Construction and Analysis Method of Coupled Dynamics Calculation Model of Trailer Suction Hoist Dredging Vessel Trailer Arm

Abstract

This invention provides a method for constructing and analyzing a coupled dynamics calculation model of the dredging arm of a trailing suction hopper dredger, comprising the following steps: S1: Establishing a multi-coordinate system and a marine environmental load model; S2: Predicting the frequency domain hydrodynamic characteristics of the target vessel based on three-dimensional potential flow theory; S3: Determining key dynamic parameters based on CFD simulation and theoretical calculation; S4: Constructing a time-domain coupled dynamics simulation model based on the OrcaFlex platform; S5: Performing multi-condition numerical calculations and safety analysis based on the time-domain coupled dynamics simulation model; S6: Developing a rapid safety distance prediction model based on a physical information neural network. This invention achieves accurate revelation of the lateral displacement mechanism of the trailing arm under crossflow conditions and quantitative definition of the safe operating boundary, providing a complete simulation analysis method for dredging operation safety assessment.

Description

Technical Field

[0001] This invention relates to the field of safety technology for ship and marine engineering dredging operations, and in particular to the construction and analysis method of a coupled dynamics calculation model of the trailing suction hopper dredging vessel's boom. Background Technology

[0002] Trailing suction hopper dredgers, with their unique operation mode of dredging while sailing, possess self-propulsion, self-dredging, self-loading, and self-unloading capabilities, and are widely used in waterway maintenance and port infrastructure construction both domestically and internationally. As port and waterway infrastructure construction continues to advance into deeper waters, the hydrological and meteorological conditions, including wind, waves, and currents, in the construction areas are becoming increasingly complex. When operating in severe sea conditions with significant crosscurrents, the draghead-boom system is prone to lateral displacement under environmental loads, leading to difficulties in controlling the draghead's attitude. In severe cases, the draghead may collide with the hull, causing draghead crushing and directly threatening construction safety.

[0003] To mitigate these risks, construction workers often rely on experience to shorten the wire ropes of the harrow winch and harrow head winch when environmental conditions deteriorate slightly, attempting to suppress harrow arm lateral movement by increasing the wire rope tension. While this measure enhances the harrow head-harrow arm system's resistance to environmental loads, it also reduces the harrow head's ground pressure, significantly decreasing grounding rate and digging efficiency. Blindly winding up the winch fails to utilize the optimal performance of equipment such as wave compensators and is a passive and limited coping strategy.

[0004] In existing technologies, some studies have established coupled dynamic models of the hull and rake head-arm system. However, most models simplify the hull and rake head-arm system into weakly coupled or independent parts, failing to fully reflect the complex dynamic constraints and interactions between them. Existing models simplify the description of the multi-segment articulated structure of the rake arm, usually treating the hull and rake arm as a whole, failing to reflect the relative rotation between the segments of the rake arm through universal joints and hinges. Furthermore, the modeling of wave compensators often uses simplified spring-damping models, lacking refined parameter determination methods based on actual working pressure and thermodynamic states. In addition, existing research mostly focuses on ideal environments such as single waves or steady flow, showing a significant deficiency in predicting the system dynamic response under the combined effects of crossflow and irregular waves. Summary of the Invention

[0005] This invention provides a method for constructing and analyzing a coupled dynamics calculation model of the dredging arm of a dredging vessel, aiming to solve the technical problems in the prior art of being unable to accurately predict the lateral displacement of the dredging arm system under the combined action of crossflow and irregular waves, and being unable to quantitatively define the boundary of safe operation.

[0006] To achieve the above objectives, the present invention adopts the following technical solution: The construction and analysis method of the coupled dynamics calculation model of the trailing suction hopper dredger's boom includes the following steps: S1: Establish a marine environmental load model, which includes the JONSWAP spectral density function and crossflow velocity vector; S2: Import the JONSWAP spectral density function into the hydrodynamic analysis software, and perform frequency domain hydrodynamic calculations on the target trailing suction hopper dredger based on the three-dimensional potential flow theory to obtain the response amplitude operator, added mass matrix and radiation damping matrix of the target trailing suction hopper dredger. S3: The roll damping coefficient of the target trailing suction dredger, as well as the axial fluid resistance coefficient and normal fluid resistance coefficient of each section of the trailing arm, are determined by computational fluid dynamics simulation. The equivalent stiffness and equivalent damping of the wave compensator, as well as the axial stiffness of the wire rope of the trailing winch and the axial stiffness of the wire rope of the trailing head winch are calculated. S4: Import the response amplitude operator, added mass matrix, radiation damping matrix, roll damping coefficient, axial fluid resistance coefficient and normal fluid resistance coefficient of each section of the rake arm, equivalent stiffness and equivalent damping of the wave compensator, axial stiffness of the wire rope of the rake winch and the rake head winch into the multibody dynamics simulation platform. In the multibody dynamics simulation platform, establish a rake arm system model including the rigid body component of the rake arm bend, the rigid body component of the upper rake arm, and the rigid body component of the lower rake arm including the rake head. Establish the hinge constraints between each rigid body component, establish the rake winch and the rake head winch and their corresponding wire ropes, establish the mechanical equivalent model of the wave compensator, and integrate all model components to form a time-domain coupled dynamics simulation model. S5: In the time-domain coupled dynamic simulation model, the ocean current parameters are set using the crossflow velocity vector, and multiple sets of wave parameters and dredging depth parameters are set to perform time-domain simulation calculations and obtain time-domain simulation data. S6: Based on time-domain simulation data, a training dataset is constructed, and a physical information neural network model is built. The physical law that the safety distance decreases monotonically with the meaningful wave height, spectral peak period and crossflow velocity is embedded as a physical constraint into the loss function of the physical information neural network model. The physical information neural network model is then trained and deployed as a rapid forecasting tool.

[0007] In summary, the present invention has at least the following beneficial effects: This invention establishes a five-dimensional coordinate system and derives the dynamic equations for the hull, the rake arm bend, the upper rake arm, and the lower rake arm, respectively. This enables a precise description of the relative rotation between the segments of the rake arm, overcoming the technical deficiency of existing technologies that treat the hull and rake arm as a whole and fail to reflect the relative motion inside the rake arm.

[0008] This invention comprehensively determines the roll damping coefficient, rake arm fluid resistance coefficient, and wave compensator equivalent stiffness and damping by combining CFD simulation, torque balance principle, and thermodynamic laws. It forms a set of key dynamic parameter determination methods that are independent of the black box parameters of commercial software, so that the input parameters of the simulation model have clear physical meaning and reproducibility.

[0009] This invention achieves an organic connection between frequency-domain hydrodynamic characteristics and nonlinear time-domain multibody dynamics through a two-stage modeling strategy that combines frequency-domain hydrodynamic prediction with time-domain full-system coupled simulation. It can simultaneously consider the random excitation of irregular waves and the geometric nonlinearity of the large-scale motion of the rake arm system.

[0010] This invention embeds the physical law that the safety distance decreases monotonically with environmental parameters into the loss function of a physical information neural network as a hard constraint, enabling the model to learn a mapping relationship that conforms to physical laws under small sample conditions, thus achieving a leap from hourly simulation to second-level prediction.

[0011] This invention, by defining safety evaluation indicators under different excavation depths, systematically reveals the transfer law of critical risk components as excavation depth increases, providing clear technical guidance for key safety monitoring points under different excavation depths. Attached Figure Description

[0012] Figure 1 A flowchart illustrating the construction and analysis method of the coupled dynamics calculation model of the dredging vessel's drag arm.

[0013] Figure 2 This is a schematic diagram showing the relative relationship between the geodetic coordinate system and the four body coordinate systems.

[0014] Figure 3 This is a schematic diagram of the force analysis on the rake head.

[0015] Figure 4 This is a schematic diagram of the components and their connections in a time-domain coupled dynamics simulation model. Detailed Implementation

[0016] like Figure 1 As shown, this embodiment provides a method for constructing and analyzing a coupled dynamics calculation model of the trailing suction hopper dredger's boom, including the following steps: S1: Establishing a multi-coordinate system and a marine environmental load model. Taking the target trailing suction hopper dredging vessel as the research object, S1 establishes a multi-coordinate system to describe the spatial motion of the hull and the trailing arm system, as well as an environmental load model to characterize the wave and current environment of the operating sea area.

[0017] S1.1: Establishing a geodetic coordinate system Geodetic coordinate system The origin Any reference point fixed on the earth, The axis points to true north. The axis points due east geographically. The axis is positively oriented with the vertical downward direction. Geodetic coordinate system. It serves as the absolute reference benchmark for all motion descriptions.

[0018] S1.2: Establish the ship's body coordinate system Hull-following coordinate system The origin Fixed at the center of gravity of the target trailing suction hopper dredging vessel. The positive direction is the axis pointing towards the bow along the length of the ship. The positive direction is the axis pointing to starboard along the width of the ship. The axis is positively oriented with the vertical downward direction. (Ship's body coordinate system) Moving along with the hull, this describes the hull's own six degrees of freedom motion and the position and attitude of each component of the rake arm system relative to the hull. (In the hull's body coordinate system) In this process, the ship's motion is decomposed into six independent degrees of freedom: along oscillation motion of the axis ,along sway motion of the axis ,along heave motion of the axis , around Rolling motion of the axis , around Pitching motion of the shaft , around The yaw motion of the shaft ,in Represents a time variable.

[0019] S1.3: Establish the body coordinate system of the rake arm bend pipe Body coordinate system of the rake arm bend The origin Fixed at the center of gravity of the rake arm bend, with the coordinate axis direction aligned with the principal inertial axis direction of the rake arm bend, the rake arm bend's body coordinate system... It moves together with the curved pipe of the rake arm.

[0020] S1.4: Establish the body coordinate system of the upper rake arm Upper rake arm body coordinate system The origin Fixed at the center of gravity of the upper rake arm, with the coordinate axes aligned with the principal inertial axes of the upper rake arm, the upper rake arm's body coordinate system... It moves along with the upper rake arm.

[0021] S1.5: Establish the body coordinate system of the lower rake arm Lower rake arm body coordinate system The origin Fixed at the combined center of gravity of the lower harrow arm and harrow head, with the coordinate axis direction aligned with the principal inertial axis of the lower harrow arm, the lower harrow arm's body coordinate system... It moves together with the lower rake arm and the rake head.

[0022] The relative relationships between the established geodetic coordinate system and the four body coordinate systems are as follows: Figure 2 As shown.

[0023] S1.6: Establish an irregular wave environmental load model. The JONSWAP spectrum is used to describe the irregular wave energy distribution in the operating area. The spectral density function of the JONSWAP spectrum... Calculated using the following formula: ; in The wave frequency is measured in Hertz (Hz). The energy scale parameter of the JONSWAP spectrum is a dimensionless quantity, obtained through... Calculation, the above The calculation formula is the standard empirical formula for the JONSWAP spectrum, which is obtained by fitting the measured data from the Joint North Sea Wave Project. The wave height is defined in meters. The peak period is expressed in seconds. is the peak factor of the JONSWAP spectrum, a dimensionless quantity with a value ranging from 1.0 to 7.0. In this invention, it is taken as... ; Let be the peak shape parameter, and be a dimensionless quantity. Time to take ,when Time to take ,in The frequency of the spectrum peak is expressed in Hertz (Hz).

[0024] S1.7: Establish the crossflow environmental load model. Crossflow is defined as a steady ocean current whose velocity magnitude and direction remain constant within the operational timescale. The crossflow velocity vector... Calculated using the following formula: ;in The crossflow velocity is expressed in meters per second. This is the cross-current direction angle, with the bow direction as the reference point and clockwise as positive, in degrees; superscript. This represents the matrix transpose operation.

[0025] S2: Prediction of the frequency domain hydrodynamic characteristics of the target ship based on three-dimensional potential flow theory. S2 uses the ship's body coordinate system established in S1.2. The JONSWAP spectral density function calculated in S1.6 Based on this, the hydrodynamic coefficients and motion response amplitude operators of the target trailing suction dredger in the frequency domain are calculated using linear three-dimensional potential flow theory.

[0026] S2.1: Establish the governing equations and boundary conditions based on potential flow theory. The fluid is assumed to be an ideal fluid that is inviscid, irrotational, and incompressible; the wave amplitude is sufficiently small relative to the wavelength to satisfy the linearization condition; and the hull motion amplitude is relatively small. Under these assumptions, the flow field can be represented by the velocity potential. This indicates that the velocity potential Satisfies the Laplace equation: ;in For the Laplace operator.

[0027] velocity potential The solution must satisfy the following boundary conditions: Surface boundary conditions: on the average wetted surface of the hull Above, the normal velocity of the fluid is equal to the normal component of the ship's velocity, that is: ;in The unit vector representing the outward normal to the hull surface. Let be the velocity vector of the ship at its center of gravity. Linearized free surface boundary conditions: on still water. Above, the linearized dynamic boundary conditions and motion boundary conditions are combined into: ;in The acceleration due to gravity is meters per second squared. Seabed boundary conditions: Assume water depth... It is a constant, at the bottom of the sea. At this point, the fluid normal velocity is zero: Radiation conditions: At infinity, waves generated by the motion of the ship or diffraction propagate only outwards, with no energy reflection.

[0028] S2.2: Decomposition of the total velocity potential based on the principle of linear superposition. Total velocity potential Decomposed into incident potential diffraction potential and radiation potential sum: ; among which the incident potential Corresponding to the incident wave field undisturbed by the ship hull; diffraction potential Describe the scattering effect of incident waves when the ship's hull is fixed, satisfying the object surface condition: ; radiation potential The radial waves generated when a ship undergoes forced motion in still water can be further decomposed into six components corresponding to the six degrees of freedom of motion.

[0029] S2.3: Calculate the first-order wave excitation force. Based on the linearized Bernoulli equation, the first-order dynamic force... for: ;in The density of seawater is expressed in kilograms per cubic meter.

[0030] Wave excitation force vector and wave excitation torque vector By examining the wet surface of the hull Points earned: ; ;in is the position vector of the integral point on the wetted surface of the hull relative to the center of gravity of the hull.

[0031] S2.4: Establish the frequency domain motion response equations of the hull and solve for the response amplitude operator. In the frequency domain, the first-order motion response of the hull is described by the following set of linear equations: ;in and Index for degrees of freedom, Corresponding to oscillation, Corresponding to horizontal movement, Corresponding to drooping, Corresponding to panning, Corresponding to pitch, Corresponding to bow roll; The angular frequency of the wave is denoted as , where . The wave frequency defined in S1.6 The unit is radians per second; The hull mass matrix is ​​the th Line number The elements of the column are in kilograms; Add the mass matrix to the hull. Line number The elements in the column, in kilograms, represent the angular frequency of the waves. The function; The first in the radiation damping matrix Line number The elements of the column, in kilograms per second, represent the wave circular frequency. The function; The first in the static water restoring force matrix Line number The elements of the column are in kilograms per second squared. For the first The complex amplitude of a motion with one degree of freedom, expressed in meters or radians, is the wave circular frequency. and wave angle The function; For the first The complex amplitude of the wave excitation force or torque in each degree of freedom, expressed in Newtons or Newton-meters, is the wave angular frequency. and wave angle The function; It is the imaginary unit.

[0032] Solving the above frequency domain motion response equations yields the hull's response at a given wave circular frequency. and wave angle The amplitude of complex motion with six degrees of freedom .

[0033] Response amplitude operator Defined as the hull first The ratio of the amplitude of the motion response per degree of freedom to the unit wave amplitude is calculated using the following formula: ;in The incident wave amplitude is expressed in meters. Represents the amplitude of complex motion The model.

[0034] S2.5: Frequency domain calculations were performed using the professional hydrodynamic analysis software OrcaWave. The wetted surface geometry model of the target trailing suction hopper dredger was imported into the OrcaWave software, and the wave frequency range was set to [value missing]. Hertz Hertz, frequency refinement is performed near the ship's natural rolling frequency. The JONSWAP spectral density function calculated in S1.6 is used. Import the wave spectrum into OrcaWave software as the wave input spectrum, wave direction angle. Five wave directions were calculated, ranging from 0° to 180° at 45° intervals. OrcaWave software, based on the potential flow theory from S2.1 to S2.4, calculated the response amplitude operators for the six degrees of freedom of the hull. Simultaneously output the additional mass matrix and radiation damping matrix .

[0035] S3: Determination of key dynamic parameters based on CFD (Computational Fluid Dynamics) simulation and theoretical calculation. S3 uses computational fluid dynamics (CFD) numerical simulation methods and theoretical mechanics analysis methods to determine the key input parameters required for the time-domain coupled dynamic simulation model.

[0036] S3.1: Calculate the hull roll damping coefficient based on the CFD method.

[0037] S3.1.1: A CFD simulation model of the target trailing suction hopper dredger's free roll decay motion is established in the STAR-CCM+ fluid dynamics simulation platform. The computational domain adopts a cuboid structure, and the longitudinal distance between the inlet boundary and the hull center is [missing information]. The longitudinal distance from the exit boundary to the center of the hull is The lateral distance from the two side boundaries to the mid-longitudinal section of the hull is The vertical distances from the top and bottom to the waterline are respectively: ,in The vertical distance between the target dredging vessel and the target suction hopper is [length]. The target width is defined for the trailing suction hopper dredging vessel. A length of [value] is set in front of all open boundaries in the computational domain. The wave damping region. In the computational domain The negative axis direction is set as the pressure outlet boundary. Both the positive and negative directions of the axis are set as symmetrical planes, the remaining boundaries are designated as velocity inlet boundaries, and the hull surface is set as a non-slip wall.

[0038] S3.1.2: Generating the computational mesh. Overlapping meshing technology is employed, with two separate systems: a background mesh and a hull mesh. Unstructured mesh generation is used, including surface reconstruction models, cut-body mesh generation, and prism layer mesh generation models. Vertical mesh refinement is implemented near the free surface. The turbulence model selected is shear stress transport. The model and boundary layer mesh satisfy... The requirements, among which The distance to the wall is dimensionless. Three computational grids with different mesh counts were designed: coarse, medium, and fine meshes. The roll period of the roll attenuation curves under the three grids was compared. Ratio of adjacent amplitudes The relative deviations of the roll period between coarse and medium grids, and between medium and fine grids, were calculated respectively. The relative deviation of the ratio of adjacent amplitudes When the relative deviation between the coarse and medium meshes is greater than the relative deviation between the medium and fine meshes, i.e. and When choosing a medium-sized grid, a convergent grid is selected. The rolling period, It is the ratio of adjacent amplitudes; This is the relative deviation of the roll period. The relative deviation of adjacent amplitude ratios; superscript This indicates the deviation between coarse and medium grids. This indicates the deviation between the medium and refined meshes; , The roll periods are calculated for coarse and medium meshes, respectively. To refine the roll period obtained from the mesh calculation; , These are the adjacent amplitude ratios calculated for coarse and medium grids, respectively. To refine the ratio of adjacent amplitudes obtained from the grid calculation.

[0039] S3.1.3: Conduct numerical simulation of the free decay motion of the roll. Set the initial roll angle of the hull as... Release the hull and allow it to undergo free rolling decay in still water, then record the rolling angle. Over time The evolution of the data was analyzed. Time step sizes of 0.02 seconds, 0.01 seconds, and 0.005 seconds were used to verify time step independence. By comparing the roll decay curves under different grid numbers and time step sizes, the convergent grid number and time step size were determined.

[0040] S3.1.4: Calculate the roll damping coefficient based on the roll annihilation curve. Plot the roll annihilation curve with the mean amplitude of the roll decay motion as the x-axis and the amplitude difference as the y-axis. Perform linear fitting on the scattered points on the annihilation curve to obtain the slope of the annihilation curve. The horizontal sway eliminates the slope of the curve. The following formula is used to obtain the fitted line: ;in It is the difference between two adjacent amplitudes. This is the average of two adjacent amplitudes. Dimensionless roll damping coefficient. Calculated using the following formula: Extract the natural angular frequency of the roll from the roll decay history curve. The unit is radians per second.

[0041] Roll damping coefficient Calculated using the following formula: ;in For the hull The moment of inertia of the shaft, expressed in kilograms per square meter, can be found in the loading manual of the target trailing suction hopper dredger. The design displacement is... and the radius of roll inertia It was calculated later. The target is the design displacement of the trailing suction hopper dredger, in tons; in calculation When doing so, it needs to be converted to the International System of Units (SI) kilogram, i.e. kilograms, then ; The radius of roll inertia is expressed in meters. .

[0042] S3.2: Calculate the fluid resistance coefficient of each section of the rake arm structure based on the CFD method.

[0043] S3.2.1: Based on the actual dimensions of the target trailing suction hopper dredging vessel's boom system, establish three-dimensional geometric models of the boom bend, upper boom, and lower boom on the STAR-CCM+ platform. Set the actual operating vertical angle of each boom section as follows: Unstructured meshes were used to generate computational meshes for the rake arm bend, upper rake arm, and lower rake arm, respectively. The turbulence model selected was... Model.

[0044] S3.2.2: Conduct numerical simulations of resistance under different flow velocities. Set the incoming flow velocity. Three typical operating conditions were described, namely 1.0 m / s, 1.5 m / s, and 2.0 m / s. The selection of these three incoming flow velocities was based on the annual average current velocity range (0.8 m / s to 2.2 m / s) of the target trailing suction hopper dredger's operating area, using approximate values ​​of the lower limit, middle value, and upper limit of this range to cover typical current velocity conditions in the operating area. Resistance was calculated in both the axial (along the length of the dredger arm) and normal (perpendicular to the length of the dredger arm) incoming flow directions.

[0045] S3.2.3: Extract resistance data and calculate the resistance coefficient. Post-processing is used to extract pressure distribution and viscous shear force data from each section of the rake arm to obtain the total resistance. ,in Indicates the bend in the rake arm, Indicates the upper rake arm, This indicates the lower rake arm.

[0046] Axial fluid resistance coefficient of each section of the rake arm structure and normal fluid drag coefficient Calculated using the following formula: ; ; in For the first The total resistance of the rake arm under axial flow, expressed in Newtons; For the first The total resistance of the rake arm under the normal flow direction, expressed in Newtons; For the first The projected area of ​​the rake arm in the axial flow direction, in square meters, is obtained by projecting the three-dimensional geometric model of the rake arm established in the STAR-CCM+ platform in step S3.2.1 onto a plane perpendicular to the axial flow direction, and then directly output by the cross-section measurement function of the STAR-CCM+ software. For the first The projected area of ​​the rake arm segment in the normal flow direction, in square meters, is obtained by projecting the three-dimensional geometric model of the rake arm established in the STAR-CCM+ platform in step S3.2.1 onto a plane perpendicular to the normal flow direction, and then directly outputting it using the cross-section measurement function of the STAR-CCM+ software. The variation range of the axial fluid resistance coefficient and normal fluid resistance coefficient of each rake arm segment under the three flow velocity conditions is less than 5%. Therefore, the average value of the resistance coefficients calculated under the three flow velocity conditions is taken to obtain the axial fluid resistance coefficient and normal fluid resistance coefficient of each rake arm segment.

[0047] S3.3: Calculate the rake head resistance based on theoretical mechanics.

[0048] S3.3.1: Calculate the horizontal frictional resistance of the rake head Horizontal frictional resistance of the rake head Calculated using the following formula: ;in Let be the coefficient of friction between the rake head and the soil layer, which is a dimensionless quantity; The supporting force of the soil layer on the rake head is expressed in Newtons.

[0049] Soil support force Calculated using the following formula: ;in The equivalent weight of the rake head in water is expressed in Newtons. The force of high-pressure water jets on the rake head is measured in Newtons. The pressure difference force generated by the suction action of the rake head is expressed in Newtons.

[0050] Rake head suction pressure difference Calculated using the following formula: ;in , is the rake head shape coefficient, a dimensionless quantity, which is obtained from the rake head design data based on the specific geometry of the rake head, and its typical value ranges from 1.5 to 3.0; Density of the mud-water mixture, expressed in kilograms per cubic meter; The flow velocity at the rake head inlet is measured in meters per second. The length of the rake head is in meters. The width of the rake head's facing surface is in meters.

[0051] S3.3.2: Calculate the cutting resistance of the rake teeth Rake tooth cutting resistance Calculated using the following formula: ;in The specific resistance of the soil, determined based on the soil type of the operating area, is expressed in Newtons per square meter (N / m²) and is obtained from geotechnical test reports or engineering geological survey data. If the original data is in kilograms per square meter (kgf / m²), it is multiplied by the acceleration due to gravity. Meters per second squared converted to Newtons per square meter; This represents the vertical displacement of the rake head, in meters. The elevation of the seabed mud surface, in meters; This refers to the depth to which the rake head penetrates the soil.

[0052] S3.3.3: Calculate the soil compaction resistance Earthwork resistance Calculated using the following formula: ; in The shear strength of the soil is determined based on the soil type of the operating sea area, and the unit is Newtons per square meter. It is obtained from the geotechnical test report or engineering geological survey data.

[0053] S3.3.4: Summarize the total resistance of the rake head. Total horizontal resistance of the rake head. Horizontal frictional resistance of the rake head Rake tooth cutting resistance and soil resistance sum: ; The total resistance of the rake head calculated in S3.3.4 This corresponds to a ground speed of 2.5 knots for the rake head. When the ship's ground speed deviates from 2.5 knots, the total resistance of the rake head is adjusted proportionally to the square of the ground speed, using the following formula: ,in The actual speed over ground is expressed in knots.

[0054] Force analysis of the rake head as follows Figure 3 As shown.

[0055] S3.4: Calculate the working pressure of the wave compensator based on the torque balance principle.

[0056] S3.4.1: Calculate the weight of the lower rake arm in water and its horizontal gravitational moment in water. The mass of the lower rake arm in air is denoted as... The unit is kilograms, obtained by multiplying the tonnage provided in the design data by 1000; the mass of the mud in the lower rake arm is denoted as... The unit is kilogram. Multiply the volume of the lower rake arm cavity by the mud density Calculated or obtained directly from design data. Gravity of the lower rake arm in water. To account for the equivalent gravity after the effect of buoyancy, the following formula is used for calculation: ; in The acceleration due to gravity is meters per second squared. This refers to the density of steel, expressed in kilograms per cubic meter. The density of seawater is expressed in kilograms per cubic meter. This refers to the density of the mud, expressed in kilograms per cubic meter.

[0057] The horizontal weight moment of the lower rake arm in water Calculated using the following formula: ;in This is the horizontal distance from the center of gravity of the lower rake arm to the cross hinge point, in meters.

[0058] S3.4.2: Calculate the geometric parameters of the rake arm. Total length of the rake arm. The length of the rake arm bend Length of the upper rake arm length of the lower rake arm and rake head length sum: ; Calculate the length of the rake arm Total length of the rake arm With the height of the rake head The hypotenuse of the right triangle formed: ;Inner angle of rake arm The interior angle subtended by the height of the rake head in the triangle: Distance from the rake head lifting point to the cross hinge point for: ;in This is the horizontal distance from the rake head lifting point to the cross hinge point, in meters; This refers to the height of the rake head lifting point, in meters. (Inner angle of the rake head lifting point) for: .

[0059] S3.4.3: Calculation of the maximum tension of the rake winch wire rope based on moment balance. With the cross hinge point as the moment center, the lower rake arm and rake head rotate around the cross hinge point, neglecting lateral constraints and hinge point friction. The total weight of the lower rake arm and rake head assembly in water is calculated. Gravity of the lower rake arm in water Gravity of the rake head in water Synthesis, in which , The mass of the rake head in air is expressed in kilograms. The unit is Newton.

[0060] First, calculate the dredging depth of the rake arm. ,in To dig deeper, Calculate the underwater depth of the suction inlet and the angle between the rake arm and the waterline. : ; The lever arm of the wire rope tension of the rake winch about the cross hinge point The distance from the rake head lifting point to the cross hinge point, calculated in S3.4.2, is determined as follows: and the inner angle of the rake head lifting point And the angle between the rake arm and the waterline calculated in S3.4.3 Find the angle between the line connecting the rake head lifting point and the cross hinge point and the direction of the waterline. . Used to calculate the direction of the line connecting the rake head lifting point and the cross hinge point, and to solve for the included angle of a triangle. The intermediate variable. The direction of the tension in the wire rope of the rake head winch is along the line connecting the rake head lifting point and the rake head A-frame lifting point, and the angle between this line and the line connecting the cross hinge point to the rake head lifting point. Using the geometric relationships of triangles and distance , and To obtain. The distance from the lifting point of the rake head A frame to the cross hinge point. The distance between the lifting point of the rake head A frame and the lifting point of the rake head. and The geometric dimensions are directly determined from the CAD drawings of the rake arm system. , ,in Here are the coordinates of the lifting point of the rake head A frame. Here are the coordinates of the cross hinge point. These are the coordinates of the rake head lifting point. Calculated using the Law of Cosines: lever arm The lever arm of the total weight about the cross hinge point The center of gravity of the lower rake arm and rake head assembly is determined by the weighted calculation of the center of gravity of the lower rake arm and the rake head, and by the angle between the line connecting the center of gravity and the cross hinge point and the waterline direction. From the coordinates of the barycenter and The calculation yielded the following result. Let the coordinates of the combined mass's center of gravity be... The coordinates of the cross hinge point are ,but: lever arm ,in The distance from the center of gravity to the cross hinge point. .in The calculation is performed in the longitudinal section of the rake arm ( Performed in the plane, laterally ( The offset in direction contributes zero to the torque at the cross hinge point, therefore it can be ignored. Coordinate differences.

[0061] According to the torque equilibrium condition: ;in This is the maximum tension in the wire rope of the rake winch, measured in Newtons. The lever arm of the wire rope tension of the rake winch about the cross hinge point, in meters; The lever arm of the total weight about the cross hinge point is expressed in meters.

[0062] S3.4.4: Calculate the average working pressure of the wave compensator. Compensation ratio. Defined as the ratio of the maximum wire rope tension borne by the wave compensator to the maximum tension when uncompensated (dimensionless, ranging from 0 to 1), the specific value is determined based on the wave compensator model and the initial charging pressure of its accumulator. In this embodiment, it is taken as... =0.5. When the compensation ratio is At that time, the tension of the rake winch wire rope From the piston force formula: ;in The diameter of the hydraulic cylinder plunger is in meters; considering the mechanical gain relationship of the movable pulley (the thrust of the hydraulic cylinder is twice the tension of the wire rope, i.e.) The average working pressure of the wave compensator can be obtained. for: .

[0063] S3.5: Calculate the equivalent stiffness and equivalent damping of the wave compensator based on the laws of thermodynamics.

[0064] S3.5.1: Calculate the volume of the gas inside the accumulator under each state. Based on the isothermal process assumption, the gas temperature inside the accumulator remains constant, and the initial pressure of the accumulator is... The initial volume is Average pressure air volume Calculated using Boyle's law: The change in hydraulic oil volume within the cylinder caused by the piston retracting to fully extend. for: ;in The stroke of the hydraulic cylinder is measured in meters.

[0065] The volume of air inside the accumulator when the piston retracts The volume of air inside the accumulator when the piston extends They are respectively: ; .

[0066] S3.5.2: Calculate the equivalent stiffness of the wave compensator. The wave compensator of the trailing suction hopper dredger adopts a movable pulley structure. One end of the wire rope is fixed to the hull, passes through the movable pulley connected to the hydraulic cylinder piston, and then connects to the lifting point of the drag head. Therefore, the transmission ratio between the displacement of the hydraulic cylinder piston and the displacement of the wire rope is 2:1 (i.e., for every unit length the piston moves, the wire rope retracts or expands by 2 units). The tension in the wire rope is half of the thrust of the hydraulic cylinder. Define the displacement of the hydraulic cylinder piston relative to the equilibrium position. (With the direction of retraction as positive), piston area The change in gas volume within the accumulator is equal to the change in oil volume within the cylinder, that is... From the isothermal process, the gas pressure Piston thrust The equivalent stiffness of the wave compensator cylinder itself. The derivative of the thrust with respect to the displacement at the equilibrium position Values ​​at: ; in This represents the equivalent stiffness of the wave compensator cylinder itself, expressed in Newtons per meter (N / m). Based on the principle of energy conservation, the displacement of the cylinder piston... Corresponding wire rope displacement The tension of the steel wire rope is Defined by stiffness After conversion .

[0067] S3.5.3: Calculate the equivalent damping of the wave compensator. Damping coefficient of the accumulator air chamber section. for: ;in is the dynamic viscosity coefficient of nitrogen, expressed in Pascal-seconds; The volume of gas inside the accumulator at average pressure, expressed in cubic meters. This refers to the cross-sectional area inside the accumulator, expressed in square meters.

[0068] Damping coefficient of the accumulator liquid chamber section for: ;in The dynamic viscosity coefficient of hydraulic oil is expressed in Pascal-seconds. The volume of hydraulic oil under average pressure inside the accumulator is expressed in cubic meters. The rated total volume of the accumulator is expressed in cubic meters, and its value is equal to the gas volume at the initial pressure of the accumulator. ,Right now The wave compensator cylinder's own equivalent damping. The sum of air cavity damping and liquid cavity damping: According to the energy method, the damping calculation and stiffness calculation follow the same transmission ratio square relationship, i.e. .

[0069] S3.6: Calculate the axial stiffness of the wire ropes of the rake winch and the rake head winch. The diameter of the rake winch wire rope is... and the diameter of the rake winch wire rope The value is determined based on the actual wire rope model. The elastic modulus of the wire rope is taken as... .

[0070] Cross-sectional area of ​​wire rope in rake winch Cross-sectional area of ​​the wire rope of the rake winch Calculate using the following formulas respectively: ; ; Axial stiffness of wire rope in rake winch Axial stiffness of rake winch wire rope Calculate using the following formulas respectively: ; .

[0071] S4: Construction of a time-domain coupled dynamics simulation model based on the OrcaFlex platform. S4 uses the hull response amplitude operator calculated by OrcaWave in S2.5. Additional mass matrix and radiation damping matrix and the roll damping coefficient determined in S3 Fluid resistance coefficient of each section of the rake arm and Total drag of the rake head after speed adjustment Equivalent stiffness of wave compensator and equivalent damping Axial stiffness of the wire rope of the rake winch Axial stiffness of rake winch wire rope Based on this, a time-domain coupled dynamic simulation model of the entire system—hull, rake arm bend, upper rake arm, lower rake arm, rake head, rake winch wire rope, rake head winch wire rope, and wave compensator—was established in the multibody dynamics simulation platform OrcaFlex.

[0072] S4.1: Create a 3D geometric model of the hull and import it into OrcaFlex. Using 3D modeling software, create a 3D geometric model of the hull based on the hull lines of the target trailing suction hopper dredger. Import the 3D geometric model of the hull into the Vessel module of OrcaFlex in STL format, and complete the basic dimensional settings of the model according to the actual principal dimension parameters of the target trailing suction hopper dredger: length between perpendiculars. , width , type depth Design draft .in The depth of the target trailing suction hopper dredging vessel, in meters; The design draft of the target dredging vessel is given in meters.

[0073] S4.2: Define the hull mass and moment of inertia properties. In the Vessel module of OrcaFlex, this is based on the design displacement of the target trailing suction hopper dredger. (Unit: kg) Set the hull mass. The hull's moments of inertia about the three coordinate axes are set as follows: Roll moment of inertia Pitch moment of inertia Bow roll moment of inertia Longitudinal position of the ship's center of gravity and vertical position of the center of gravity The loading manual of the target trailing suction hopper dredging vessel was used to determine this.

[0074] S4.3: Load the hull frequency domain hydrodynamic data. Use the response amplitude operators for the six degrees of freedom of the hull calculated by OrcaWave in S2.5. Additional mass matrix and radiation damping matrix Import hydrodynamic data from the Vessel module of OrcaFlex in tabular form. OrcaFlex uses the S2.5-based response amplitude operator. The wave excitation force time history is generated within the linear range. The roll damping coefficient calculated in S3.1.4 is used. The roll component is input as a constant additional linear damping matrix to OrcaFlex (here). The nonlinear viscous damping correction value obtained from CFD simulation is used to compensate for the inadequacy of linear radiation damping in potential flow theory during strongly nonlinear roll motion. The two are superimposed to form the total roll damping. In the Additional LinearDamping Matrix of the Vessel module in OrcaFlex, it is input as the diagonal element corresponding to the roll degree of freedom (the 4th degree of freedom). The convention for degree of freedom numbering in OrcaFlex is consistent with the degree of freedom numbering in step S2.4 of this invention, that is, the 4th degree of freedom corresponds to the roll motion, which is consistent with the value obtained from CFD simulation. The time-domain radiation damping forces obtained by convolution integration are superimposed to form the total damping term in the roll motion equation.

[0075] S4.4: Establish a multibody dynamics model of the rake arm system. Use the OrcaFlex 6D Buoy module to create three rigid body components of the rake arm system: the rake arm bend, the upper rake arm, and the lower rake arm containing the rake head.

[0076] S4.4.1: Define the physical properties of the rake arm bend. The mass of the rake arm bend is set to... The position of the center of gravity of the rake arm bend is determined based on the actual structure of the rake arm bend, and the moment of inertia of the rake arm bend is calculated based on its geometric dimensions. The axial fluid resistance coefficient of the rake arm bend calculated in S3.2.3 is then used. and normal fluid drag coefficient Assign water resistance coefficient parameters to the 6DBuoy module.

[0077] S4.4.2: Define the physical properties of the upper rake arm. The mass of the upper rake arm is set to... The center of gravity position of the upper rake arm is determined based on its actual structure, and the moment of inertia is calculated based on its geometric dimensions. The axial fluid resistance coefficient of the upper rake arm calculated in S3.2.3 is used. and normal fluid drag coefficient The water resistance coefficient parameter is assigned to the 6D Buoy module.

[0078] S4.4.3: Define the physical properties of the combined lower harrow arm and harrow head. The mass of the combined lower harrow arm and harrow head is set to... The center of gravity of the lower rake arm and rake head assembly is determined based on the actual structure of the lower rake arm and rake head, and the moment of inertia of the lower rake arm and rake head assembly is calculated based on the geometric dimensions of the lower rake arm and rake head. The axial fluid resistance coefficient of the lower rake arm calculated in S3.2.3 is used. and normal fluid drag coefficient Assign water resistance coefficient parameters to the 6D Buoy module. The total drag of the rake head after speed adjustment in S3.3.4 is then considered. As the soil resistance input for the interaction between the rake head at the end of the lower rake arm and the seabed, this resistance acts horizontally at the center of the bottom of the rake head, in the opposite direction to the ship's speed. Furthermore, a rake head rubber pad is defined on the outer surface of the rake head facing the ship for subsequent safety distance monitoring; the rake head rubber pad is an elastic buffer, and its spatial coordinates can be output from the simulation.

[0079] S4.5: Define the hinge constraints between the components of the rake arm system. Use the OrcaFlex Constraint module to define the connection relationships and motion constraints between the rigid body components. The rotational friction coefficient of each hinge constraint is set to... .

[0080] S4.5.1: Establish a longitudinal rotational hinge constraint between the hull and the rake arm bend, the longitudinal rotational hinge constraint allowing the rake arm bend to rotate about the hull's transverse axis in a vertical plane, and constraining the relative motion of the other five degrees of freedom.

[0081] S4.5.2: Establish a vertical horizontal rotational hinge constraint between the rake arm bend and the upper rake arm. The vertical horizontal rotational hinge constraint allows the upper rake arm to rotate in the horizontal plane about the vertical axis at the end of the rake arm bend, and constrains the relative motion of the other five degrees of freedom.

[0082] S4.5.3: Establish a cross-hing constraint between the upper and lower rake arms. This cross-hing constraint allows the lower rake arm to rotate relative to the upper rake arm about two mutually perpendicular axes, while constraining the relative motion of the other four degrees of freedom. Define a cross-hing rubber pad on the outer surface of the cross-hing constraint facing the hull side for safe distance monitoring; the cross-hing rubber pad is an elastic buffer, and its spatial coordinates can be output from the simulation.

[0083] S4.6: Establish the model of the hoisting system and wave compensator.

[0084] S4.6.1: Use the Winch module of OrcaFlex to create the harrow winch and harrow head winch, and generate the harrow winch wire rope connected to the harrow winch and the harrow head winch wire rope connected to the harrow head winch. Calculate the axial stiffness of the harrow winch wire rope obtained in S3.6. Axial stiffness of rake winch wire rope Assign the corresponding Winch module. Set the diameter of the winch wire rope in the rake to [value]. The diameter of the rake winch wire rope is set to .

[0085] S4.6.2: Use the Link module of OrcaFlex to create a mechanical equivalent model of the wave compensator. Apply the equivalent stiffness of the wave compensator calculated in S3.5.2 and S3.5.3, converted to the wire rope suspension point. and equivalent damping The Link module is then assigned to the hull. One end of the Link module is connected to the wave compensator base point on the hull, and the other end is connected to the corresponding lifting point of the rake arm system.

[0086] S4.7: Defines the ship's motion and external environmental conditions. In OrcaFlex's Primary Motion module, the ship's ground speed is set to... The heading is 0°. The JONSWAP spectral density function calculated in S1.6 is... Import wave environment settings and set the significant wave height. Spectral peak period and wave angle As a variable parameter, the crossflow velocity vector calculated in S1.7 is used. Import ocean current environment settings and set the crossflow velocity. and crossflow direction angle As a variable parameter, wave force is generated in the time domain based on frequency-domain hydrodynamic data calculated using OrcaWave in S2 via a linear superposition method.

[0087] S4.8: Integration and verification of the time-domain coupled dynamics simulation model. All model components established in S4.1 to S4.7 are integrated in the OrcaFlex platform to form a complete time-domain coupled dynamics simulation model of the hull-rake arm bend-upper rake arm-lower rake arm-rake head-middle winch wire rope-rake head winch wire rope-wave compensator. The time-domain coupled dynamics simulation model of the hull-rake arm bend-upper rake arm-lower rake arm-rake head-middle winch wire rope-rake head winch wire rope-wave compensator is verified by the following three indicators: the initial position error of the hinge constraint is less than... The root mean square value of the wave load deviates less than the analytical solution of the linear potential flow theory. The deviation between the roll period and the theoretical natural period in free decay motion is less than 10%. The initial position error of the hinge constraints is obtained by comparing the Euclidean distance between the initial coordinates of each hinge constraint point in OrcaFlex and the coordinates in the CAD design. ,in These are the initial coordinates of the hinge constraint points in OrcaFlex. The CAD design coordinates for the corresponding hinge constraint points; the deviation between the root mean square value of the wave load and the analytical solution of the linear potential flow theory is obtained by statistically calculating the time-domain wave force history and comparing it with the theoretical value; free decay motion is achieved by applying force to the hull. The initial roll angle is determined by releasing the hull and recording the roll duration. The theoretical inherent period of the roll is then determined. Through formula Calculation, theoretical inherent period The roll natural circular frequency extracted by CFD simulation in step S3.1.3 For the same physical quantity, the relationship between the two is as follows: Consistency deviation in the verification indicators refers to... with by Calculated period The relative deviation between them. The additional roll moment of inertia, in kilograms per square meter, is calculated from the additional mass matrix output by OrcaWave in step S2.5. The element corresponding to the mid-roll degree of freedom is obtained, that is... The corresponding additional mass value; This refers to the ship's displacement volume in cubic meters, obtained from the target ship's loading manual. The initial stability height is in meters, obtained from the target ship's loading manual.

[0088] The components and their connections in the time-domain coupled dynamics simulation model are as follows: Figure 4 As shown.

[0089] S5: Multi-condition numerical calculation and safety analysis based on time-domain coupled dynamic simulation model. S5 uses the time-domain coupled dynamic simulation model of hull-dredge arm bend-upper rake arm-lower rake arm-dredge head-middle winch wire rope-dredge head winch wire rope-wave compensator established in S4.8 as a tool to carry out systematic numerical simulation calculations under multiple wave, ocean current and dredging depth conditions. Taking the lateral safety clearance between the rake head-dredge arm system and the hull side as the core evaluation index, it analyzes the influence of environmental parameters and operating parameters on the risk of rake damage and quantitatively defines the safe operation boundary under different dredging depths and sea conditions.

[0090] S5.1: Set up the multi-condition simulation parameter matrix. Based on historical environmental statistics of the target trailing suction hopper dredger operating in the Yangtze River estuary, this data covers typical wave and current conditions in the four seasons of this sea area, with the ground speed fixed at [value missing]. Knot and heading are fixed at 0°; significant wave height Two horizontal levels were selected: 1.0 meter and 2.0 meter; spectral peak period. Three horizontal time points were selected: 5.0 seconds, 6.0 seconds, and 7.0 seconds; crossflow velocity. Eight horizontal wave speeds were selected: 0.5 m / s, 0.75 m / s, 1.0 m / s, 1.25 m / s, 1.5 m / s, 2.0 m / s, 2.5 m / s, and 3.0 m / s. In this invention, the wave direction is assumed to be consistent with the ocean current direction (i.e.,...). This assumption is based on the statistical regularity that wave and current directions are usually similar under crossflow conditions in engineering practice. Wave-current direction angle. Three levels were selected: 0°, 45°, and 90°. For special cases where the wave direction and current direction have a significant angle, extended analysis can be performed by setting the wave and current direction angles independently. The JONSWAP spectral peak factor was fixed at [value missing]. To analyze the parameter influence of S5.4, under all combinations of excavation depth and wave angle, the following parameters were applied: The simulation was performed at eight levels. The dataset used for training the fast prediction model in S6.1 only selected the self-mining depth... The operating condition combination is a 0° wave direction angle. The operational measures evaluation of S5.5 is performed with its simulation conditions set separately.

[0091] S5.2: Perform time-domain coupled dynamics simulation calculations. For each working condition selected in S5.1, run the simulation in the time-domain coupled dynamics simulation model established in S4.8, which includes the hull-rake arm bend-upper rake arm-lower rake arm-rake head-rake winch wire rope-rake head winch wire rope-wave compensator. The total time-domain simulation duration for each working condition is set to... To eliminate initial transient effects, seconds are extracted from the simulation time history. Instant Statistical analysis was performed on the data within the 600-second interval. The initial transient decay time of 600 seconds is approximately 55 times the natural period of the roll (approximately 11 seconds), which is sufficient for the initial transient to completely decay. The simulation time step was set to... seconds, the time step The selection of seconds is based on a wave cycle range of 5.0 to 7.0 seconds, ensuring that each wave cycle contains more than 500 time steps. This satisfies the numerical stability requirements of the implicit integration scheme and guarantees the analytical accuracy of high-frequency wave components. OrcaFlex employs an implicit time-domain dynamic integration scheme, recalculating the system geometry at each time step. On a computing platform configured with an Intel Xeon 3.6GHz processor and 128GB of memory, simulation of a single operating condition takes approximately 6 to 8 hours.

[0092] S5.3: Define the safety evaluation indicators for the rake head-rake arm system and the hull side. All distance calculations are performed in the geodetic coordinate system. The following steps will be taken. The reference point for the hull side will be a fixed point located at the outer plating of the midship side, and its geodetic coordinates will be... It is determined by the hull position and hull lines.

[0093] S5.3.1: When the excavation depth At the meter level, the rake head rubber pad and the cross hinge rubber pad are the components closest to the hull side. The distance from the rake head rubber pad to the hull side is defined as... The rake head rubber pad defined in S4.4.3 is the calculation... The location of the rake head rubber pad used, and the distance from the cross hinge rubber pad to the side of the hull are: . Through formula Calculation, where For the rake head rubber pad at all times The geodetic coordinates. The cross-hinged rubber pad defined in S4.5.3 is the one calculated in this step. The location of the cross-hinged rubber pad used. The calculation method and The same, except that the rake head rubber pad is replaced with a cross-hinged rubber pad, i.e. ,in For the cross hinge rubber pad at time The geodetic coordinates.

[0094] S5.3.2: When the excavation depth At this point, the rake head is located below the hull bottom, making the rake head winch wire rope and the cross-hing rubber pad the focal point of risk. The distance from the rake head winch wire rope to the hull side is defined as... Alternative As a safety evaluation indicator for the rake head system, the distance from the cross hinge rubber pad to the hull side is... It remains unchanged. The calculation method and Similarly, the rubber pad on the rake head is replaced with the point on the rake head winch wire rope closest to the hull side. The geodetic coordinates of this point are obtained by wire rope alignment interpolation. The wire rope alignment interpolation uses a piecewise linear interpolation method, dividing the wire rope into 50 equal linear segments based on the coordinates of the nodes at both ends. The spacing between adjacent interpolation points is approximately 2% of the total wire rope length. This number of segments is sufficient to ensure that the interpolation error is less than [a certain value]. Calculate the coordinates of each interpolation point, and take the point closest to the side of the ship among all interpolation points on the wire rope as the representative position of the wire rope.

[0095] S5.3.3: When the excavation depth At this point, both the rake head and the cross hinge are located below the hull bottom, making the rake head winch wire rope and the rake center winch wire rope the main risk components. The distance from the rake center winch wire rope to the hull side is defined as... Alternative As a safety evaluation indicator for the rake arm system, the distance from the rake head winch wire rope to the side of the hull is... It remains unchanged. The calculation method and Similarly, replace the cross-hinged rubber pad with the point on the winch wire rope closest to the hull side. When any safety evaluation index... (in When this occurs, it is determined that there is a risk of being crushed or collided.

[0096] S5.4: Analyze the influence of environmental parameters on the system's dynamic response. Extract the time-history data of safety evaluation indicators under various operating conditions from the simulation results of S5.2. Calculate the statistical values ​​of safety evaluation indicators under each working condition, including the maximum value. Minimum value and average The method of controlling variables was used to analyze the meaningful wave heights. Spectral peak period Wave direction angle Crossflow velocity and digging depth Impact on safety evaluation indicators. When analyzing a parameter using the controlled variable method, the other parameters are fixed at their baseline value—sense waveform height. Meter, spectral peak period Seconds, crossflow velocity meters per second, wave direction angle , dig deeper Meters. When the target parameter of the analysis changes, only the value of that parameter is changed, while the other parameters remain unchanged at the above baseline values. For multi-parameter interaction effects, range analysis using orthogonal experimental design is employed, with the minimum value of the safety evaluation index as the criterion. For the response quantity. In the orthogonal experimental design, the levels of each parameter are divided as follows: significant wave height Two horizontal levels were selected: 1.0 meter and 2.0 meter; spectral peak period. Three horizontal time points were selected: 5.0 seconds, 6.0 seconds, and 7.0 seconds; crossflow velocity. Four horizontal wave speeds of 0.5 m / s, 1.0 m / s, 1.5 m / s, and 2.0 m / s are used. In this invention, the wave direction is assumed to be consistent with the ocean current direction, i.e., the wave-current direction angle is... Three levels were selected: 0°, 45°, and 90°; the excavation depth was... Three levels—12.5 meters, 20 meters, and 30 meters—were used to represent shallow, medium, and deep water conditions, respectively, covering the main dredging depth range of the target trailing suction hopper dredger. Data from the 15-meter and 25-meter conditions were used as supplementary conditions for the controlled variable method analysis and were not included in the range analysis to avoid making the orthogonal experimental design too complex. The level selections of each parameter in the above orthogonal experimental design correspond to some levels in the full-condition simulation parameter matrix of S5.1. The range analysis used simulation data from S5.1 corresponding to the above level combinations for calculation. Condition data not included in the orthogonal design level combinations (such as crossflow velocity) were excluded. Operating conditions of 1.5 m / s and 3.0 m / s are not included in the range analysis. Calculate the range values ​​of each parameter. ,in For the first The parameter in the first... At each level, other parameters are taken as the arithmetic mean of the minimum values ​​of the safety evaluation indicators for all combinations of levels. ,in Indicates the first The parameter takes the first one. The first level, other parameters are taken as the first... Minimum value of safety evaluation index when there are multiple levels. The possible values ​​of ; For all level combinations of the other parameters, Indicates the first of the other parameters Level combinations; range The larger the value, the more significant the effect of the parameter. The orthogonal experimental design of this invention uses an orthogonal array that does not consider interaction effects, analyzing only the main effects of each parameter. This invention employs a mixed-level orthogonal array. , which includes 1 two-level column and 7 three-level columns. The significant wave height (two-level) occupies one two-level column; the spectral peak period (three-level), the wave-current direction angle (three-level) and the excavation depth (three-level) each occupy one three-level column; the cross-current velocity (four-level) is processed by splitting the four levels into two three-level factors, occupying two three-level columns. This orthogonal table requires a total of 18 groups of experimental combinations and can accommodate the existing level configurations of all parameters. The interaction effect between parameters is qualitatively analyzed through the simulation results of multi-parameter combined working conditions in the controlled variable method.

[0097] S5.5: Evaluate the improvement effect of operational measures on safety. For the working condition with excavation depth meters, significant wave height meters, spectral peak period seconds, and wave-current direction angle , four sets of retraction working conditions for the rake head winch steel wire rope and the mid-rake winch steel wire rope are set. Working condition A is the reference working condition, in which neither the rake head winch steel wire rope nor the mid-rake winch steel wire rope is retracted; working condition B is that the rake head winch steel wire rope is retracted by 1.0 meter and the mid-rake winch steel wire rope is retracted by 0.55 meter; working condition C is that the rake head winch steel wire rope is retracted by 2.0 meters and the mid-rake winch steel wire rope is retracted by 1.1 meters; working condition D is that the rake head winch steel wire rope is retracted by 3.0 meters and the mid-rake winch steel wire rope is retracted by 1.65 meters. The steel wire rope retraction amount of working condition D is determined according to the geometric size of the rake arm system of the target ship, by calculating the required winch rope retraction amount when the upper rake arm reaches the horizontal state (that is, the limit state where the horizontal angle of the upper rake arm approaches 0°), and this state corresponds to the retraction amount of the mid-rake winch steel wire rope reaching its mechanical stroke limit of 1.65 meters. Under the condition that the cross-current velocity is 0.75 meters per second, 1.0 meters per second, 1.25 meters per second and 1.5 meters per second respectively, the simulation calculation is carried out. Extract the statistical values of the safety evaluation indexes under each working condition from the simulation results and , and determine the safety flow velocity threshold under each steel wire rope retraction working condition . The safety flow velocity threshold is defined as the maximum cross-current velocity value that makes the minimum value of the safety evaluation index and hold simultaneously, that is . Calculate the improvement amount of the safety flow velocity threshold after the rake lifting operation .

[0098] S5.6: Quantitatively define the safe operation boundary. By integrating the analysis results of S5.4 and S5.5, taking the minimum value of the safety evaluation index as the criterion, for the combination of different excavation depth , significant wave height , spectral peak period and wave-current direction angle , determine the corresponding cross-current velocity safety threshold This establishes the safe operating boundaries for the target trailing suction hopper dredger under typical working conditions. These safe operating boundaries are output in tabular form, with rows corresponding to different dredging depths. and wave direction angle The combination of columns corresponds to different meaningful wave heights. Spectral peak period The combination of these values, and the data in the table, represent the corresponding crossflow velocity safety thresholds. .

[0099] S6: Development of a rapid prediction model for safe distance based on a physical information neural network. S6 uses simulation data obtained from multi-condition simulation calculations in S5.2 as the training dataset to construct a rapid prediction model for the safe distance between the rake head-rake arm system and the ship hull based on a physical information neural network.

[0100] S6.1: Constructing the training dataset. From the simulation results in S5.2, for a fixed excavation depth... Meter, wave direction angle Under the operating conditions, 21 sets of simulation data with different combinations of environmental parameters were selected to form a training dataset. Select the wave direction angle. The data from the operating conditions were used for training. Based on the analysis results of the influence of the wave-current direction angle on the safety clearance in step S5.4, in Crossflow velocity under wave-direction conditions The contribution to the safety clearance exceeds 85% (based on the calculation results of the range analysis in step S5.4, crossflow velocity). range The proportion of these parameters exceeding 85% of the sum of their ranges, while changes in the wave direction angle have a relatively small impact on safety clearances, therefore... Wave direction conditions can be used as representative conditions to establish forecasting models, covering typical scenarios where crossflow is the main risk factor in engineering projects. The specific combination is: significant wave height. Spectral peak period at meter time Take 5.0 seconds, 6.0 seconds, and 7.0 seconds, and measure the crossflow velocity. Take 0.5, 1.0, 1.5, and 2.0 meters per second (12 sets in total); Rice time, Set the time to 5.0 seconds, 6.0 seconds, and 7.0 seconds. Take 0.5, 1.0, and 1.5 meters per second (9 groups in total). When the significant wave height... crossflow velocity at meters The operating condition of meters per second has an extremely low probability of occurrence (less than 1%) in the historical environmental data of the target operating area, and the range analysis in step S5.4 shows that the simulation results of this operating condition combination are consistent with... The difference in operating conditions per meter per second was not significant in the safety clearance assessment (range). The variation rate is less than 5%, therefore it is not included in the training dataset. Since the training data all come from deep mining... For the meter operating condition, the specific output parameters here correspond to: The relevant indicators are those in S5.3.1. , The relevant indicators are those in S5.3.1. This forecasting model is applicable to For shallow excavation conditions (>12.8 meters), simulation data for the corresponding excavation depth should be obtained according to method S5.2, and the model parameters should be retrained according to steps S6.1 to S6.4.

[0101] Input parameter vector Defined as: ;in For the first The meaningful wave height of the group of working conditions, in meters; For the first The peak period of the spectrum under the working condition, in seconds; For the first The crossflow velocity under the working conditions is expressed in meters per second.

[0102] Output parameter vector Defined as: ;in , , The first The maximum, minimum, and average distances between the rake arm system (here, the cross-hinged rubber pad) and the hull side under the following working conditions, in meters; , , The first The maximum, minimum, and average distances between the rake head system (here, the rake head rubber pad) and the hull side under the following working conditions are given in meters.

[0103] S6.1.1: Perform Z-score normalization on the input and output parameters. Input parameters (in Corresponding to , , Standardized value of ) Calculated using the following formula: ;in For the first The mean of each input parameter across all training samples. For the first The standard deviation of each input parameter over all training samples This represents the total number of training samples.

[0104] Output parameters (in The standardized values ​​(corresponding to the six output safety distance indicators respectively) Calculated using the following formula: ;in For the first The mean of each output parameter across all training samples. For the first The standard deviation of each output parameter across all training samples.

[0105] S6.2: Constructing the Physical Information Neural Network Model Architecture. The physical information neural network adopts a feedforward fully connected neural network structure, which includes one input layer, two hidden layers, and one output layer.

[0106] S6.2.1: Define the input layer. The input layer receives the standardized 3D input parameter vector. .

[0107] S6.2.2: Define the first hidden layer. The first hidden layer contains... 10 neurons, using the hyperbolic tangent activation function The output of the first hidden layer for: ;in This is the weight matrix of the first hidden layer. This is the bias vector for the first hidden layer. Hyperbolic tangent activation function. It acts on each element of the vector.

[0108] S6.2.3: Define the second hidden layer. The second hidden layer contains... 10 neurons, using the hyperbolic tangent activation function The output of the second hidden layer. for: ;in This is the weight matrix for the second hidden layer. This is the bias vector for the second hidden layer.

[0109] S6.2.4: Define the output layer. The output layer contains... Each neuron uses a linear activation function. The network output... for: ;in The weight matrix of the output layer. This is the bias vector for the output layer. Network output. The predicted safety clearance is obtained by inverse normalization transformation, which yields the predicted value in physical space. : ;in .

[0110] S6.3: Design the loss function for a physical information neural network. Total loss function. Loss due to data fitting and physical constraint loss It consists of two parts: ;in These are the physical constraint weighting coefficients.

[0111] S6.3.1: Define the data fitting loss The data fitting loss is expressed in the form of mean squared error: ;in For physical information neural networks to the first Standardized input samples The predicted output, For the first Standardized true output of each sample This represents the Euclidean norm of a vector.

[0112] S6.3.2: Define physical constraint loss Based on the physical laws revealed in S5.4 through the control variable method and range analysis—sense wave height. Larger, spectral peak period The longer the crossflow velocity As the size increases, the safe distance between the rake head-rake arm system and the hull decreases monotonically—this monotonically decreasing law is embedded as a physical constraint into the physical information neural network.

[0113] For the training samples By applying positive perturbations to the three input parameters respectively, the perturbed input vectors are obtained. (in The positive perturbation factor of 1.1 is chosen so that the perturbed input parameters produce a 10% increment relative to the original parameters. This 10% increment is physically sufficient to cause a identifiable change in the safety clearance, while ensuring that the perturbed parameters remain within a physically reasonable range. ; ; Physical constraint loss Defined as: ;in For physical information neural networks, the original input The Each output predicted value, For physical information neural networks, the input after perturbation The One output predicted value; Indicates taking and The larger value in the range. Physical constraint loss. Only predictive behaviors that violate the law of monotony and diminishing returns are punished.

[0114] S6.4: Training the physical information neural network model.

[0115] S6.4.1: Initialize network parameters. The weights of the physical information neural network are initialized using a Xavier normal distribution, and the biases are initialized to zero.

[0116] S6.4.2: A two-stage training strategy is adopted. The first stage is the standard supervised learning stage, and the optimization algorithm uses an adaptive moment estimator optimizer with an initial learning rate of 100%. The number of training rounds is set to Batch size set to The learning rate decays every 200 epochs with a decay factor of 0.5. The first stage only optimizes the data fitting loss. Physical constraint weighting coefficient . Selected from the commonly used learning rate range of adaptive moment estimation optimizers. The convergence of the loss curve is determined based on the monitoring validation set.

[0117] The second stage is the physical constraint fine-tuning stage, where the gradient descent algorithm is used for optimization, with an initial learning rate of 100%. The number of training rounds is set to Full-batch training is used. The second stage simultaneously optimizes the data fitting loss. and physical constraint loss The physical constraint weight coefficient is set to . Selected from the commonly used learning rate range for the gradient descent algorithm. Based on the convergence of the loss curve on the validation set, training is stopped when the loss on the validation set no longer decreases after 50 consecutive rounds.

[0118] S6.4.3: Leave-one-out cross-validation is used to evaluate model performance. Each of the 21 datasets is used as the validation set, and the remaining 20 datasets are used as the training set. The training process described in S6.4.1 and S6.4.2 is repeated 21 times. After each training iteration, the prediction error of the physical information neural network model on the validation set, including the root mean square error, is calculated. Mean absolute error and average relative error : ; ; ;in To determine the number of samples in the validation set, Let represent the Manhattan norm of the vector. In 21 cross-validations, the minimum root mean square error was 0.004 meters and the maximum was 0.043 meters; the minimum mean absolute error was 0.003 meters and the maximum was 0.038 meters; and the mean root mean square error was... Meters, with an average relative error not exceeding .

[0119] S6.5: Deploy the trained physical information neural network model as a fast prediction tool. After completing 21 leave-one-out cross-validations and confirming that the model performance meets the requirements, retrain the physical information neural network model using all 21 sets of training data to obtain the final network parameters for deployment. The trained physical information neural network model includes the network weight matrix. , , and bias vector , , and standardized parameters , , , Save as a model file. For any new environmental conditions input... First, Z-score standardization is performed to obtain... Then, the standardized prediction output is obtained through forward propagation calculation using a physical information neural network. Finally, the predicted safe distance in physical space is obtained through inverse normalization transformation. .

[0120] In the above technical solution, the OrcaWave frequency-domain hydrodynamic calculation described in S2.5 and the OrcaFlex time-domain coupled simulation described in S4.8 constitute the core implementation path of this invention. As an alternative implementation path, a joint simulation approach using MATLAB / Simulink and ADAMS can also be adopted, where MATLAB / Simulink performs frequency-domain hydrodynamic calculations and solves time-domain motion equations, and ADAMS performs multibody dynamics modeling of the rake arm system, achieving bidirectional data exchange through a joint simulation interface; or a joint simulation approach using the OpenFOAM and Chrono open-source platforms can be adopted, where OpenFOAM solves the viscous flow field around the hull and obtains hydrodynamic coefficients, and Chrono establishes the multibody system model of the rake arm, achieving data transfer through a custom coupling interface.

Claims

1. A method for constructing and analyzing a coupled dynamics calculation model of a trailing suction hopper dredging vessel's boom, characterized in that... Includes the following steps: S1: Establish a marine environmental load model, which includes the JONSWAP spectral density function and crossflow velocity vector; S2: Import the JONSWAP spectral density function into the hydrodynamic analysis software, and perform frequency domain hydrodynamic calculations on the target trailing suction hopper dredger based on the three-dimensional potential flow theory to obtain the response amplitude operator, added mass matrix and radiation damping matrix of the target trailing suction hopper dredger. S3: The roll damping coefficient of the target trailing suction dredger, as well as the axial fluid resistance coefficient and normal fluid resistance coefficient of each section of the trailing arm, are determined by computational fluid dynamics simulation. The equivalent stiffness and equivalent damping of the wave compensator, as well as the axial stiffness of the wire rope of the trailing winch and the axial stiffness of the wire rope of the trailing head winch are calculated. S4: Import the response amplitude operator, added mass matrix, radiation damping matrix, roll damping coefficient, axial fluid resistance coefficient and normal fluid resistance coefficient of each section of the rake arm, equivalent stiffness and equivalent damping of the wave compensator, axial stiffness of the wire rope of the rake winch and the rake head winch into the multibody dynamics simulation platform. In the multibody dynamics simulation platform, establish a rake arm system model including the rigid body component of the rake arm bend, the rigid body component of the upper rake arm, and the rigid body component of the lower rake arm including the rake head. Establish the hinge constraints between each rigid body component, establish the rake winch and the rake head winch and their corresponding wire ropes, establish the mechanical equivalent model of the wave compensator, and integrate all model components to form a time-domain coupled dynamics simulation model. S5: In the time-domain coupled dynamic simulation model, the ocean current parameters are set using the crossflow velocity vector, and multiple sets of wave parameters and dredging depth parameters are set to perform time-domain simulation calculations and obtain time-domain simulation data. S6: Based on time-domain simulation data, a training dataset is constructed, and a physical information neural network model is built. The physical law that the safety distance decreases monotonically with the meaningful wave height, spectral peak period and crossflow velocity is embedded as a physical constraint into the loss function of the physical information neural network model. The physical information neural network model is then trained and deployed as a rapid forecasting tool.

2. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, Before establishing the marine environmental load model, a multi-coordinate system was also established, including a geodetic coordinate system, a ship-body coordinate system, a sniper arm bend sniper coordinate system, an upper sniper arm sniper coordinate system, and a lower sniper arm sniper coordinate system. The origin of the geodetic coordinate system was fixed at any reference point on Earth, with the x-axis pointing due north, the y-axis pointing due east, and the z-axis pointing vertically downwards. The origin of the ship-body coordinate system was fixed at the center of gravity of the target dredging vessel, with the x-axis pointing towards the bow along the ship's length, the y-axis pointing towards the starboard side along the ship's beam, and the z-axis pointing vertically downwards. In the ship's body coordinate system, the ship's motion is decomposed into six degrees of freedom: pitching, swaying, heaving, rolling, pitching, and bowing. The origin of the rake arm bend's body coordinate system is fixed at the center of gravity of the rake arm bend, the origin of the upper rake arm's body coordinate system is fixed at the center of gravity of the upper rake arm, and the origin of the lower rake arm's body coordinate system is fixed at the combined center of gravity of the lower rake arm and the rake head. The calculation of the roll damping coefficient is carried out in the ship's body coordinate system. The rigid body components of the rake arm bend, upper rake arm, and lower rake arm are established in the rake arm bend's body coordinate system, the upper rake arm's body coordinate system, and the lower rake arm's body coordinate system, respectively.

3. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The JONSWAP spectral density function is established as follows: the significant wave height, peak period, and peak factor of the operating sea area are obtained; the energy scale parameter of the JONSWAP spectrum is calculated based on the significant wave height, peak period, and peak factor; and the JONSWAP spectral density function is calculated based on the significant wave height, peak period, energy scale parameter, and peak factor. The crossflow velocity vector is established as follows: the magnitude of the crossflow velocity and the crossflow direction angle of the operating sea area are obtained, and the crossflow velocity vector is calculated based on the magnitude of the crossflow velocity and the crossflow direction angle. The crossflow direction angle is based on the bow direction as 0° and the clockwise direction is positive.

4. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The S2 mid-frequency domain hydrodynamic calculation further includes: establishing the Laplace equation satisfied by the velocity potential based on the assumptions that the fluid is inviscid, irrotational, and incompressible; setting the boundary conditions of the object surface, linearized free surface, seabed, and radiation; decomposing the velocity potential into the sum of incident potential, diffraction potential, and radiation potential; calculating the first-order dynamic pressure based on the velocity potential and integrating it over the wetted surface of the hull to obtain the wave excitation force and wave excitation torque; establishing the frequency domain motion response equation of the hull and solving it to obtain the response amplitude operator.

5. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The roll damping coefficient in S3 is obtained as follows: A simulation model of the target trailing suction hopper dredger's free roll decay motion is established in a hydrodynamic simulation platform. The initial roll angle of the target dredger is set and the hull is released. The change of the roll angle over time is recorded. A roll annihilation curve is plotted with the mean amplitude of the roll decay motion as the abscissa and the amplitude difference as the ordinate. Linear fitting is performed to obtain the slope of the annihilation curve. The dimensionless roll damping coefficient is calculated based on the slope of the annihilation curve. The natural circular frequency of the roll is extracted from the change of the roll angle over time. The roll damping coefficient is calculated based on the dimensionless roll damping coefficient, the hull's moment of inertia about the x-axis, and the natural roll frequency. The axial and normal fluid resistance coefficients of each section of the rake arm are obtained by establishing three-dimensional geometric models and computational meshes for the rake arm bend, upper rake arm, and lower rake arm, respectively, setting multiple incoming flow velocities to calculate the total resistance in the axial and normal incoming flow directions, and calculating the axial and normal fluid resistance coefficients of each section of the rake arm based on the total resistance, seawater density, incoming flow velocity, and projected area.

6. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The equivalent stiffness and equivalent damping of the wave compensator in S3 are obtained in the following ways: the average working pressure of the wave compensator is calculated based on the torque balance principle; the average pressure air volume is calculated based on the average working pressure and the initial pressure and initial volume of the gas in the accumulator; the change in hydraulic oil volume in the cylinder is calculated based on the cylinder plunger diameter and cylinder stroke; the equivalent stiffness of the cylinder is calculated based on the average working pressure, the average pressure air volume, and the piston area, and then converted to the wire rope suspension point based on the mechanical gain of the pulley system to obtain the equivalent stiffness; the equivalent damping of the cylinder itself is obtained based on the sum of the damping coefficient of the accumulator air chamber and the damping coefficient of the accumulator liquid chamber, and then converted to the wire rope suspension point based on the mechanical gain of the pulley system to obtain the equivalent damping.

7. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The time-domain coupled dynamics simulation model formed in S4 further includes: importing the three-dimensional geometric model of the target trailing suction hopper dredger into the Vessel module of the multibody dynamics simulation platform and setting the model scale according to the principal scale parameters; setting the hull mass according to the design displacement of the target trailing suction hopper dredger and setting the hull's own moment of inertia about the three coordinate axes; establishing longitudinal rotational hinge constraints between the hull and the rake arm bend; establishing vertical horizontal rotational hinge constraints between the rake arm bend and the upper rake arm; establishing cross hinge constraints between the upper rake arm and the lower rake arm; setting the rotational friction coefficient of each hinge constraint; creating the mid-drake winch and the head rake winch through the Winch module and generating the wire ropes of the mid-drake winch and the head rake winch connected to the mid-drake winch; creating the mechanical equivalent model of the wave compensator through the Link module and assigning the equivalent stiffness and equivalent damping to the Link module; and setting the primary model of the multibody dynamics simulation platform. In the Motion module, the ship's ground speed and heading are set, the JONSWAP spectral density function is imported into the wave environment settings, and the crossflow velocity vector is imported into the ocean current environment settings to complete the integration and verification of the time-domain coupled dynamic simulation model.

8. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, After obtaining the time-domain simulation data, S5 also extracts the lateral safety clearance between the rake head-rake arm system and the hull side as a safety evaluation index. The safety evaluation index is defined differently depending on the digging depth. When the digging depth is less than or equal to 12.8 meters, the distance from the rake head rubber pad to the hull side is used as the safety evaluation index for the rake head system, and the distance from the cross-hing rubber pad to the hull side is used as the safety evaluation index for the rake arm system. When the digging depth is greater than 12.8 meters but less than 21.5 meters, the distance from the rake head winch wire rope to the hull side is used as the safety evaluation index. The distance to the side of the rake head rubber pad is used as the safety evaluation index for the rake head system, replacing the distance from the rake head rubber pad to the side of the hull. The distance from the cross hinge rubber pad to the side of the hull is used as the safety evaluation index for the rake arm system. When the digging depth is greater than or equal to 21.5 meters, the distance from the rake head winch wire rope to the side of the hull is used as the safety evaluation index for the rake head system, and the distance from the rake winch wire rope to the side of the hull is used as the safety evaluation index for the rake arm system, replacing the distance from the cross hinge rubber pad to the side of the hull. When the value of any safety evaluation index is less than zero, it is determined that there is a risk of rake crushing or collision.

9. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The physical information neural network model in S6 further includes: extracting meaningful wave height, spectral peak period, and crossflow velocity from multiple sets of working conditions as input parameters from time-domain simulation data; extracting the maximum, minimum, and average values ​​of the distance between the rake arm system and the ship's side, as well as the maximum, minimum, and average values ​​of the distance between the rake head system and the ship's side, as output parameters; standardizing the input and output parameters; and constructing a feedforward fully connected neural network containing one input layer, two hidden layers, and one output layer. The activation function of the hidden layers uses the hyperbolic tangent function. The loss function of the physical information neural network model consists of two parts: data fitting loss and physical constraint loss. The data fitting loss uses the mean square error form, and the physical constraint loss is constructed by applying positive perturbations to the meaningful wave height, spectral peak period, and crossflow velocity respectively and penalizing predictions where the predicted value after perturbation is greater than the original predicted value. The physical information neural network model adopts a two-stage training strategy: the first stage optimizes the data fitting loss, and the second stage simultaneously optimizes the data fitting loss and the physical constraint loss.

10. The method for constructing and analyzing the coupled dynamics calculation model of the trailing suction hopper dredger's boom according to claim 1, characterized in that, The settings for multiple sets of wave parameters, dredging parameters, and ocean current parameters set using cross-current velocity vectors include: a fixed ground speed of 2.5 knots, a fixed heading of 0°, two levels for significant wave height (1.0 m and 2.0 m), three levels for spectral peak periods (5.0 s, 6.0 s, and 7.0 s), eight levels for cross-current velocities (0.5 m / s, 0.75 m / s, 1.0 m / s, 1.25 m / s, 1.5 m / s, 2.0 m / s, 2.5 m / s, and 3.0 m / s), three levels for wave and current direction angles (0°, 45°, and 90°), and five levels for dredging depths (12.5 m, 15 m, 20 m, 25 m, and 30 m). The total duration of the time-domain simulation calculation is set to 10800 seconds, and data from the interval between 600 seconds and 10799 seconds is extracted for statistical analysis. The simulation time step is set to 0.01 seconds.

Images