A water conservancy flood control pipe intelligent siltation detection method based on physical rule guidance

Abstract

This invention discloses an intelligent siltation detection method for water conservancy and flood control pipelines guided by physical rules, belonging to the field of water conservancy engineering and municipal pipeline maintenance technology. This invention constructs a high-dimensional heterogeneous sensing spatiotemporal tensor field based on Riemannian manifold geometry, utilizes a physical information neural network combined with the Herschel-Bulkley constitutive equation to invert the non-Newtonian fluid rheological properties of the silt, and then integrates robot multibody dynamics, Rayleigh-Plesset cavitation dynamics, and a solid-liquid two-phase flow critical velocity model for parameter optimization. Finally, it achieves global optimal control of the dredging operation by solving the Hamilton-Jacobi-Bellman equation, effectively solving the problems of difficult pipeline siltation diagnosis, low dredging efficiency, and poor operational stability in complex flow field environments.

Description

Technical Field

[0001] This invention belongs to the field of water conservancy engineering and municipal pipeline maintenance technology, and particularly relates to an intelligent siltation detection method for water conservancy flood control pipelines based on physical rules. Background Technology

[0002] Currently, using robots to replace manual labor for pipeline dredging has become an industry consensus. However, in practical engineering applications, existing dredging equipment and methods still face the following significant technical bottlenecks:

[0003] First, the robot's motion mechanism lacks adaptability and stability to complex pipe shapes. Existing pipeline dredging robots mostly employ a single-unit rigid chassis structure. When facing curved pipes with small radii of curvature or unstructured terrain, this single-unit structure is prone to jamming or tipping over. Furthermore, during high-pressure water jetting or powerful mechanical breaking operations, the robot body is easily subjected to enormous recoil forces. The lack of an effective multi-body dynamics stability control mechanism makes it difficult to maintain the robot's operating posture, severely impacting dredging accuracy.

[0004] Second, the perception of sedimentation lacks physical mechanism support, making it difficult to achieve accurate inversion of rheological properties. Existing sedimentation detection technologies mostly rely on simple geometric measurements (such as thickness and volume) of visual or acoustic data, primarily processing data in Euclidean space. However, the inner wall of a pipe has typical non-Euclidean geometric features, and sediment often exhibits complex non-Newtonian fluid properties (such as yield stress and shear thinning). Traditional pure data-driven deep learning algorithms ("black box" models) lack the constraints of rheological physical equations (such as the Herschel-Bulkley constitutive equation), making it impossible to accurately distinguish whether the sediment is in a "sludge-like" or "hard-shell" state. This leads to a lack of basis for setting dredging parameters, easily resulting in phenomena such as "excessive breaking of soft sludge, wasting energy" or "hard shell, which cannot be cleared."

[0005] Third, dredging operations lack coordinated optimization and closed-loop control involving multiple physical fields. Existing dredging operations often employ open-loop control, meaning that once fixed ultrasonic power, jet pressure, and suction speed are set, no further adjustments are made. However, the dredging process involves strong coupling of multiple physical processes, including acoustics (ultrasonic cavitation), fluid dynamics (jet impact), and solid-liquid two-phase flow (slurry transport). For example, the matching of ultrasonic frequency and jet pressure directly determines the collapse efficiency of cavitation bubbles (Rayleigh-Plesset dynamics), while if the suction speed is below the critical flow velocity (Durand-Condolios theory), it can easily cause blockage of the sewage pipe. Current technologies lack a globally optimal control algorithm (such as a control law based on the HJB equation) that can unify and integrate the above physical dynamic models, making it impossible to achieve intelligent adaptive closed-loop control from perception to execution.

[0006] In summary, the key to solving the above problems lies in developing a physical knowledge-guided intelligent dredging equipment and method for water conservancy and flood control pipelines that adopts a split adaptive structure and integrates physical information neural network (PINN), cavitation dynamics, and global optimal control theory.

[0007] Patent application CN120291607A discloses a suction-type intelligent dredging robot and dredging method for underground drainage pipes. It can automate and intelligently diagnose and dredge siltation in underground drainage pipe networks within a certain maintenance area. However, the robot relies heavily on algorithms and mechanically breaks up pipes using an electric hammer, without further algorithmic improvements to enhance operational safety at pipe defects. Furthermore, the robot's single-unit structure makes it less adaptable to bends in pipes compared to a modular design. Summary of the Invention

[0008] To address the problems of difficulty in diagnosing pipeline siltation, low dredging efficiency, and poor operational stability in existing technologies under complex flow field environments, this invention provides an intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules.

[0009] The present invention is achieved through the following technical solutions.

[0010] This invention provides an intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules, comprising the following steps:

[0011] S1, the deployment phase: the robot is placed into the target water conservancy and flood control pipeline through the rainwater well, and the operator issues the operation command through the ground terminal to drive the robot to start.

[0012] S2, Model Building Stage: The robot performs flow velocity detection. The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines combines the design parameters of water conservancy and flood control pipelines with real-time flow field data to adjust the robot to its working posture. At the same time, it scans the water conservancy and flood control pipelines to initially construct a three-dimensional distribution map of siltation in the water conservancy and flood control pipelines.

[0013] S3, Intelligent Diagnosis Stage: Run the intelligent diagnosis algorithm for siltation in water conservancy and flood control pipelines, input the data collected by the operation robot, and combine it with the reaction kinetic model of the chemical ablation reagent to calculate the optimal operation parameters;

[0014] S4, parameter adjustment stage: adjust the working parameters of the robot according to the optimal working parameters obtained by the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines.

[0015] S5, the collaborative dredging execution phase, involves using a robot to spray chemical ablation reagents to soften the silt, ultrasonically crush the silt, and assist in stirring and pulverizing the silt. The robot then pumps and transports the crushed silt.

[0016] S6, Dynamic Adaptation Stage: During the operation, the flow field changes in the water conservancy and flood control pipeline are monitored in real time. If the flow velocity changes abruptly, the posture of the operating robot and the suction power of the silt are adjusted. The condition of the inner wall of the water conservancy and flood control pipeline is also detected. If new silt is found, the new silt is targeted for dredging, and the dredging of a section of the water conservancy and flood control pipeline is completed.

[0017] S7, Secondary Inspection Phase: After completing the dredging of a section of the flood control pipeline, the operation robot scans the undredged section of the flood control pipeline again. The intelligent siltation diagnosis module of the flood control pipeline compares the data before and after dredging. If the siltation residue rate is high, secondary dredging is initiated; if the residue rate is low, the operation robot moves to the next section of the flood control pipeline and repeats the process of S2-S6.

[0018] S8, the shutdown and recovery phase: After the entire section of the water conservancy and flood control pipeline is dredged, the intelligent diagnosis module for siltation in the water conservancy and flood control pipeline generates a dredging report and transmits the report to the ground terminal. The operator controls the operation robot to return to the rainwater well, and finally shuts down the operation robot to recover the dredging equipment and the stored silt.

[0019] Preferably, the intelligent diagnostic stage of S3 includes the following steps:

[0020] S31: Construct a high-dimensional heterogeneous sensing spatiotemporal tensor field based on Riemannian manifold geometry; construct a system state tensor on Riemannian manifold space based on the non-Euclidean geometric features of the inner wall of a hydraulic flood control pipeline. First, using the point cloud data collected by the robot, a manifold metric tensor is defined to establish a body-orthogonal curvilinear coordinate system for the water conservancy and flood control pipeline. Secondly, the flow field velocity vector collected by the robot is... , Sediment thickness field and acoustic impedance spectrum Spatiotemporal registration and dimensionless processing are performed to generate the input feature tensor. :

[0021] ;

[0022] in, for The input feature tensor at time step; This is to iterate over any specific moment within the time window from the start time to the current time; For time-series feature aggregation operators, it means that from Time's up A collection of time window data at any given moment; for The velocity vector of the flow field acquired by the time-of-flight Doppler flow velocity sensor; The maximum modulus of the flow velocity; For multimodal feature splicing operators; for The thickness field of the sedimentary layer collected by the constantly operating robot; The diameter of the water conservancy and flood control pipeline; Impedance spectrum data; The standard acoustic impedance of water is used as the reference value for dimensionless processing. The six-DOF attitude vectors of the robot are provided by a three-axis gyroscope;

[0023] S32: Inversion of non-Newtonian fluid rheological properties based on a physical information neural network; The feature tensor generated in S31 is used... The input is fed into a physical information neural network, which introduces the Herschel-Bulkley non-Newtonian fluid constitutive equation as a physical constraint layer, and minimizes the hybrid loss function. To invert the true rheological parameters of the sediment;

[0024] S33: Construct operational stability boundary constraints incorporating the dynamics of the operational robot; the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines introduces multibody dynamics equations to perform overall modeling of the operational robot and calculate the anti-overturning stability domain under complex flow fields. Establish the Lagrange dynamic equations:

[0025] ;

[0026] In the formula, For the generalized mass matrix of the robot; A generalized coordinate vector representing the spatial displacement and attitude angle of a robot. For generalized velocity vectors, The first derivative; It is a generalized acceleration vector. The second derivative; The matrix represents the Coriolis force and the centrifugal force. Here is the nonlinear stiffness matrix of hose 26; It is an external generalized force vector, which includes the omnidirectional high-pressure nozzle recoil force and the crusher head excitation force; Let this be the generalized gravity vector of the system; For the transpose of the contact Jacobian matrix; The contact force between the robot and the inner wall of the pipe;

[0027] In the aforementioned Lagrange dynamic equations, the generalized coordinate vector The spatial displacement data is obtained based on the displacement distance of the drive wheel 2; the attitude angle data is provided by the three-axis gyroscope 12 in real time by sensing the six degrees of freedom motion state of the robot.

[0028] S34: Optimization of breakup parameters based on Rayleigh-Plesset cavitation dynamics; Establishment of an acoustic-fluid coupled cavitation efficiency model for the coordinated operation of ultrasonic breakup of sediment by a robot and a nozzle; Introduction of Rayleigh-Plesset equations to describe the dynamic behavior of microbubbles under the combined action of ultrasound and jet.

[0029] The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines finds the optimal ultrasonic frequency by numerically solving the above differential equations. With jet pressure This causes the peak pressure of the microjets generated when the cavitation bubbles collapse to reach its maximum value. Maximize, and satisfy:

[0030] ;

[0031] In the formula, The internal bonding strength of the silt;

[0032] S35: Transfer efficiency control based on critical velocity of solid-liquid two-phase flow; for sediment transfer systems, a critical velocity control model based on Durand-Condolios theory is established to prevent blockage of the sediment discharge pipe used by the robot for sediment suction and transport; the critical non-sludge velocity is calculated. :

[0033] ;

[0034] Durand's empirical coefficient; It is the acceleration due to gravity; The diameter of the pipe of the sludge suction device used by the robot for sludge suction and transportation; The density of the solid particles in the sediment; The density of the fluid used to transport sediment;

[0035] The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines is based on the real-time mixing density of silt at the inlet of the silt extractor. Dynamically adjust the negative pressure power of the vacuum sewage pump Ensure the actual flow rate within the sediment extractor pipe. Always within the safe zone:

[0036] ;

[0037] In the formula, For safety margin coefficient, Maximum flow rate limit to prevent excessive wear of the sediment suction pipe;

[0038] S36: Global optimal control closed loop based on Hamilton-Jacobi-Bellman equations; couple the subsystems in S31 to S35 to construct the Hamiltonian function of the whole system. Solve for the global energy-optimal control law:

[0039] ;

[0040] In the formula, It is an energy functional; This is the system state vector; To control the input vector; The transpose of the costate variable vector represents the shadow price of the system state change;

[0041] The system dynamics equations describe the rate of change of the system state over time;

[0042] ;

[0043] In the formula, It is an energy functional; To optimize the weighting coefficients of the objective; The rate of removal of sediment per unit time; This represents the total energy consumption of the system. This is the current control input vector; To smooth the control reference trajectory and prevent drastic fluctuations in the control input; These are the start and end times of the task.

[0044] Preferably, in S34, the mixture loss function is minimized. To invert the true rheological parameters of the sediment:

[0045] ;

[0046] in, These are the weighting coefficients for data-driven items; The flow field state or rheological parameters predicted by the neural network; The actual data observed by the sensor; These are the weighting coefficients for the physical constraint terms; For incompressible fluid Navier-Stokes dynamic residuals; The rheological physical residual is used to determine the hardness of sediment and whether it is in the form of soft mud or hard crust. Its definition is as follows:

[0047] ;

[0048] In the formula, For the shear stress tensor, Shear rate, To retrieve the yield stress of the sedimentary deposits from the inversion output, This is the consistency coefficient. The rheological index;

[0049] at the same time, For incompressible fluid Navier-Stokes dynamic residuals, ensure that the predicted flow field conforms to momentum conservation:

[0050] ;

[0051] in, The fluid velocity vector is denoted by g; g is the acceleration due to gravity. This represents the time corresponding to the current momentum conservation equation for the flow field environment. The density of the sediment fluid; For fluid pressure; Unit tensor; ∠ is the viscous stress tensor; ▽ is the bitwise operation symbol;

[0052] Preferably, in step S33, the Rayleigh-Plesset equation is introduced to describe the dynamic behavior of microbubbles under the combined action of ultrasound and jet, and the calculation formula is as follows:

[0053] ;

[0054] In the formula, The instantaneous radius of the cavitation bubble; Let be the first derivative of the bubble radius with respect to time, representing the velocity. Let be the second derivative of the bubble radius with respect to time, representing acceleration; The density of the sediment fluid; This is the saturated vapor pressure inside the liquid; Environmental pressure at infinity; The additional static pressure generated by the jet; The surface tension coefficient of the liquid; The dynamic viscosity of the liquid.

[0055] Preferably, in step S36, the optimal control vector is obtained by solving the Hamilton-Jacobi-Bellman partial differential equation:

[0056] ;

[0057] In the formula, The optimal ultrasonic frequency; To achieve the optimal jet pressure; To achieve the optimal stirring speed; For optimal negative pressure power, This is the transpose symbol.

[0058] Preferably, the collaborative dredging execution phase of S5 includes the following steps:

[0059] S51: Silt Softening: Based on the three-dimensional distribution map of siltation in water conservancy and flood control pipelines, the intelligent diagnosis module for siltation in water conservancy and flood control pipelines controls the swing angle of the nozzle of the operation robot and the extension stroke of its hydraulic telescopic nozzle to spray chemical ablation reagent onto the siltation surface and crevices; the nozzle adopts a fan-shaped spray pattern to complete the silt softening;

[0060] S52: Ultrasonic Breaking: After the silt softening is completed, the ultrasonic hammer of the working robot outputs high-frequency vibration to break the softened silt block; and the breaking effect is monitored in real time. If large residual silt blocks are detected, the intelligent diagnosis module for siltation in water conservancy and flood control pipelines will extend the breaking time and increase the ultrasonic power of the ultrasonic hammer.

[0061] S53: Mixing Assistance: The robot mixes the broken-up silt and coagulated blocks with clean water to form a slurry, preventing particle sedimentation; combined with the flaw detection data of the robot, it enables the robot to avoid defective areas in the water conservancy and flood control pipeline during operation;

[0062] S54: Silt Extraction: The silt transfer system is activated. Based on the flow field data and the concentration of the stirred sludge, the intelligent silt diagnosis module of the water conservancy and flood control pipeline adjusts the negative pressure power of the vacuum pump. The sludge-like silt is extracted from the water conservancy and flood control pipeline through the suction port of the silt extractor. The nozzle of the operation robot continuously sprays clean water, forming a pushing water flow in the water conservancy and flood control pipeline, which helps the sludge to converge towards the suction port of the silt extractor and avoids silt residue. The extracted sludge is output through the silt discharge pipe, completing the dredging.

[0063] The beneficial effects of this invention are as follows:

[0064] This invention constructs a high-dimensional heterogeneous sensing spatiotemporal tensor field, combines physical information neural networks to invert the characteristics of silt deposits, and integrates robot multibody dynamics, cavitation dynamics, and critical flow velocity models for parameter optimization. This effectively solves the problems of difficult pipeline silt diagnosis, low dredging efficiency, and poor operational stability in complex flow field environments, and achieves global optimal control of dredging operations. Attached Figure Description

[0065] Figure 1 This is a flowchart of an embodiment of the present invention;

[0066] Figure 2 This is a connection diagram of the adaptive robot and the operation control box provided in this embodiment of the invention;

[0067] Figure 3This is an assembly diagram of the adaptive operation robot provided in an embodiment of the present invention;

[0068] Figure 4 This is a side view of the adaptive work robot provided in an embodiment of the present invention;

[0069] Figure 5 This is a bottom view of the adaptive work robot provided in this embodiment of the invention;

[0070] Figure 6 This is a front view of the adaptive work robot provided in an embodiment of the present invention;

[0071] Figure 7 This is a rear view of the adaptive work robot provided in an embodiment of the present invention;

[0072] Figure 8 This is a hardware structure diagram of the operation control box provided in the embodiment of the present invention;

[0073] In the diagram, 1-telescopic support arm, 2-drive wheel, 3-driven wheel, 4-multi-channel ultrasonic crushing rotating head, 5-high pressure nozzle, 6-stirring brush, 7-sludge suction device, 8-sludge discharge pipe, 9-vacuum pump, 10-Doppler flow velocity sensor, 11-central command controller, 12-three-axis gyroscope, 13-signal interface, 14-annular sliding groove, 15-ultrasonic thickness sensor, 16-lidar, 17-detection component, 18-box structure, 19-box cover, 20-power supply device, 21-non-contact ultrasonic level gauge, 22-wireless signal transceiver module, 23-signal antenna, 24-cable, 25-development board, 26-hose, 27-front vehicle, 28-rear vehicle. Detailed Implementation

[0074] The technical solution of the present invention is further described below, but the scope of protection is not limited to what is described.

[0075] Example:

[0076] like Figures 1 to 8 As shown, a smart siltation detection method for water conservancy and flood control pipelines based on physical rules includes the following steps:

[0077] S1, the deployment phase: the robot is placed into the target water conservancy and flood control pipeline through the rainwater well. The operator issues the operation command through the ground terminal. The operation control box, which is placed on the ground and fixed, receives the operation control command and the fluid state index data transmitted by the cable 24. The operation control box then transmits the signal to drive the robot's motor system to start.

[0078] S2, Model Building Stage: The ultrasonic flaw detection sensor and pressure sensor are activated, the Doppler flow velocity sensor 10 starts flow velocity detection, and the intelligent diagnosis algorithm for siltation in water conservancy and flood control pipelines combines the design parameters (pipe diameter, slope) and real-time flow field data of water conservancy and flood control pipelines to adjust the telescopic support arm 1 so that the operating robot is in the best operating posture. At the same time, the multi-channel ultrasonic crushing rotating head 4 drives the ultrasonic thickness measuring sensor 15 and the lidar 16 to start rotating and scanning the water conservancy and flood control pipelines, and initially constructs a three-dimensional distribution map of siltation in water conservancy and flood control pipelines.

[0079] S3, Intelligent Diagnosis Stage: Development Board 25 runs the intelligent diagnosis algorithm for siltation in water conservancy and flood control pipelines, inputs the data collected by the operation robot, and calculates the optimal operation parameters by combining the reaction kinetic model of the chemical ablation reagent.

[0080] S4, Parameter Adjustment Stage: Based on the optimal operating parameters obtained by the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines, adjust and set the ultrasonic frequency of the multi-channel ultrasonic crushing rotating head 4, the jet pressure of the high-pressure nozzle 5, the stirring speed of the stirring brush 6, and the negative pressure power of the vacuum pump 9.

[0081] S5, the collaborative dredging execution phase, uses a high-pressure nozzle 5 to spray chemical ablation reagent to soften the silt, uses a multi-channel ultrasonic crushing rotating head 4 to ultrasonically crush the silt, uses a stirring brush 6 to assist in stirring and crushing the silt, and uses a silt transfer system to pump and transport the crushed silt.

[0082] S6, Dynamic Adaptation Stage: During the operation, the flow field changes inside the water conservancy and flood control pipeline are monitored in real time. If the flow velocity changes abruptly, the posture of the robot and the suction power of the sediment transfer system are adjusted. The condition of the inner wall of the water conservancy and flood control pipeline is detected by ultrasonic flaw detection sensors and pressure sensors. If new deposits are found, the high-pressure nozzle 5 and the multi-channel ultrasonic crushing rotating head 4 are triggered to start, and the new deposits are targeted for dredging, thus completing the dredging of a section of the water conservancy and flood control pipeline.

[0083] S7, Secondary Inspection Stage: After dredging a section of the flood control pipeline, the ultrasonic thickness sensor 15 and the lidar 16 scan the undragged section of the flood control pipeline again. The intelligent siltation diagnosis module of the flood control pipeline compares the data before and after dredging. If the siltation residue rate is high, secondary dredging is initiated; if the residue rate is low, the control system is activated to move the robot to the next section of the flood control pipeline and repeat the process of S2-S6.

[0084] S8, during the shutdown and recovery phase, after the entire section of the water conservancy and flood control pipeline has been dredged, the intelligent diagnosis module for siltation in the water conservancy and flood control pipeline generates a dredging report, which is transmitted to the ground terminal via the wireless signal transceiver module 22. The operator controls the work robot to return to the rainwater well, and the high-pressure nozzle 5 switches to the clean water spraying mode to clean the multi-channel ultrasonic crushing rotating head 4, the stirring brush 6, and the suction port of the silt suction device 7 to prevent siltation from affecting the next use. Finally, the work robot is shut down, and the dredging equipment and stored silt are recovered.

[0085] The stirring brush 6 is a metal stirring brush, and the wireless signal transceiver module 22 is a 5G wireless signal transceiver module.

[0086] The intelligent diagnostic phase of S3 includes the following steps:

[0087] S31: Construct a high-dimensional heterogeneous sensing spatiotemporal tensor field based on Riemannian manifold geometry; the development board 25 sends data collected by the robot's sensors through the signal interface 13, and constructs the system state tensor on the Riemannian manifold space based on the non-Euclidean geometric features of the inner wall of the water conservancy flood control pipeline. First, using point cloud data collected by LiDAR 16, a manifold metric tensor is defined to establish a body-orthogonal curvilinear coordinate system for the water conservancy and flood control pipeline. Secondly, the flow field velocity vector collected by the Doppler flow velocity sensor 10 The thickness field of the sediment layer collected by the ultrasonic thickness sensor 15 And the acoustic impedance spectrum acquired by the ultrasonic flaw detection sensor and pressure sensor array Spatiotemporal registration and dimensionless processing are performed to generate the input feature tensor. :

[0088] ;

[0089] in, for The input feature tensor at time step; This is to iterate over any specific moment within the time window from the start time to the current time; For time-series feature aggregation operators, it means that from Time's up A collection of time window data at any given moment; for The velocity vector of the flow field acquired by the time-of-flight Doppler flow velocity sensor; The maximum modulus of the flow velocity; For multimodal feature splicing operators; for The thickness field of the sedimentary layer collected by the constantly operating robot; The diameter of the water conservancy and flood control pipeline; Impedance spectrum data; The standard acoustic impedance of water is used as the reference value for dimensionless processing. The six-DOF attitude vectors of the robot are provided by a three-axis gyroscope;

[0090] S32: Inversion of non-Newtonian fluid rheological properties based on Physical Information Neural Network (PINN); The feature tensor generated in S31... The input is fed into a physical information neural network, which introduces the Herschel-Bulkley non-Newtonian fluid constitutive equation as a physical constraint layer, and minimizes the hybrid loss function. To invert the true rheological parameters of the sediment:

[0091] ;

[0092] in, These are the weighting coefficients for data-driven items; The flow field state or rheological parameters predicted by the neural network; The actual data observed by the sensor; These are the weighting coefficients for the physical constraint terms; For incompressible fluid Navier-Stokes dynamic residuals; The rheological physical residual is used to determine the hardness of sediment and whether it is in the form of soft mud or hard crust. Its definition is as follows:

[0093] ;

[0094] In the formula, For the shear stress tensor, Shear rate, To retrieve the yield stress of the sedimentary deposits from the inversion output, This is the consistency coefficient. The rheological index;

[0095] at the same time, For incompressible fluid Navier-Stokes dynamic residuals, ensure that the predicted flow field conforms to momentum conservation:

[0096] ;

[0097] in, The fluid velocity vector is denoted by g; g is the acceleration due to gravity. This represents the time corresponding to the current momentum conservation equation for the flow field environment. The density of the sediment fluid; For fluid pressure; Unit tensor; ∠ is the viscous stress tensor; ▽ is the bitwise operation symbol;

[0098] S33: Construct operational stability boundary constraints incorporating the dynamics of the operational robot; the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines introduces multibody dynamics equations to perform overall modeling of the front vehicle 27, rear vehicle 28, and hose 26 of the operational robot, and calculates the anti-overturning stability domain under complex flow fields. Establish the Lagrange dynamic equations:

[0099] ;

[0100] In the formula, For the generalized mass matrix of the robot; A generalized coordinate vector representing the spatial displacement and attitude angle of a robot. For generalized velocity vectors, The first derivative; It is a generalized acceleration vector. The second derivative; The matrix represents the Coriolis force and the centrifugal force. Here is the nonlinear stiffness matrix of hose 26; It is an external generalized force vector, which includes the omnidirectional high-pressure nozzle recoil force and the crusher head excitation force; Let this be the generalized gravity vector of the system; For the transpose of the contact Jacobian matrix; The contact force between the robot and the inner wall of the pipe;

[0101] In the aforementioned Lagrange dynamic equations, the generalized coordinate vector The spatial displacement data is obtained based on the displacement distance of the drive wheel 2; the attitude angle data is provided by the three-axis gyroscope 12 in real time by sensing the six degrees of freedom motion state of the robot.

[0102] S34: Optimization of breakup parameters based on Rayleigh-Plesset cavitation dynamics; establishment of an acoustic-fluid coupled cavitation efficiency model for the coordinated operation of the multi-channel ultrasonic breakup rotating head 4 and the high-pressure nozzle 5; introduction of Rayleigh-Plesset equations to describe the dynamic behavior of microbubbles under the combined action of ultrasound and jet:

[0103] ;

[0104] In the formula, The instantaneous radius of the cavitation bubble; Let be the first derivative of the bubble radius with respect to time, representing the velocity. Let be the second derivative of the bubble radius with respect to time, representing acceleration; The density of the sediment fluid; It is the saturated vapor pressure inside the liquid. The environmental driving pressure at infinity is mainly determined by the ultrasonic field; The additional static pressure generated by the jet; The surface tension coefficient of the liquid; The dynamic viscosity of the liquid;

[0105] The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines finds the optimal ultrasonic frequency by numerically solving the above differential equations. With jet pressure This causes the peak pressure of the microjets generated when the cavitation bubbles collapse to reach its maximum value. Maximize, and satisfy:

[0106] ;

[0107] In the formula, The internal bonding strength of the silt;

[0108] Meanwhile, based on the degree of freedom of the annular sliding groove 14, the cutting angle of the stirring brush 6 is optimized. To achieve the combined effect of mechanical shearing and cavitation erosion;

[0109] S35: Transfer efficiency control based on critical velocity of solid-liquid two-phase flow; for sediment transfer systems, a critical velocity control model based on Durand-Condolios theory is established to prevent blockage of sediment discharge pipe 8; the critical non-sludge velocity is calculated. :

[0110] ;

[0111] This is the Durand empirical coefficient, which is related to particle size and concentration; It is the acceleration due to gravity; The diameter of the pipe of the sediment extractor 7; The density of the solid particles in the sediment; The density of the fluid used to transport sediment;

[0112] The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines is based on the real-time mixing density of silt at the inlet of the silt extractor 7. Dynamically adjust the negative pressure power of vacuum sewage pump 9 Ensure the actual flow rate within the silt extractor 7 pipe. Always within the safe zone:

[0113] ;

[0114] In the formula, For safety margin coefficient, Maximum flow rate limit to prevent excessive wear of the silt extractor 7 pipe;

[0115] S36: Global optimal control closed loop based on Hamilton-Jacobi-Bellman (HJB) equations; coupling the subsystems in S31 to S35 to construct the Hamiltonian function of the whole system. Solve for the global energy-optimal control law:

[0116] ;

[0117] In the formula, It is an energy functional; This is the system state vector; To control the input vector; The transpose of the costate variable vector represents the shadow price of the system state change;

[0118] The system dynamics equations describe the rate of change of the system state over time;

[0119] ;

[0120] In the formula, It is an energy functional; To optimize the weighting coefficients of the objective; The rate of removal of sediment per unit time; This represents the total energy consumption of the system. This is the current control input vector; To smooth the control reference trajectory and prevent drastic fluctuations in the control input; These are the start and end times of the task;

[0121] The optimal control vector is obtained by solving the Hamilton-Jacobi-Bellman partial differential equation:

[0122] ;

[0123] In the formula, The optimal ultrasonic frequency; To achieve the optimal jet pressure; To achieve the optimal stirring speed; For optimal negative pressure power, It is the transpose symbol;

[0124] The above operations complete the closed-loop control from perception, inversion, decision-making to execution.

[0125] The collaborative dredging execution phase of S5 includes the following steps:

[0126] S51: Silt Softening: The 360-degree swingable high-pressure nozzle 5 is activated. Based on the three-dimensional distribution map of siltation in the water conservancy and flood control pipeline, the intelligent siltation diagnosis module of the water conservancy and flood control pipeline controls the swing angle of the high-pressure nozzle 5 and the extension stroke of its hydraulic telescopic nozzle to accurately spray the chemical ablation reagent onto the siltation surface and crevices. The nozzle adopts a fan-shaped spray pattern to ensure uniform reagent coverage and complete silt softening.

[0127] S52: Ultrasonic Breakup: After the silt softening is completed, the multi-channel ultrasonic breakup rotating head 4 is activated. The cylindrical rotating head drives 4 ultrasonic hammers to rotate at high speed. At the same time, the ultrasonic hammers output high-frequency vibrations to break the softened silt and coagulated blocks into small pieces. The ultrasonic thickness sensor 15 and the lidar 16 monitor the breaking effect in real time. If large silt and coagulated blocks are detected, the intelligent diagnosis module for siltation in the water conservancy and flood control pipeline extends the breaking time and increases the ultrasonic power of the multi-channel ultrasonic breakup rotating head 4.

[0128] S53: Mixing Assist: When the mixing brush 6 is activated, it rotates and mixes with the bottom of the water conservancy and flood control pipeline, mixing the broken up silt and small clumps with a small amount of water to form a slurry, thus preventing particle settling; at the same time, the bristles of the mixing brush 6 can reach into the gaps in the inner wall of the water conservancy and flood control pipeline to remove the attached silt. With the help of the flaw detection data of the operation robot, the mixing brush 6 can avoid the defective areas of the water conservancy and flood control pipeline, thus improving the safety of the operation.

[0129] S54: High-efficiency suction: The sediment transfer system is activated. Based on the flow field data and the concentration of the mud after stirring, the intelligent diagnosis module for sedimentation in the water conservancy and flood control pipeline adjusts the negative pressure power of the vacuum sludge pump 9. The mud-like sediment is quickly sucked out of the water conservancy and flood control pipeline through the suction port of the sediment suction device 7. The high-pressure nozzle 5 continuously sprays clean water, forming a pushing water flow in the water conservancy and flood control pipeline, which helps the mud to converge towards the suction port of the sediment suction device 7, avoiding sediment residue. The suctioned mud is output to the external storage device through the sediment discharge pipe 8, completing the dredging.

[0130] The measuring equipment used to perform the measurement method includes a working robot and a work control box. The working robot includes a hardware structure, which includes a front vehicle 27, a rear vehicle 28, and a fuselage shell structure, a motor system, a circuit system, an information sensing system, a sediment crushing system, and a sediment transfer system mounted on the front vehicle 27 and the rear vehicle 28. The front vehicle 27 and the rear vehicle 28 are connected by a flexible hose 26, which is a universal metal flexible hose that can rotate in any direction to facilitate the front vehicle 27 and the rear vehicle 28 to pass through curves. The work control box includes a software module, which includes an intelligent diagnosis module for sedimentation in water conservancy and flood control pipelines. The working robot and the work control box are connected by a cable 24, which contains signal transmission lines and power transmission lines to complete the transmission of signals and power.

[0131] The motor system is provided in multiple sets, and the multiple sets of motor systems are arranged in a centrally symmetrical manner with the central axis of the fuselage shell structure of the front vehicle 27 and the rear vehicle 28 as the center of symmetry. The motor system includes a telescopic support arm 1, a drive wheel 2 and a driven wheel 3. The drive wheel 2 and the driven wheel 3 are both fixed to the fuselage shell structure of the front vehicle 27 and the rear vehicle 28 through the telescopic support arm 1.

[0132] The circuit system includes a central command controller 11, a three-axis gyroscope 12, and a signal interface 13. The central command controller 11 and the three-axis gyroscope 12 are located inside the fuselage shell structure of the front vehicle 27, and the signal interface 13 is located at the upper rear of the fuselage shell structure of the rear vehicle 28. The three-axis gyroscope 12 is used to sense the motion state of the robot working in the water conservancy flood control pipeline and send attitude signals to the central command controller 11 to adjust the attitude and movement of the robot.

[0133] The sediment crushing system includes a multi-channel ultrasonic crushing rotary head 4, a high-pressure nozzle 5, a stirring brush 6, and an annular sliding groove 14, all mounted on the outer shell structure of the front vehicle 27. The multi-channel ultrasonic crushing rotary head 4 is connected to the center of the front outer shell structure of the front vehicle 27. The multi-channel ultrasonic crushing rotary head 4 is provided with an annular sliding groove 14, allowing it to move left and right and forward and backward, and to rotate based on the annular sliding groove 14. The high-pressure nozzle 5 is located on the upper front side of the outer shell structure of the front vehicle 27 and is connected to the outer shell structure of the front vehicle 27 via a metal pipe. The stirring brush 6 is located at the lower part of the front vehicle 27 and can rotate at the lower part of the front vehicle 27. It consists of multiple metal blades and its height can be adjusted vertically.

[0134] The information sensing system includes a Doppler flow velocity sensor 10, an ultrasonic thickness sensor 15, a lidar 16, and a detection component 17; three Doppler flow velocity sensors 10 are arranged circumferentially along the outer wall of the front vehicle 27; the ultrasonic thickness sensor 15 and the lidar 16 are arranged on the multi-channel ultrasonic crushing rotary head 4; the detection component 17 includes an ultrasonic flaw detection sensor and a pressure sensor, and the detection component 17 is arranged between the driving wheel 2 and the driven wheel 3 of the front vehicle 27.

[0135] The sludge transfer system includes a sludge suction device 7, a sludge discharge pipe 8, and a vacuum pump 9. The sludge suction device 7 is located at the bottom rear end of the rear vehicle 28's outer shell structure. The sludge discharge pipe 8 is located at the center of the rear part of the rear vehicle 28's outer shell structure. The vacuum pump 9 is located inside the rear vehicle 28's outer shell structure at a slightly rearward position. The inlet end of the vacuum pump 9 is connected to the sludge suction device 7, and the outlet end is connected to the sludge discharge pipe 8, for transferring the crushed sludge and sludge to the outside.

[0136] The central command controller 11 is responsible for controlling the motors of the telescopic support arm 1, the drive wheel 2, the multi-channel ultrasonic crushing rotary head 4, the high-pressure nozzle 5, the stirring brush 6, the sediment suction device 7, the sediment discharge pipe 8, the vacuum sludge pump 9, the Doppler flow velocity sensor 10, the three-axis gyroscope 12 for attitude control, the annular sliding groove 14, the ultrasonic thickness sensor 15, the lidar 16, and the detection component 17.

[0137] The operation control box includes a box structure 18, a box cover plate 19, a power supply device 20, a pipeline non-contact ultrasonic level gauge 21, a wireless signal transceiver module 22, a signal antenna 23, a cable 24, and a development board 25.

[0138] The power supply device 20, the non-contact ultrasonic level gauge 21, the wireless signal transceiver module 22, the signal antenna 23, and the development board 25 are all installed and fixed inside the box structure 18; the intelligent diagnostic module for siltation in water conservancy and flood control pipelines runs on the development board 25.

[0139] The wireless signal transceiver module 22 is used to receive the operation control commands received by the signal antenna 23 and the fluid state index data transmitted by the cable 24, and transmit the above data to the development board 25 as data input for the intelligent diagnosis algorithm of siltation in water conservancy and flood control pipelines.

[0140] The stirring brush 6, hose 26, and annular sliding groove 14 are all made of metal. The signal interface 13 is a multi-signal interface for preventing the fall off of the aviation plug. The signal antenna 23 is a high-power anti-interference signal antenna. The cable 24 is a towed floating cable. The cable 24 is also connected to the pipeline non-contact ultrasonic level gauge 21 and the wireless signal transceiver module 22.

[0141] The wireless signal transceiver module 22 is not a single radio frequency component, but a comprehensive data gateway terminal integrating a physical communication interface. For example... Figure 8 As shown, after the cable 24 enters the enclosure structure 18, it directly connects to the wired data interface of the non-contact ultrasonic level gauge 21, the wireless signal transceiver module 22, and the development board 25 via the terminal block to establish physical contact. The wireless signal transceiver module 22 is responsible not only for transmitting and receiving wireless signals but also acts as a front-end data aggregation node, receiving fluid state index data transmitted from the cable 24. Data from the non-contact ultrasonic level gauge 21 and the wireless signal transceiver module 22 are transmitted to the development board 25 via the cable 24 to achieve electrical signal transmission. Furthermore, the non-contact ultrasonic level gauge 21 and the wireless signal transceiver module 22 adopt a compact modular layout within the enclosure, sharing the aforementioned terminal block for equipment wiring.

Claims

1. A water conservancy flood control pipeline intelligent siltation detection method based on physical rule guidance, characterized by: Includes the following steps: S1, the deployment phase: the robot is placed into the target water conservancy and flood control pipeline through the rainwater well, and the operator issues the operation command through the ground terminal to drive the robot to start. S2, Model Building Stage: The robot performs flow velocity detection. The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines combines the design parameters of water conservancy and flood control pipelines with real-time flow field data to adjust the robot to its working posture. At the same time, it scans the water conservancy and flood control pipelines to initially construct a three-dimensional distribution map of siltation in the water conservancy and flood control pipelines. S3, Intelligent Diagnosis Stage: Run the intelligent diagnosis algorithm for siltation in water conservancy and flood control pipelines, input the data collected by the operation robot, and combine it with the reaction kinetic model of the chemical ablation reagent to calculate the optimal operation parameters; S4, parameter adjustment stage: adjust the working parameters of the robot according to the optimal working parameters obtained by the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines. S5, the collaborative dredging execution phase, involves using a robot to spray chemical ablation reagents to soften the silt, ultrasonically crush the silt, and assist in stirring and pulverizing the silt. The robot then pumps and transports the crushed silt. S6, Dynamic Adaptation Stage: During the operation, the flow field changes in the water conservancy and flood control pipeline are monitored in real time. If the flow velocity changes abruptly, the posture of the operating robot and the suction power of the silt are adjusted. The condition of the inner wall of the water conservancy and flood control pipeline is also detected. If new silt is found, the new silt is targeted for dredging, and the dredging of a section of the water conservancy and flood control pipeline is completed. S7, Secondary Inspection Phase: After completing the dredging of a section of the flood control pipeline, the operation robot scans the undredged section of the flood control pipeline again. The intelligent siltation diagnosis module of the flood control pipeline compares the data before and after dredging. If the siltation residue rate is high, secondary dredging is initiated; if the residue rate is low, the operation robot moves to the next section of the flood control pipeline and repeats the process of S2-S6. S8, during the shutdown and recovery phase, after the entire section of the water conservancy and flood control pipeline has been dredged, the intelligent diagnosis module for siltation of the water conservancy and flood control pipeline generates a dredging report and transmits the report to the ground terminal. The operator controls the operation robot to return to the rainwater well, and finally shuts down the operation robot and recovers the dredging equipment and stored silt. The intelligent diagnostic phase of S3 includes the following steps: S31: Construct a high-dimensional heterogeneous sensing spatiotemporal tensor field based on Riemannian manifold geometry; construct a system state tensor on Riemannian manifold space based on the non-Euclidean geometric features of the inner wall of a hydraulic flood control pipeline. First, using the point cloud data collected by the robot, a manifold metric tensor is defined to establish a body-orthogonal curvilinear coordinate system for the water conservancy and flood control pipeline. Secondly, the flow field velocity vector collected by the robot is... , Sediment thickness field and acoustic impedance spectrum Spatiotemporal registration and dimensionless processing are performed to generate the input feature tensor. : ； in, for The input feature tensor at time step; This is to iterate over any specific moment within the time window from the start time to the current time; For time-series feature aggregation operators, it means that from Time's up A collection of time window data at any given moment; for The velocity vector of the flow field acquired by the time-of-flight Doppler flow velocity sensor; The maximum modulus of the flow velocity; For multimodal feature splicing operators; for The thickness field of the sedimentary layer collected by the constantly operating robot; The diameter of the water conservancy and flood control pipeline; Impedance spectrum data; The standard acoustic impedance of water is used as the reference value for dimensionless processing. The six-DOF attitude vectors of the robot are provided by a three-axis gyroscope; S32: Inversion of non-Newtonian fluid rheological properties based on a physical information neural network; The feature tensor generated in S31 is used... The input is fed into a physical information neural network, which introduces the Herschel-Bulkley non-Newtonian fluid constitutive equation as a physical constraint layer, and minimizes the hybrid loss function. To invert the true rheological parameters of the sediment; S33: Construct operational stability boundary constraints incorporating the dynamics of the operational robot; the intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines introduces multibody dynamics equations to perform overall modeling of the operational robot and calculate the anti-overturning stability domain under complex flow fields. Establish the Lagrange dynamic equations: ； In the formula, For the generalized mass matrix of the robot; A generalized coordinate vector representing the spatial displacement and attitude angle of a robot. For generalized velocity vectors, The first derivative; It is a generalized acceleration vector. The second derivative; The matrix represents the Coriolis force and the centrifugal force. Here is the nonlinear stiffness matrix of the flexible hose (26); It is an external generalized force vector, which includes the omnidirectional high-pressure nozzle recoil force and the crusher head excitation force; The generalized gravity vector of the system; For the transpose of the contact Jacobian matrix; The contact force between the robot and the inner wall of the pipe; In the aforementioned Lagrange dynamic equations, the generalized coordinate vector The spatial displacement data contained therein is obtained based on the displacement distance of the drive wheel (2); the attitude angle data is provided by the three-axis gyroscope (12) in real time sensing the six degrees of freedom motion state of the robot. S34: Optimization of breakup parameters based on Rayleigh-Plesset cavitation dynamics; Establishment of an acoustic-fluid coupled cavitation efficiency model for the coordinated operation of ultrasonic breakup of sediment by a robot and a nozzle; Introduction of Rayleigh-Plesset equations to describe the dynamic behavior of microbubbles under the combined action of ultrasound and jet. The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines finds the optimal ultrasonic frequency by numerically solving the above differential equations. With jet pressure This causes the peak pressure of the microjets generated when the cavitation bubbles collapse to reach its maximum value. Maximize, and satisfy: ； In the formula, The internal bonding strength of the silt; S35: Transfer efficiency control based on critical velocity of solid-liquid two-phase flow; For sediment transfer systems, a critical velocity control model based on Durand-Condolios theory is established to prevent blockage of the sediment discharge pipe (8) used by the robot for sediment suction and transport; the critical non-sludge velocity is calculated. : ； Durand's empirical coefficient; It is the acceleration due to gravity; The diameter of the pipe of the silt suction device (7) used by the robot for silt suction and transportation; The density of the solid particles in the sediment; The density of the fluid used to transport sediment; The intelligent diagnostic algorithm for siltation in water conservancy and flood control pipelines is based on the real-time mixing density of silt at the inlet of the silt extractor (7). Dynamically adjust the negative pressure power of the vacuum sewage pump (9) Ensure the actual flow rate within the silt extractor (7) pipe. Always within the safe zone: ； In the formula, For safety margin coefficient, To prevent excessive wear of the silt extractor (7) pipe, the maximum flow rate is limited; S36: Global optimal control closed loop based on Hamilton-Jacobi-Bellman equations; couple the subsystems in S31 to S35 to construct the Hamiltonian function of the whole system. Solve for the global energy-optimal control law: ； In the formula, It is an energy functional; This is the system state vector; To control the input vector; The transpose of the costate variable vector represents the shadow price of the system state change; The system dynamics equations describe the rate of change of the system state over time; ； In the formula, It is an energy functional; To optimize the weighting coefficients of the objective; The rate of removal of sediment per unit time; This represents the total energy consumption of the system. This is the current control input vector; To smooth the control reference trajectory and prevent drastic fluctuations in the control input; These are the start and end times of the task.

2. The intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules as described in claim 1, characterized in that: In step S34, by minimizing the mixture loss function To invert the true rheological parameters of the sediment: ； in, These are the weighting coefficients for data-driven items; The flow field state or rheological parameters predicted by the neural network; The actual data observed by the sensor; These are the weighting coefficients for the physical constraint terms; For incompressible fluid Navier-Stokes dynamic residuals; The rheological physical residual is used to determine the hardness of sediment and whether it is in the form of soft mud or hard crust. Its definition is as follows: ； In the formula, For the shear stress tensor, For shear rate, To retrieve the yield stress of the sedimentary deposits from the inversion output, This is the consistency coefficient. The rheological index; at the same time, For incompressible fluid Navier-Stokes dynamic residuals, ensure that the predicted flow field conforms to momentum conservation: ； in, The fluid velocity vector is denoted by g; g is the acceleration due to gravity. This represents the time corresponding to the current momentum conservation equation for the flow field environment. The density of the sediment fluid; For fluid pressure; Unit tensor; is the viscous stress tensor; ▽ is the bitwise operation symbol.

3. The intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules as described in claim 1, characterized in that: In S33, the Rayleigh-Plesset equation is introduced to describe the dynamic behavior of microbubbles under the combined action of ultrasound and jet, and the calculation formula is as follows: ； In the formula, The instantaneous radius of the cavitation bubble; Let be the first derivative of the bubble radius with respect to time, representing the velocity. Let be the second derivative of the bubble radius with respect to time, representing acceleration; The density of the sediment fluid; This is the saturated vapor pressure inside the liquid; Environmental pressure at infinity; The additional static pressure generated by the jet; The surface tension coefficient of the liquid; The dynamic viscosity of the liquid.

4. The intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules as described in claim 1, characterized in that: In step S36, the optimal control vector is obtained by solving the Hamilton-Jacobi-Bellman partial differential equation: ； In the formula, The optimal ultrasonic frequency; To achieve the optimal jet pressure; To achieve the optimal stirring speed; For optimal negative pressure power, This is the transpose symbol.

5. The intelligent siltation detection method for water conservancy and flood control pipelines based on physical rules as described in claim 1, characterized in that: The collaborative dredging execution phase of S5 includes the following steps: S51: Silt Softening: Based on the three-dimensional distribution map of siltation in water conservancy and flood control pipelines, the intelligent diagnosis module for siltation in water conservancy and flood control pipelines controls the swing angle of the nozzle of the operation robot and the extension stroke of its hydraulic telescopic nozzle to spray chemical ablation reagent onto the siltation surface and crevices; the nozzle adopts a fan-shaped spray pattern to complete the silt softening; S52: Ultrasonic Breaking: After the silt softening is completed, the ultrasonic hammer of the working robot outputs high-frequency vibration to break the softened silt block; and the breaking effect is monitored in real time. If large residual silt blocks are detected, the intelligent diagnosis module for siltation in water conservancy and flood control pipelines will extend the breaking time and increase the ultrasonic power of the ultrasonic hammer. S53: Mixing Assistance: The robot mixes the broken-up silt and coagulated blocks with clean water to form a slurry, preventing particle sedimentation; combined with the flaw detection data of the robot, it enables the robot to avoid defective areas in the water conservancy and flood control pipeline during operation; S54: Silt suction: The silt transfer system is started. Based on the flow field data and the concentration of the mud after stirring, the intelligent diagnosis module for siltation in the water conservancy and flood control pipeline adjusts the negative pressure power of the vacuum sludge pump (9) and sucks the mud-like silt out of the water conservancy and flood control pipeline through the suction port of the silt suction device (7). The nozzle of the operation robot continuously sprays clean water, forming a pushing water flow in the water conservancy and flood control pipeline, which helps the mud to converge towards the suction port of the silt suction device (7) to avoid siltation residue. The suctioned mud is output through the silt discharge pipe (8) to complete the silt removal.

Images