure, stone, etc.) and installation environments (dry and wet areas, high and low temperature areas, water flow impact areas). Negative pressure and magnetic attraction are suitable for flat metal surfaces, and are easy to install and disassemble without damage; clips are suitable for scenarios with pre-set mounting slots, providing a firm fixation and strong shock resistance.

[0037] Optionally, the mounting module 2 includes a suction cup and a vacuum pump. The vacuum pump is connected to the suction cup and is used to evacuate the suction cup after it is attached to the building to fix the mounting module 2.

[0038] Optionally, when the building has a surface that can be magnetically attracted by a magnet, the mounting module 2 may also include an electromagnet, through which the mounting module 2 achieves the attraction and removal of the system.

[0039] In one embodiment, the installation module 2 includes a vacuum pump and at least one suction cup. During installation, the vacuum pump is turned on so that the entire system is adsorbed onto the side or bottom of the building. After the monitoring task is completed, the vacuum pump is turned on again, and the entire system automatically detaches from the structure, completing rapid dismantling and recycling.

[0040] Optionally, the system also includes a leveling module, on which the sensor is fixed and one side of the leveling module is connected to the mounting module 2. The leveling module is used to adjust the attitude of the sensor, wherein the sensor can be adjusted to a horizontal attitude or an attitude that allows the sensor to accurately acquire data through the leveling module.

[0041] Optionally, the leveling module includes a housing 51 and a telescopic unit 52. The sensor is located on the side of the housing 51 away from the mounting module 2. The telescopic unit 52 is located on the side of the housing 51 away from the sensor and includes a telescopic column and a telescopic rod. The telescopic column is used to adjust the height of the housing 51, and the telescopic rod is used to adjust the housing 51 to a horizontal state.

[0042] Optionally, the mounting module 2 can be installed on top of the telescopic column and / or telescopic rod, with the bottom of the telescopic column connected to the housing 51, and the bottom of the telescopic rod also connected to the housing 51. The height of the housing 51 and its distance from the building can be adjusted by raising and lowering the telescopic column. Multiple telescopic rods can be used; by adjusting the lengths of different rods, the attitude of the housing 51 can be changed, thereby adjusting the attitude of the sensors on the housing 51.

[0043] Optionally, the telescopic column and telescopic rod can be hydraulic devices. One end of the telescopic column can be rotatably connected to the housing 51. The number of telescopic rods can be four or other quantities. The installation positions of different telescopic rods are different. The sensor attitude can be adjusted by adjusting the height of different telescopic rods.

[0044] Optionally, the sensor can also be connected to a telescopic rod, through which the sensor's attitude (e.g., using multiple sensors for attitude adjustment) and height can be adjusted to ensure the accuracy of the sensor's monitoring data.

[0045] Optionally, to enable automatic installation and movement of the monitoring system, the system further includes a power module 4 and a positioning module, which are mounted on the housing 51. The positioning module is used to acquire current positioning information; the power module 4 is used to move the installation module 2 to a predetermined installation position based on the positioning information during installation. This predetermined installation position can be pre-stored in the power module 4 or the data processing module 3.

[0046] Optionally, the positioning module can determine its installation location using the BeiDou positioning system or a sonar laser measuring device. The power module 4 can be equipped with a propeller and a ballast water tank, which drives the installation module 2 and the leveling module to move.

[0047] In one embodiment, the underwater structure can be a submerged tube. The data processing unit can determine the maximum scouring and silting area through simulation analysis and engineering experience, calculate the coordinates of the predetermined installation position from the submerged tube location, and move the installation module 2 to the predetermined installation position using the power module 4. The power module 4 includes a small propeller and a ballast water tank. The small propeller drives the installation module 2, and the buoyancy and submersion of the system are achieved by the intake and drainage of water from the ballast water tank. After the installation module 2 moves to the predetermined installation position, the installation module 2 is activated, and the entire system is installed on the surface of the structure.

[0048] Optionally, the data processing module 3 is used to receive sensor data (siltation data) transmitted by the data transmission module, predict siltation trends based on a siltation model (such as an LSTM neural network), and generate early warning information based on the prediction results.

[0049] Optionally, the system can also be connected to an early warning and visualization platform that supports web and mobile access, and displays monitoring data, forecast curves and early warning levels in real time. Users can quickly understand the siltation situation and achieve rapid siltation treatment through the displayed information.

[0050] Alternatively, the sediment deposition model can be a machine learning model, which can be trained using an LSTM neural network.

[0051] In one embodiment, the sediment deposition model can be an LSTM model, which employs a multivariate LSTM time-series prediction structure, including an input layer, an LSTM hidden layer, a Dropout regularization layer, a fully connected layer, and an output layer. The input layer receives a multidimensional feature sequence composed of sensor data collected from multiple sources within a continuous time window; the LSTM hidden layer extracts the long-term and short-term dependencies of parameters such as flow velocity, sediment concentration, and deposition thickness over time; the Dropout layer randomly deactivates some neurons to reduce the risk of overfitting; the fully connected layer maps the high-dimensional time-series features extracted by the LSTM to the target prediction space; and the output layer outputs the deposition thickness, deposition rate, or erosion / deposition risk level for one or more future time points.

[0052] The multidimensional feature sequence of this model consists of feature parameters (obtained from sensor data) from multiple consecutive sampling times. These feature parameters include: (1) Hydrodynamic characteristics: instantaneous flow velocity, sliding average flow velocity, flow direction angle, and rate of change of water level; (2) Sediment characteristics: sediment volume concentration, mass concentration, concentration change rate, and cumulative sediment transport index; (3) Characteristics of scouring and sedimentation: sedimentation thickness, sedimentation rate, scouring depth, and volume change of scouring and sedimentation; (4) Structural response characteristics: pressure change value, stress fluctuation amplitude; (5) Auxiliary environmental characteristics: water temperature, particle size distribution parameters, bed slope and local topographic relief.

[0053] The sediment deposition model uses a sliding time window to organize the input samples. Let the input time step be T. Then, at the prediction time t, the input sequence formed by the input samples is:

[0054] in, Given the input sequence, This represents the vector composed of multidimensional feature parameters at the t-th sampling time.

[0055] Optionally, the output of the sedimentation model can be a single-step prediction or a multi-step rolling prediction; the output includes the sedimentation thickness within a preset future time period. siltation rate Changes in siltation volume And the corresponding risk level.

[0056] Before system deployment, a training sample set can be constructed using historical monitoring data, engineering survey data, and typical siltation disaster case data to train the LSTM model offline. During training, the sample data can be divided into a training set, a validation set, and a test set, with an optimal ratio of 80%:10%:10%. (The last part, "=0.1", appears to be an unrelated typo and can be omitted.)

[0057] Optionally, mean squared error can be used as the loss function during training, and the loss function is:

[0058] in, This represents the true value corresponding to the i-th test sample. Let N be the predicted value of the model for the i-th test sample, and N be the total number of test samples.

[0059] During training, the Adam optimization algorithm can be used to iteratively update the parameters of the sediment deposition model. The number of training epochs is 50–300, the batch size is 16–128, and the learning rate is 0.0001–0.01. Furthermore, an early stopping mechanism is introduced during training: when the validation set loss no longer decreases within a preset number of consecutive epochs, training is stopped and the model parameters with the best validation performance are retained. Simultaneously, Dropout and L2 regularization are combined to suppress overfitting.

[0060] Optionally, to reduce manufacturing costs, the entire system structure, except for essential components such as monitoring and transmission parts which use metal components, is constructed using high-performance composite materials. Furthermore, the structures of each module can be directly manufactured using 3D printing, further reducing production costs. The system's sealing interfaces employ IP68-level waterproofing technology, effectively preventing corrosion and withstanding certain water pressure, ensuring stable data transmission in harsh environments and avoiding data interruptions due to equipment damage.

[0061] Optionally, data processing module 3 can visually display the sedimentation data collected by the sensors, the sedimentation trend information predicted by the model, the working status of each module, and early warning information. This information can be displayed through charts, curves, map annotations, etc. Managers can use this information to quickly determine the system's operating status, sedimentation information, and early warning levels to quickly perform operations such as sedimentation location positioning and cause investigation.

[0062] Optionally, the data processing module 3 can also perform functions such as data query, statistical analysis, and automatic report generation according to the instructions of the management personnel to simplify the operation and maintenance management process and reduce the workload of the management personnel.

[0063] This application has the following advantages: (1) Achieve fully automated monitoring and significantly reduce manual intervention: Completely change the traditional model of siltation monitoring that relies on manual on-site sampling, measurement, and recording. Through real-time sensor acquisition, automatic system processing, and real-time data transmission, achieve real-time and continuous monitoring of siltation. No manual supervision is required, which can effectively avoid problems such as low efficiency, long cycle, data dispersion, and great influence from the environment in manual monitoring, reduce labor costs and operational risks, and at the same time ensure the continuity and timeliness of monitoring data, providing comprehensive data coverage for siltation analysis.

[0064] (2) Intelligent prediction and hierarchical early warning to enhance decision-making and response capabilities: Relying on advanced prediction models and big data analysis technology, it has the ability to accurately predict siltation in advance, and can predict risk trends before siltation disasters occur, breaking the passive situation of "post-disaster handling". Through the hierarchical early warning mechanism, the response strategies for different risk levels are clarified, providing sufficient decision-making time for engineering management departments, helping them to scientifically formulate dredging plans, adjust operating parameters, and activate emergency plans, thereby improving the ability to prevent and control siltation risks and response efficiency, and reducing disaster losses.

[0065] (3) Modular design adapts to diverse scenarios, and installation and maintenance are convenient and efficient: The system adopts a modular and standardized design concept. The interfaces of each component are unified and highly compatible. According to the monitoring needs of different engineering scenarios, the sensors, data transmission units and other components can be flexibly combined to achieve personalized configuration. At the same time, the modular structure greatly simplifies the installation, inspection and maintenance process. When a single module fails, it can be replaced independently without overall shutdown, reducing the difficulty of operation and maintenance and downtime losses, and adapting to complex and ever-changing engineering application scenarios.

[0066] (4) Enhanced Operation and Maintenance Efficiency through Visualization Platform: The supporting visualization management platform displays real-time monitoring data, characteristic parameters, prediction results, and early warning information in intuitive forms such as charts, curves, and map annotations, making the data information readily apparent. Managers can remotely monitor the system's operating status, sludge dynamics, and early warning situations in real time through the platform, quickly locate the problem location, and analyze the causes of sludge. At the same time, the platform supports functions such as data query, statistical analysis, and automatic report generation, simplifying the operation and maintenance management process and improving the level of refinement and intelligence in management.

[0067] (5) Provide scientific data support and extend the service life of structures: The real-time monitoring data and predictive analysis results accumulated by the system can provide accurate scientific basis for engineering design, construction and subsequent maintenance. In the design stage, the engineering structure layout can be optimized based on data and reasonable anti-siltation schemes can be formulated; in the construction stage, siltation can be monitored in real time and construction technology and progress can be adjusted; in the maintenance stage, dredging cycles and schemes can be accurately formulated to avoid structural damage caused by excessive dredging or untimely dredging. Through scientific management, the erosion and load of siltation on engineering structures can be effectively reduced, the service life of buildings can be extended and the overall operating cost of the project can be reduced.

[0068] According to one aspect of the embodiments of this application, a method for monitoring sediment deposition in underwater structures based on multi-source sensing is provided. This method is used in the monitoring system described in the above embodiments. Figures 1-5 As shown, the monitoring method includes: S101: Use sensor module 1 to acquire siltation data and preprocess the siltation data.

[0069] Optionally, the sediment deposition data includes water flow velocity, sediment concentration, sediment deposition thickness, water level, and flow direction. Various types of sensors in the sensing module can be used to collect multi-dimensional data such as water flow velocity, sediment concentration, sediment deposition thickness, water level, and flow direction in real time. The acquisition frequency can be flexibly set according to requirements, achieving continuous data capture throughout the entire time period.

[0070] Optionally, the sediment deposition data collected by the sensor can be transmitted to the data processing unit via the data transmission module. Alternatively, the data processing unit can be directly connected to the sensor to obtain the sediment deposition data collected by the sensor.

[0071] Optionally, when the data processing module 3 preprocesses the siltation data, it can remove invalid data such as environmental interference and equipment noise through filtering algorithms, and supplement missing data by methods such as linear interpolation and nonlinear fitting.

[0072] In one embodiment, after receiving sediment deposition data, the data processing module 3 removes invalid data such as environmental interference and equipment noise using a filtering algorithm. For missing data, adaptive completion is performed based on the duration of the missing data and the correlation status of multiple source sensors: for short-term missing data, linear interpolation or spline interpolation of adjacent time points is used for completion; for long-term missing or nonlinear fluctuation data, nonlinear fitting or time series prediction methods based on historical time windows are used for reconstruction; for data missing due to the failure of a single sensor, flow velocity, flow direction, sediment concentration, water level, and pressure data collected by other sensors are combined, and cross-compensation is performed using multivariate regression models or machine learning models to achieve the reconstruction of missing data. For data segments that cannot be reliably completed, the system marks them as missing and reduces their weight in subsequent analysis. After completion, data standardization and normalization are performed to ensure the accuracy, completeness, and consistency of the data, providing high-quality data support for subsequent analysis.

[0073] S102: Extract feature parameters from the sediment deposition data, input the feature parameters and preprocessed sediment deposition data into the pre-trained sediment deposition model, and obtain sediment deposition trend prediction information.

[0074] Optionally, techniques such as statistical analysis and machine learning feature engineering can be used to accurately extract key feature parameters closely related to sediment deposition. Specifically, key feature parameter extraction can be performed based on preprocessed sediment deposition data. First, time-domain statistical analysis was performed on the original time-series data, including flow velocity, flow direction, sediment concentration, water level, pressure, and sediment thickness, to extract basic features such as instantaneous values, mean, extreme values, variance, rate of change, fluctuation amplitude, and cumulative amount within a sliding time window. Second, the underwater surface morphology data acquired by sonar sedimentation sensors and laser scanners were spatially reconstructed to extract morphological features such as sediment thickness, scour depth, sediment area, scour volume, slope, and topographic gradient. Furthermore, by combining the spatiotemporal correspondence between flow velocity, flow direction, sediment concentration, and sediment thickness, coupled characteristic parameters characterizing the sediment transport and local scour evolution mechanism were constructed, including sediment transport intensity per unit time, sediment growth rate, flow velocity-concentration coupling coefficient, and hydrodynamic index. Finally, correlation analysis, principal component analysis, recursive feature elimination, or tree-based feature importance evaluation methods were used to screen the extracted candidate features, retaining key input variables highly correlated with the sediment deposition prediction target, and constructing a feature input set for the prediction model, thereby improving the model's prediction accuracy and stability. The feature parameters include water flow velocity (instantaneous velocity, average velocity), sediment concentration (volume concentration, mass concentration), and sedimentation thickness (cumulative thickness, sedimentation rate). At the same time, auxiliary feature parameters such as water temperature, hydrodynamic force, and sediment particle size distribution in the water flow are also taken into account to construct a comprehensive feature system, providing effective input variables for the prediction model and improving the model's prediction accuracy.

[0075] Optionally, training the sedimentation model includes: constructing a training sample set based on historical data, including historical monitoring data, engineering survey data, and sedimentation disaster case data; determining the model to be trained, training the model using the training sample set, and adjusting the model's hyperparameters according to the model type; and obtaining the sedimentation model based on the adjusted hyperparameters and the model training results.

[0076] Optionally, engineering survey data may include water flow velocity, sediment concentration, sediment gradation, and annual riverbed siltation.

[0077] Alternatively, the sediment deposition model can be a neural network model, a support vector machine model, a regression analysis model, or other types of machine learning models.

[0078] Alternatively, cross-validation, grid search, or other methods can be used to adjust the model hyperparameters.

[0079] In one embodiment, when using a Long Short-Term Memory (LSTM) neural network as a sediment deposition model, the hyperparameters to be adjusted include the input time step, the number of hidden layers, the number of hidden units per layer, the learning rate, the batch size, the number of training epochs, and the Dropout ratio. The input time step characterizes the length of historical monitoring sequences utilized by the model; the number of hidden units controls the model's ability to express temporal features; the learning rate controls the gradient update step size; and the Dropout ratio suppresses overfitting. The adjusted hyperparameters can be: setting the input time step to 6–48 sampling steps, the number of hidden layers to 1–4, the number of hidden units per layer to 16–256, the learning rate to 0.0001–0.01, the batch size to 16–128, and the Dropout ratio to 0–0.5. The optimal configuration is obtained through parameter combination search to reduce the risk of overfitting and underfitting, and improve the model's generalization ability and prediction reliability.

[0080] Optionally, after obtaining the sedimentation model, the preprocessed real-time monitoring data and extracted feature parameters can be input into the trained and optimized prediction model. The model combines historical data patterns with real-time operating conditions to quickly output the sedimentation trend over a future period (from a few hours to several months), including key information such as sedimentation location, sedimentation rate, and cumulative sedimentation volume.

[0081] Optionally, to improve model accuracy, when the model is not outputting results, only short-term (1-7 days) predictions are obtained using the sedimentation model, and these predictions are compared in real-time with the measured data obtained in the next step. The prediction model is then adjusted in real-time based on the measured data. When output results are needed, the desired forecast time can be directly input, and the data processing module 3 will make a forecast based on the latest model output. It will also intuitively output underwater sedimentation prediction topographic maps, sedimentation thickness maps, and other information based on the sedimentation information to obtain the most accurate prediction results.

[0082] S103: Issue early warnings based on siltation trend predictions and actual siltation data.

[0083] Optionally, early warnings can be issued based on siltation trend prediction information and actual siltation data, including: obtaining the predicted siltation height and the actual siltation height corresponding to the actual siltation data based on the siltation trend prediction information; and issuing early warnings based on the distance between the actual siltation height, the predicted siltation height, and the predetermined location of the building.

[0084] Optionally, the predetermined location can be the bottom of a building or other locations. The actual distance and predicted distance between the sediment and the predetermined location can be obtained based on the actual siltation height and the predicted siltation height. If at least one of the actual distance and the predicted distance meets a preset condition (such as being less than a preset distance), it is determined that an early warning needs to be issued.

[0085] In one embodiment, multiple warning thresholds (mild warning, moderate warning, severe warning) can be preset in the 1-7 day forecast results output by the model in real time, and a comprehensive judgment can be made by combining the forecast results with the actual siltation situation. Specifically, when the structure is an immersed tube, the distance between the top of the silt and the bottom of the immersed tube section can be calculated based on the siltation height. For example, an actual distance of <10cm is a toxicity warning, an actual distance of 10-20cm is a moderate warning, and 20-500cm is a mild warning.

[0086] The prediction results can include the siltation height 7 days later. Taking a siltation depth of 20cm from the top of the siltation to the bottom of the pipe section as an example, a severe warning is issued if the predicted depth reaches 20cm within 2 days, a moderate warning within 5 days, and a mild warning within 7 days. An alarm will be triggered if either the predicted result or the actual siltation situation is met; if both are met, the alarm intensity is stronger. When the corresponding warning threshold is reached, the tiered warning mechanism is automatically triggered. Warning information is simultaneously released through multiple channels, including a visualization platform, SMS, email, and audible and visual alarms, clearly informing users of the warning level, scope of impact, potential risks, and response suggestions. It can also be linked with the engineering scheduling system and operation and maintenance management platform to support rapid decision-making and emergency response, minimizing the impact of siltation disasters.

[0087] The terms "first," "second," "third," "fourth," "1," "2," etc. (if present) in the specification, claims, and accompanying drawings of this application are used to distinguish similar objects and are not necessarily used to describe a specific order or sequence. It should be understood that such data can be interchanged where appropriate so that the embodiments of this application described herein can be implemented in a sequence other than that shown in the illustrations or text descriptions.

[0088] It should be understood that although arrows indicate various operation steps in the flowcharts of this application's embodiments, the order in which these steps are implemented is not limited to the order indicated by the arrows. Unless explicitly stated herein, in some implementation scenarios of this application's embodiments, the implementation steps in each flowchart can be executed in other orders as required. Furthermore, some or all steps in each flowchart, based on the actual implementation scenario, may include multiple sub-steps or multiple stages. Some or all of these sub-steps or stages can be executed at the same time, and each sub-step or stage can also be executed at different times. In scenarios where execution times differ, the execution order of these sub-steps or stages can be flexibly configured according to requirements, and this application's embodiments do not limit this.

[0089] The above description is only an optional implementation method for some implementation scenarios of this application. It should be noted that for those skilled in the art, other similar implementation methods based on the technical concept of this application without departing from the technical concept of this application also fall within the protection scope of the embodiments of this application.

Claims

1. A multi-source sensing-based underwater structure sediment deposition monitoring system, characterized in that, include: The sensor module has sensors deployed around or below the building. The sensor module is used to collect data on siltation around the building. The sensors include at least two of the following: sonar siltation sensor, laser scanner, flow velocity and direction sensor, shear force sensor, silt concentration sensor, and pressure sensor. A data transmission module, which is connected to the sensor module, is used to transmit the siltation data; The data processing module is used to receive the sedimentation data transmitted by the data transmission module, predict the sedimentation trend based on the pre-trained sedimentation model and the sedimentation data, and issue an early warning based on the prediction results.

2. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 1, characterized in that, It includes an installation module, the sensor module and the data transmission module are connected to the installation module and fixed to the building through the installation module.

3. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 2, characterized in that, The mounting module can be fixed by at least one of negative pressure, magnetic attraction, bolts, and clips, and the fixing method of the mounting module corresponds to the installation environment and the surface material of the building.

4. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 3, characterized in that, The installation module includes a suction cup and a vacuum pump. The vacuum pump is connected to the suction cup and is used to evacuate the suction cup to fix the installation module after the suction cup is attached to the building.

5. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 2, characterized in that, The system includes a leveling module, on which the sensor is fixed. One side of the leveling module is connected to the mounting module. The leveling module is used to adjust the orientation of the sensor.

6. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 5, characterized in that, The leveling module includes a housing and a telescopic unit, and the sensor is disposed on the side of the housing away from the mounting module. The telescopic unit is located on the side of the housing away from the sensor, and includes a telescopic column and a telescopic rod. The telescopic column is used to adjust the height of the housing, and the telescopic rod is used to adjust the housing to a horizontal state.

7. The underwater structure sediment deposition monitoring system based on multi-source sensing according to claim 6, characterized in that, It includes a power module and a positioning module, wherein the power module and the positioning module are mounted on the housing; The positioning module is used to obtain current positioning information; The power module is used to move the installation module to the predetermined installation position according to the positioning information during installation.

8. A method for monitoring sediment deposition in underwater structures based on multi-source sensing, characterized in that, For the underwater structure sediment deposition monitoring system based on multi-source sensing as described in any one of claims 1-7, the method comprises: The sensor module is used to acquire sediment deposition data, and the sediment deposition data is preprocessed. The sediment deposition data includes water flow velocity, sediment concentration, sediment deposition thickness, water level, and flow direction. Extract the feature parameters from the sediment deposition data, and input the feature parameters and the preprocessed sediment deposition data into the pre-trained sediment deposition model to obtain sediment deposition trend prediction information; Early warnings are issued based on the predicted siltation trend information and actual siltation data.

9. The underwater structure sediment deposition monitoring method based on multi-source sensing according to claim 8, characterized in that, The training of the sediment deposition model includes: A training sample set is constructed based on historical data, including historical monitoring data, engineering survey data, and siltation disaster case data. The model to be trained is determined, the model is trained using the training sample set, and the hyperparameters of the model are adjusted according to the type of the model. The sediment deposition model is obtained based on the adjusted hyperparameters and model training results.

10. The underwater structure sediment deposition monitoring method based on multi-source sensing according to claim 8, characterized in that, The early warning based on the siltation trend prediction information and actual siltation data includes: Based on the siltation trend prediction information, the predicted siltation height and the actual siltation height corresponding to the actual siltation data are obtained; Early warnings are issued based on the actual siltation height, the predicted siltation height, and the distance between the building's predetermined location.

Images