A prefabricated wave retaining wall construction process simulation and risk early warning system

Abstract

The present application relates to the technical fields of ocean engineering construction simulation, and discloses a kind of assembled wave breakwater construction process simulation and risk early warning system, including offline data preprocessing module, environmental perception and state initialization module, closed loop coupling calculation module and phase space early warning intervention module.System constructs hydrodynamic response tensor field and perturbation matrix offline;Collection state is converted into initial generalized coordinate and generalized momentum;Extract reference hydrodynamic combined with perturbation matrix correction, utilize Hamiltonian function and regular equation output current step generalized coordinate and generalized momentum;Accordingly, construct dimensionality reduction phase space state vector and dynamic bounding box sequence and carry out intersection operation, when non-empty, send brake signal.The present application decouples hydrodynamic calculation, solves the problem of long time consumption of online calculation;Introduce Hamiltonian to realize the conservation calculation of coupled system;Establish dynamic interference prediction to calculate collision time in advance, solve the brake lag defect caused by lack of kinematic foresight.

Description

Technical Field

[0001] This invention relates to the field of marine engineering construction simulation technology, specifically to a simulation and risk warning system for the construction process of prefabricated wave-breaking walls. Background Technology

[0002] The installation of prefabricated wave barriers in complex marine environments is a typical high-risk offshore construction operation. Before and during actual operations, computer simulation systems are typically used to simulate the motion of large components and, combined with relevant mechanisms, provide collision warnings to ensure the safety of lifting equipment and components. However, existing construction simulation and early warning systems still have significant technical shortcomings in practical applications.

[0003] When simulating wave and flow forces acting on structural members, existing systems typically solve complex fluid control equations online within each numerical integration time step to obtain hydrodynamic parameters such as radiation damping and added mass. This direct calculation method consumes a significant amount of computing power, resulting in lengthy online calculations of the hydrodynamic response. This slows down the overall system's coupled solution efficiency and makes it difficult to meet the real-time requirements of on-site construction operations.

[0004] In calculating the motion state of components, traditional dynamic models often employ numerical integration of Newton's equations based on displacement and velocity variables. However, when dealing with long-term simulations of complex coupling effects between marine fluids and hoisted components, this conventional integration method struggles to maintain the system's energy conservation. Energy drift or numerical dissipation easily occurs during the calculation, directly reducing the accuracy of the component's spatial pose and motion state calculations.

[0005] In terms of safety control and risk warning, most existing collision detection mechanisms rely on static geometric overlap determination of the component's current instantaneous position. These models lack dynamic forward-looking consideration of the component's kinematic state and fail to effectively introduce velocity compensation and time gradient factors in bounding box interference detection. This limitation causes the warning system to be unable to accurately calculate the specific remaining collision time, often triggering an alarm only when the component is already close to the obstacle. This results in a lag in the braking control of the lifting equipment and fails to provide a sufficient safety intervention window for on-site operations. Summary of the Invention

[0006] To address the shortcomings of existing technologies, this invention provides a simulation and risk warning system for the construction process of prefabricated wave-breaking walls. This system solves the problems of long online calculation time for hydrodynamic response, low efficiency of coupled solution, and braking lag caused by the lack of kinematic dynamic foresight in collision warning models during the hoisting of large components in complex sea conditions.

[0007] To achieve the above objectives, the present invention provides the following technical solution: a simulation and risk warning system for the construction process of a prefabricated wave barrier wall, comprising: The offline data preprocessing module constructs a hydrodynamic response tensor field based on spatial pose parameters and marine environmental parameters, and calculates the first-order radiation damping perturbation matrix and the second-order added mass perturbation matrix. The environmental perception and state initialization module collects marine environmental parameters and the initial motion state of components, and combines them with a preset generalized mass matrix to convert them into initial generalized coordinates and initial generalized momentum. The closed-loop coupled solution module extracts the reference hydrodynamic vector in the hydrodynamic response tensor field based on the marine environmental parameters and the generalized coordinates of the previous time step. It then combines the generalized momentum, the first-order radiation damping perturbation matrix, and the second-order additional mass perturbation matrix of the previous time step to obtain the actual hydrodynamic action vector. The Hamiltonian function and the canonical equation are established, and the generalized coordinates and generalized momentum of the current step are output. The phase space early warning and intervention module constructs a reduced-dimensional phase space state vector based on the generalized coordinates of the current step and the generalized velocity vector mapped from the generalized momentum of the current step. It generates a dynamic phase space bounding box sequence and performs an intersection operation with a preset static obstacle bounding box. When the operation result is not empty, it sends a braking signal to the programmable logic controller.

[0008] Preferably, the spatial pose parameters include the linear translation coordinates of the component's center of mass along the X, Y, and Z axes in the global geodetic coordinate system, and the yaw, pitch, and roll angles of the component's rotation around the X, Y, and Z axes in the local connected coordinate system, forming a six-dimensional spatial pose state vector; the marine environmental parameters include the significant wave height extracted from the sea state statistical model, the peak period of the wave spectrum, and the average horizontal current velocity within the construction water depth range, forming a three-dimensional marine environmental parameter vector.

[0009] Preferably, the specific steps for constructing the hydrodynamic response tensor field are as follows: Based on a preset spatial step size, the physical boundary limits of the component's translational dimension and the angular limits of yaw, pitch, and roll angles are divided into equidistant or variable-distance segments to generate a set of spatial pose discrete nodes; upper and lower limits for wave height, wave period, and flow velocity are set to generate a set of environmental discrete nodes; the spatial pose discrete node set and the environmental discrete node set are combined using a Cartesian product operation to form a nine-dimensional global computational network; the computational fluid dynamics solver engine is invoked, and the spatial position coordinates and attitude angles corresponding to the discrete nodes are input into the numerical water tank model as fixed rigid body boundary conditions; the wave height, wave period, and flow velocity values ​​are input as wave-generating inlet boundary conditions and flow field loads; translational hydrodynamic components and rotational hydrodynamic torque components are extracted and combined to form the reference hydrodynamic vectors corresponding to the discrete nodes; and the reference hydrodynamic vectors corresponding to all discrete nodes are summarized to construct the hydrodynamic response tensor field.

[0010] Preferably, the specific steps for calculating the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix are as follows: based on the three-dimensional linear potential flow theory, the boundary element method is used to solve the Laplace equation in the water space domain to obtain the radiation velocity potential distribution of the flow field around the component; the radiation velocity potential is integrated on the wet surface of the component to calculate the pressure; the fluid reaction force obtained by the integration calculation is decomposed into components in phase with the component's acceleration and components in phase with the component's velocity; the components in phase with the component's acceleration are combined according to the six-degree-of-freedom cross-coupling relationship to form the second-order additional mass perturbation matrix, and the components in phase with the component's velocity are combined according to the six-degree-of-freedom cross-coupling relationship to form the first-order radiation damping perturbation matrix.

[0011] Preferably, the process of extracting the reference hydrodynamic vector is as follows: a nine-dimensional target independent variable is constructed from marine environmental parameters and spatial pose parameters corresponding to the generalized coordinates of the previous time step; the value of the nine-dimensional target independent variable is compared with a preset discrete node sequence in the hydrodynamic response tensor field to determine the lower and upper limit discrete nodes surrounding the target value; the difference between the target value and the lower limit discrete node value is extracted, and the ratio is calculated with the interval span of the upper and lower limit discrete node values ​​to obtain the one-dimensional interpolation weight coefficient; the hypercube mesh cells surrounding the nine-dimensional target independent variable are determined in the hydrodynamic response tensor field, the reference hydrodynamic vector corresponding to the hypercube mesh cells is extracted, and multilinear interpolation is performed using the one-dimensional interpolation weight coefficient to output the current reference hydrodynamic vector.

[0012] Preferably, the process of obtaining the actual hydrodynamic action vector is as follows: the difference between the generalized momentum of the previous time step and the generalized momentum retained in the time step before that is calculated and the ratio is calculated with a preset discrete time step parameter to obtain the discrete rate of change vector of the generalized momentum with respect to time; the inverse matrix of the preset generalized mass matrix is ​​used to multiply the generalized momentum of the previous time step and the discrete rate of change vector respectively to map them into a generalized velocity vector and a generalized acceleration vector; the radiation damping compensation term is calculated by left-multiplying the generalized velocity vector with the first-order radiation damping perturbation matrix, and the additional mass compensation term is calculated by left-multiplying the generalized acceleration vector with the second-order additional mass perturbation matrix; the radiation damping compensation term and the additional mass compensation term are superimposed on the reference hydrodynamic vector to output the actual hydrodynamic action vector.

[0013] Preferably, the specific steps for establishing the Hamiltonian function and the canonical equation include: defining the total kinetic energy as half of the product of the transpose of the generalized momentum, the inverse of the preset generalized mass matrix, and the generalized momentum; defining the total potential energy as the sum of the gravitational potential energy of the component and the elastic deformation potential energy of the cable calculated based on the linear Hooke's law; constructing a scalar energy function with generalized coordinates and generalized momentum as independent scalar variables by adding the total kinetic energy and the total potential energy, which is the Hamiltonian function; setting the first canonical equation as the derivative of the generalized coordinate column vector with respect to time being exactly the same as the partial derivative of the Hamiltonian function with respect to the generalized momentum; setting the second canonical equation as the sum of the derivative of the generalized momentum column vector with respect to time being the same as the sum of the negative values ​​of the actual hydrodynamic action vector and the partial derivative of the Hamiltonian function with respect to the generalized coordinate column vector.

[0014] Preferably, the specific steps for constructing the reduced phase space state vector are as follows: multiply the generalized momentum of the current step by the inverse of the preset generalized mass matrix to map it into a generalized velocity vector; extract the position coordinates along the horizontal X-axis and Y-axis of the global geodetic coordinate system and the yaw angle around the vertical Z-axis from the generalized coordinates of the current step to form a reduced position vector; extract the translational velocity component and the yaw angular velocity component along the horizontal X-axis and Y-axis from the generalized velocity vector to form a reduced velocity vector; concatenate the reduced position vector and the reduced velocity vector in sequence, and write the horizontal X-axis coordinate, horizontal Y-axis coordinate, yaw angle, horizontal X-axis velocity, horizontal Y-axis velocity and yaw angular velocity values ​​into memory in sequence to form a six-dimensional reduced phase space state vector.

[0015] Preferably, the logic of the intersection operation is as follows: construct a phase space topological interference index model, calculate the geometric projection distance between each dynamic bounding box in the phase space dynamic bounding box sequence and the preset static obstacle bounding box on the unit normal vector of the separation axis; subtract the velocity compensation term obtained by multiplying the phase space velocity sensitivity coefficient by the absolute value of the projection of the relative velocity vector on the unit normal vector of the separation axis from the geometric projection distance, and extract the minimum value among all axial differences as the phase space topological interference index; when the index is not greater than zero, it is determined that topological interference has occurred, and the specific predicted time point of interference is output.

[0016] Preferably, the logic for sending the braking signal is as follows: the specific predicted time point output by the intersection operation is defined as the remaining collision time, and compared with the preset warning intervention time threshold and the limit intervention time threshold; when the remaining collision time is not greater than the warning intervention time threshold but greater than the limit intervention time threshold, a speed limit command message is sent to the hoist frequency converter of the lifting equipment; when the remaining collision time is not greater than the limit intervention time threshold, the minimum braking acceleration is calculated by quotienting the instantaneous composite velocity of the component's current horizontal plane extracted from the state vector of the reduced phase space with the effective time window, and an emergency braking control word containing the minimum braking acceleration is written to the programmable logic controller of the lifting equipment.

[0017] This invention provides a simulation and risk warning system for the construction process of prefabricated wave-breaking walls. It has the following beneficial effects: 1. This invention constructs the hydrodynamic response tensor field offline and pre-calculates the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix. During online operation, it uses multidimensional interpolation to obtain the reference hydrodynamic force and performs algebraic compensation correction. This avoids repeatedly solving complex fluid control equations in each integration step, reduces the online calculation load while ensuring the accuracy of fluid forces, solves the problem of long online calculation time for hydrodynamic response during the hoisting of large components, and improves the real-time performance of construction simulation.

[0018] 2. This invention introduces Hamiltonian functions to establish a system of canonical equations, and uses generalized coordinates and generalized momentum to replace traditional displacement and velocity variables for closed-loop coupled solution. It unifies the kinetic and potential energy of the component motion process in an independent scalar energy function, controls the energy drift or dissipation phenomenon that is prone to occur in the traditional dynamic numerical integration process, realizes the conservation calculation of multibody systems under the coupled action of complex marine environment, and improves the accuracy of state solution.

[0019] 3. This invention constructs a dynamic bounding box sequence in a reduced phase space containing a velocity compensation term. By using a topological interferometry index model, it transforms the traditional static geometric position overlap detection into a dynamic interferometric prediction that includes a time gradient. This allows for the prediction of the specific remaining collision time in advance, and based on this, it outputs matching speed limit or minimum braking acceleration commands to the control system of the lifting equipment. This solves the braking lag problem caused by the lack of kinematic foresight in previous early warning models, and improves the safety intervention capability of actual hoisting operations. Attached Figure Description

[0020] Figure 1 This is a schematic diagram of the system functional architecture according to an embodiment of the present invention; Figure 2 This is a schematic diagram of the system operation method according to an embodiment of the present invention; Figure 3 This is a schematic diagram comparing the total energy drift rate of the dynamic calculation system according to an embodiment of the present invention; Figure 4 This is a schematic diagram comparing the distance evolution during the anti-collision braking process under extreme sea conditions according to an embodiment of the present invention. Detailed Implementation

[0021] The technical solutions in the embodiments of the present invention will be clearly and completely described below with reference to the accompanying drawings. Obviously, the described embodiments are only some embodiments of the present invention, and not all embodiments. Based on the embodiments of the present invention, all other embodiments obtained by those skilled in the art without creative effort are within the scope of protection of the present invention.

[0022] Please see the appendix Figure 1 This invention provides a simulation and risk warning system for the construction process of prefabricated wave barrier walls, including an offline data preprocessing module, an environmental perception and state initialization module, a closed-loop coupled solution module, and a phase space early warning intervention module.

[0023] The offline data preprocessing module is deployed on the computing server and is used to perform dimensionality reduction calculations and store computational fluid dynamics data during the construction preparation phase. The offline data preprocessing module converts the calculation process of high-dimensional physical fields into a matrix data structure.

[0024] The environmental perception and state initialization module communicates with multi-source sensing devices at the construction site. These multi-source sensing devices include wave buoys deployed on the sea surface, underwater acoustic Doppler current profilers, and encoders and inertial measurement units installed on the crane vessel components. The environmental perception and state initialization module is used to continuously collect environmental parameters and the initial motion state of the components during the lifting operation.

[0025] The closed-loop coupled solution module is deployed in the edge computing industrial control computer, establishing data connections with the offline data preprocessing module and the environment perception and state initialization module, respectively. The closed-loop coupled solution module receives environmental parameters and performs low-level multibody dynamics evolution calculations.

[0026] The phase space early warning and intervention module is connected to the closed-loop coupled solution module to obtain the state output results of the dynamic evolution calculation. The output of the phase space early warning and intervention module is connected to the programmable logic controller of the on-site lifting equipment to issue collision avoidance intervention commands when the trigger conditions are met.

[0027] See attached document Figure 2 Based on the above system architecture, the overall workflow of the prefabricated wave barrier construction process simulation and risk early warning system is as follows: First, the offline data preprocessing module uses the spatial pose parameters of the component and the marine environmental parameters as independent variables to extract the baseline hydrodynamic forces and moments through numerical solutions, thereby constructing the hydrodynamic response tensor field. Simultaneously, the offline data preprocessing module calculates the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix. These two matrices reside in the system's memory as constant operators.

[0028] Once the online construction phase begins, the environmental perception and state initialization module reads the current wave height, wave period, and flow velocity data according to the set time sequence. Combining this with the mass distribution matrix, the module transforms the initial motion parameters of the component into the initial generalized coordinates and initial generalized momentum required for the underlying calculations, and inputs this state data into the closed-loop coupled solution module.

[0029] Within each time step, the closed-loop coupled solution module extracts the reference hydrodynamic vector from the hydrodynamic response tensor field based on the data input from the environment perception and state initialization module. The closed-loop coupled solution module also extracts the generalized momentum output from the integral calculation of the previous time step.

[0030] The closed-loop coupled solution module uses the generalized momentum from the previous time step, as well as the radiation damping perturbation matrix and the additional mass perturbation matrix obtained from offline calculation, to perform algebraic correction on the reference hydrodynamic vector and calculate the actual hydrodynamic action vector that includes the radiation effect of component motion.

[0031] Based on the actual hydrodynamic action vector, the closed-loop coupled solution module establishes the Hamiltonian and canonical equations for the component and cable system. The module employs an explicit symplectic integral algorithm to discretize and advance the canonical equations, outputting the generalized coordinates and generalized momentum at the current time step. The module then feeds back the output generalized momentum to the hydrodynamic correction calculation at the next time step, forming a computational closed loop within the system.

[0032] Finally, the phase space early warning and intervention module receives the generalized coordinates output by the closed-loop coupled solution module and the velocity coordinates obtained from the generalized momentum mapping. The phase space early warning and intervention module concatenates the position and velocity to construct a reduced-dimensional phase space state vector of the hoisting component.

[0033] The phase space early warning and intervention module constructs a sequence of dynamic bounding boxes in phase space for the hoisting components based on the reduced-dimensional phase space state vector. This module then performs a high-dimensional Boolean intersection operation between the dynamic bounding box sequence and the inherent set of bounding boxes in phase space of the target wave-breaking wall. When the Boolean intersection result is not empty, the module sends a braking signal to the programmable logic controller (PLC) of the hoisting equipment to perform cable self-locking or yaw angle correction operations.

[0034] During the offline data preprocessing stage, the system performs dimensionality reduction operations on the fluid dynamics data in the computing server. To realize the transformation of the high-dimensional physical field calculation process into a matrix data structure, the system establishes a parametric definition model for the state space and environment space.

[0035] The system constructs a global geodetic coordinate system and a locally connected coordinate system for the components to determine the spatial pose state vector of the prefabricated wave barrier components. The spatial pose state vector is represented by a six-degree-of-freedom generalized coordinate system composed of the component's three-dimensional spatial position coordinates and three-dimensional spatial attitude data. In a specific embodiment, the three-dimensional spatial position coordinates refer to the position components of the component's center of mass in the global geodetic coordinate system. The three-dimensional spatial attitude data is characterized using Euler angles to determine the component's rotational attitude in three-dimensional space.

[0036] The spatial pose state vector is defined as a six-dimensional column vector. Each component of this six-dimensional column vector includes, in turn, the linear translation coordinates of the component's center of mass along the X, Y, and Z axes in the global geodetic coordinate system, and the attitude angle values ​​of the component's rotation around the X, Y, and Z axes in the local connected coordinate system. The specific rotation attitude angles correspond to yaw, pitch, and roll angles, respectively. These six components constitute a six-dimensional state space in the real number domain, used to fully characterize the absolute position and motion attitude of the prefabricated component.

[0037] For the rules for establishing the coordinate system and the calculation of the transformation matrix from the local coordinate system to the global coordinate system, those skilled in the art can use the conventional rigid body kinematic matrix transformation method. The Euler angle transformation equation and coordinate mapping principle are well-known technologies in this field and will not be elaborated here.

[0038] Based on the defined state space, the system further defines a marine environmental parameter vector. This vector consists of physical quantities characterizing the wave and current features of the construction area. In a specific embodiment, the marine environmental parameter vector includes three core parameters: wave height, wave period, and ocean current velocity.

[0039] The marine environmental parameter vector is defined as a three-dimensional column vector. The specific components of this three-dimensional column vector include, in order, a wave height parameter representing wave scale, a wave period parameter representing wave time characteristics, and a current velocity parameter representing the overall ocean current flow. In specific implementation, the wave height parameter is configured as the effective wave height extracted from the sea state statistical model, the wave period parameter is configured as the spectral peak period of the wave spectrum, and the current velocity parameter corresponds to the average horizontal current velocity within the construction water depth range. These three components constitute a three-dimensional environmental space in the real number domain, used to quantify the dynamic boundary conditions required for system solution.

[0040] For the selection of specific statistical models for marine environmental parameters and the measurement methods for hydrological characteristic data, those skilled in the art can calibrate and extract parameters in accordance with engineering hydrological survey specifications. The definition rules of hydrological elements are well-known technologies in this field and will not be elaborated here.

[0041] After defining the parameters of the state space and environment space, the offline data preprocessing module performs calculations to construct the reference hydrodynamic response tensor field based on the established six-degree-of-freedom spatial pose state vector and three-dimensional marine environment parameter vector. This construction process aims to pre-convert the complex partial differential equation hydrodynamic solution into an offline data structure, thereby providing a data foundation for millisecond-level addressing in the online phase.

[0042] The system performs grid-based discretization on the spatial pose range of the prefabricated wave barrier components and the range of marine environmental parameters encountered in the construction area. Specifically, the system sets physical boundary limits for the components in the translational dimensions along three coordinate axes, and angular limits in the rotational dimensions of yaw, pitch, and roll. The system divides these limit ranges into equidistant or variable-distance intervals according to a preset spatial step size, generating a set of discrete nodes for spatial pose. The system obtains historical hydrological extreme values ​​of the construction area, sets upper and lower limits for wave height, wave period, and current velocity, and divides the environmental intervals according to a preset environmental parameter step size, generating a set of discrete environmental nodes. The system performs a Cartesian product operation on the set of discrete spatial pose nodes and the set of discrete environmental nodes, forming a global computational network with nine independent variable dimensions.

[0043] For each multidimensional discrete node in the global computational network, the offline data preprocessing module calls the computational fluid dynamics (CFD) solver engine to configure static and fluid dynamic boundary conditions. Under each independent solution condition, the CFD solver engine inputs the node's corresponding 3D spatial coordinates and 3D spatial attitude angle as fixed rigid body boundary conditions for the prefabricated component into the numerical water tank model. The CFD solver engine also inputs the node's corresponding wave height, wave period, and velocity values ​​as wave-generating inlet boundary conditions and flow field loads for the fluid domain.

[0044] After configuring the boundary conditions, the computational fluid dynamics (CFD) solver engine solves the governing equations for the incompressible fluid in the numerical pool model. By performing area and volume integrals on the hydrodynamic pressure and viscous shear stress at the mesh nodes on the component surface, the system calculates and extracts the steady-state fluid forces acting on the component under this specific orientation and environmental parameter combination. The extracted physical quantities include translational hydrodynamic components along three orthogonal axes of the global geodetic coordinate system, and rotational hydrodynamic torque components around three orthogonal axes of the component's local body coordinate system. These six mechanical components are combined sequentially to form a six-degree-of-freedom reference hydrodynamic vector strictly corresponding to the discrete node.

[0045] The system traverses all discrete nodes in the global computing network, completing numerical solutions and extracting reference hydrodynamic vectors for all discrete working conditions. The offline data preprocessing module arranges and combines all extracted six-degree-of-freedom reference hydrodynamic vectors according to the dimensional index order of their corresponding spatial pose state vectors and marine environmental parameter vectors, constructing a high-dimensional tensor field structure in memory. This high-dimensional tensor field achieves a deterministic data mapping from a nine-dimensional input space of independent variables to a six-dimensional output space of mechanical dependent variables. During the online construction phase, this high-dimensional tensor field is serialized and loaded into the memory of the edge computing industrial control computer, serving as a static data base for table lookup and addressing.

[0046] For the solution methods of fluid control equations, the setting of absorption conditions at the boundary of numerical pools, and the surface pressure integration algorithm involved in offline computational fluid dynamics models, those skilled in the art can use general finite volume method fluid simulation models for processing. The flow field solution mechanism and mesh generation rules are well-known technologies in this field and will not be elaborated here.

[0047] After constructing the baseline hydrodynamic response tensor field, the offline data preprocessing module simultaneously performs the pre-calculation of the perturbation matrix. The baseline hydrodynamic response tensor field reflects the fluid forces acting on the component at a specific orientation and environmental boundary, but does not include the disturbance reaction forces generated by the component's transient motion on the surrounding flow field. To achieve online dimensionality reduction algebraic computation of fluid-structure interaction effects, the system pre-calculates the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix.

[0048] The system imports the 3D geometric model of the prefabricated wave barrier components and meshes the wetted surface of the components at their external boundaries. The system defines the unit motion modes of the components in six degrees of freedom in a local connected coordinate system, including unit translational modes along the three orthogonal axes and unit rotational modes about the three orthogonal axes. The system sets a discrete frequency set covering the typical wave frequency range of the construction area.

[0049] For each discrete frequency and each unit motion mode, the offline data preprocessing module, based on three-dimensional linear potential flow theory, sets boundary value problems that satisfy free surface conditions, seabed boundary conditions, and the impenetrable condition of the wetted surface of the component. The offline data preprocessing module calls the boundary element method to solve the Laplace equation in the water space domain to obtain the radiation velocity potential distribution of the flow field around the component.

[0050] The offline data preprocessing module performs pressure integration calculations on the wetted surface of the component using the obtained radiation velocity potential. The system then decomposes the fluid reaction force obtained from the integration calculation into two independent components according to phase: a component in phase with the component's acceleration and a component in phase with the component's velocity.

[0051] The system extracts and combines the components in phase with the acceleration according to their cross-coupling relationships in the six degrees of freedom, constructing a 6x6 second-order additional mass perturbation matrix. This second-order additional mass perturbation matrix quantifies the inertial drag characteristics generated by the surrounding attached water body during the accelerated motion of the component.

[0052] The system extracts and combines the components in phase with the velocity according to the cross-coupling relationships in the six degrees of freedom directions, constructing a first-order radiation damping perturbation matrix of six rows and six columns. This first-order radiation damping perturbation matrix quantifies the system energy dissipation characteristics caused by the radiation waves generated and propagating outward when the component moves at a certain velocity.

[0053] To reduce the computational dimensionality in the online phase and adapt to the time-domain solver engine, the system extracts the average value of matrix elements within the discrete frequency range, or extracts the matrix element values ​​corresponding to the characteristic frequencies of the design waves, as independent constant physical operators. The calculated first-order radiation damping perturbation matrix and second-order additional mass perturbation matrix are serialized and stored in the memory of the computing device for online phase compensation correction of the reference hydrodynamics.

[0054] For the establishment of the Laplace equation in the three-dimensional potential flow theory, the discrete solution process of the Green's function in the boundary element method, and the pressure integration algorithm of the wetted surface, those skilled in the art can consult the basic literature on marine engineering hydrodynamics to build numerical models. The numerical derivation mechanism of the radiation velocity potential is a well-known technology in this field and will not be elaborated here.

[0055] During the hoisting and construction of the prefabricated wave barrier, the environmental perception and state initialization module is started and runs continuously, performing real-time access operations for multi-source heterogeneous environmental data.

[0056] The system internally establishes a unified global discrete time series as the time benchmark for the collaborative work of various software modules. In a specific embodiment, the discrete calculation time is defined as the product of the time step size and the discrete time step index, where the discrete time step index is a non-negative integer starting from zero and incrementing. This global discrete time series is used throughout the subsequent closed-loop coupled solution and phase space early warning intervention process.

[0057] The environmental perception and state initialization module establishes physical and logical connections with various sensing devices deployed in the work area through a field communication network. The system receives real-time wave surface elevation and wave period data streams from wave buoys and horizontal current velocity data streams measured by an underwater acoustic Doppler current profiler. The system also synchronously receives raw kinematic electrical signals from attitude and displacement measurement devices such as inertial measurement units and encoders installed on lifting equipment and suspension components. Due to differences in internal structure and working principles, the data streams output by these sensing devices differ in format, sampling frequency, and timestamps, forming a multi-source heterogeneous dataset.

[0058] To address data inconsistency issues, the environment perception and state initialization module incorporates a data buffer pool and a timestamp alignment mechanism. The system extracts nodes based on a set global discrete computation time and performs resampling operations on heterogeneous data. For high-frequency physical signals with sampling frequencies higher than the global discrete time series (such as angular velocity and acceleration output from the inertial measurement unit), the system employs low-pass filtering or average downsampling to extract smooth state values ​​aligned with the current computation time, thereby filtering out high-frequency hardware noise. For low-frequency environmental data with update frequencies lower than the global discrete time series (such as wave period statistics updated minute-by-minute by wave buoys), the system uses a zero-order hold or a time-domain polynomial interpolation algorithm to estimate environmental characteristics at the current specific computation time.

[0059] Through the aforementioned time base unification and heterogeneity elimination, the system successfully constructs a real-time marine environmental parameter vector that is strictly aligned with each specific discrete calculation moment. This vector specifically includes the wave height, wave period, and current velocity values ​​after synchronous processing at the current moment. The system synchronously outputs the measured values ​​of the absolute displacement and Euler angles of the components at the current moment, providing a complete and time-consistent external boundary input for the underlying closed-loop coupled calculation.

[0060] For the configuration of the underlying communication interface (such as RS485 bus or industrial Ethernet) of multi-source sensing devices and the hardware clock synchronization technology based on network time protocol, those skilled in the art can use the conventional data bus communication standard of industrial IoT to build it. Its underlying message parsing and clock calibration mechanism are well known technologies in this field and will not be described in detail here.

[0061] After completing the real-time access to multi-source heterogeneous environmental data, the environmental perception and state initialization module performs an initial state mapping operation on the prefabricated wave barrier components at the start calculation time of the hoisting operation. This operation converts the conventional kinematic parameters output by the sensing devices into the generalized phase space initial values ​​required by the underlying symplectic geometric dynamics engine.

[0062] The system reads the measurement data output by the inertial measurement unit and encoder on the crane vessel at the initial moment. The system extracts the three-dimensional spatial coordinates of the prefabricated wave-breaking wall component in the global geodetic coordinate system at that moment, as well as the three Euler angles reflecting its three-dimensional rotational attitude. The system sequentially combines these six position and attitude components to construct a six-dimensional initial generalized coordinate column vector. Simultaneously, the system extracts the three-dimensional translational velocity component of the component's center of mass at that initial moment, as well as the angular velocity components of rotation around the three axes of the local connected coordinate system. The system concatenates these six velocity components to construct a six-dimensional initial generalized velocity column vector.

[0063] To establish the physical mapping between kinematic and dynamic parameters, the system calls a pre-built generalized mass matrix of the prefabricated wave-breaking wall component from memory. This generalized mass matrix is ​​defined as a 6x6 real square matrix. The first three elements on the main diagonal of the generalized mass matrix represent the inherent physical mass of the component, and the last three elements represent the principal moments of inertia of the component about the local integral coordinate axes. The off-diagonal elements of the matrix are used to characterize the inertia product characteristics caused by the asymmetry of the component's mass distribution. This matrix fully quantifies the magnitude of the component's inertia in resisting external forces and torques to maintain its original state of motion in six degrees of freedom.

[0064] The system multiplies the constructed six-dimensional initial generalized velocity column vector by the sixth-order generalized mass matrix. By performing this linear algebraic matrix multiplication, the system calculates the initial generalized momentum column vector of the prefabricated wave barrier component at the initial moment. The calculated initial generalized momentum column vector consists of three linear momentum components and three angular momentum components, characterizing the absolute motion state of the component at the initial moment from the dimensions of energy and momentum.

[0065] Building upon this, the environment perception and state initialization module packages and encapsulates the constructed initial generalized coordinate column vector and initial generalized momentum column vector. The system uses the encapsulated state dataset as the initial conditions for dynamic evolution calculations, transmitting it unidirectionally and inputting it into the closed-loop coupled solution module. The underlying multibody dynamic evolution engine uses this initial state as the absolute starting point for time-domain integration, initiating the subsequent numerical recursive solution process over the time series.

[0066] For the time-domain integral conversion method of the original acceleration and angular velocity signals of the inertial measurement unit, and the theoretical calculation method of the rotational inertia tensor and product of inertia of complex irregular rigid bodies, those skilled in the art can deduce and physically calibrate them based on the conventional criteria of theoretical mechanics and spatial rigid body dynamics. The acquisition of the geometric properties of rigid body mass and the conversion rules of basic mechanical parameters are well-known technologies in this field and will not be elaborated here.

[0067] After entering the online closed-loop coupled solution stage, the closed-loop coupled solution module performs real-time addressing of the reference hydrodynamic vector for each discrete calculation moment.

[0068] The closed-loop coupled solution module receives the spatial pose state vector of the prefabricated wave-breaking wall component at the current moment, as well as the corresponding marine environmental parameter vector. The spatial pose state vector contains the component's current three-dimensional spatial position coordinates and three-dimensional spatial attitude data. The marine environmental parameter vector contains the wave height, wave period, and current velocity at the current calculation moment. These nine parameters together constitute the nine-dimensional target independent variables for addressing the system in the hydrodynamic response tensor field at the current moment.

[0069] The closed-loop coupled solution module inputs the nine-dimensional target independent variable into a high-dimensional hydrodynamic response tensor field pre-loaded in memory. For any specific dimension among the nine dimensions, the system compares the specific value of the current target independent variable with the preset discrete node sequence for that dimension in the tensor field. The system determines two adjacent boundary nodes surrounding the target value in the discrete node sequence, defining them as the lower and upper bound discrete nodes for that dimension, respectively.

[0070] The system synchronously calculates the relative bias ratio of the target value between the lower and upper discrete nodes. The system extracts the difference between the target value and the lower discrete node value and divides it by the interval span between the upper and lower discrete node values. The calculated ratio is the normalized distance of the target value in that specific dimension, and this normalized distance is configured by the system as the single-dimensional interpolation weight coefficient for subsequent multi-dimensional operations.

[0071] After locating the upper and lower boundary discrete nodes in all nine dimensions, the system determines a unique hypercube mesh cell within the nine-dimensional tensor field computational network that encloses the current nine-dimensional target independent variable by combining the coordinates of these boundary nodes. This high-dimensional hypercube mesh cell consists of 512 vertex nodes. Using the multidimensional coordinate combination of these 512 vertex nodes as a memory address index, the system extracts the corresponding 512 offline pre-calculated baseline hydrodynamic vectors from the underlying data structure of the tensor field.

[0072] The closed-loop coupled solution module invokes a preset tensor interpolation engine, utilizing the previously calculated single-dimensional interpolation weight coefficients for each dimension to perform multilinear interpolation operations on the extracted 512 reference hydrodynamic vectors. The system performs linear weighted summation operations on adjacent node vectors level by level according to the nine-dimensional spatial hierarchy. After multiple dimensionality reduction combinations and weighted calculations, the system outputs a six-dimensional column vector that strictly corresponds to the current continuous spatial pose and environmental parameter state. This output six-dimensional column vector is defined as the reference hydrodynamic vector at the current moment. The reference hydrodynamic vector characterizes the basic fluid forces and moments experienced by the prefabricated component under the current transient pose and wave-current environment boundary, excluding the influence of its own absolute motion.

[0073] For the array index traversal optimization algorithm in the multidimensional grid search process, and the continuous address reading and vectorized multiplication and addition operation mechanism of high-dimensional tensor multilinear interpolation in the underlying hardware memory, those skilled in the art can use conventional computational mathematical addressing methods and computer memory management technology for software configuration. The principle of high-dimensional array index and discrete data linear smoothing is a well-known technology in this field, and will not be elaborated here.

[0074] After obtaining the reference hydrodynamic vector at the current discrete calculation moment, the closed-loop coupled solution module calls the momentum feedback perturbation compensation model to perform the reverse correction calculation of the fluid-structure interaction effect.

[0075] The closed-loop coupled solution module extracts the generalized momentum vector from the integral output of the previous time step and the generalized momentum vector retained from the time step before that from the memory area of ​​the underlying dynamics solution engine. The module employs a first-order finite-difference numerical algorithm to subtract the generalized momentum vector from the previous time step from the generalized momentum vector from the time step before that. The system divides this difference by a preset discrete time step parameter to calculate the discrete rate of change vector of the generalized momentum with respect to time. This discrete rate of change vector is used to characterize the transient evolution trend of the component's motion state from an algebraic perspective.

[0076] The closed-loop coupled solution module calls the generalized mass matrix established during the state initialization phase in memory and performs the inverse matrix calculation operation of this generalized mass matrix. Based on this inverse matrix, the system constructs two independent state mapping operators. The first operator performs linear matrix multiplication, left-multiplying the extracted generalized momentum vector from the previous time step by the inverse matrix of the generalized mass matrix, thereby reducing the dimensionality of the momentum parameters in the dynamic phase space to a generalized velocity column vector in Cartesian coordinates. The second operator executes the same logical structure, left-multiplying the calculated discrete rate of change of generalized momentum by the inverse matrix, mapping it to a generalized acceleration column vector at the corresponding time step.

[0077] After completing the aforementioned state mapping, the closed-loop coupling solution module reads the high-dimensional matrix constants resident in memory during system initialization, specifically including the first-order radiation damping perturbation matrix and the second-order added mass perturbation matrix. Based on the generalized velocity column vector and generalized acceleration column vector obtained from the state mapping, and combined with the perturbation matrix, the system performs real-time algebraic compensation calculations on the reference hydrodynamic vector. This calculation logic constitutes a closed-loop perturbation compensation model, and the calculated output includes the actual hydrodynamic action vector containing radiation damping and added mass effects. The specific expression for the actual hydrodynamic action vector is as follows: ; in, For the current time step The actual hydrodynamic vector output below; For the current discrete computation time, where For discrete time step index; For the current time step The baseline hydrodynamic vector extracted by multilinear interpolation addressing; The first-order radiation damping perturbation matrix is ​​pre-calculated and kept constant; For the previous time step Extracted generalized momentum vector; This is the first operator that converts generalized momentum into generalized velocity, and its overall output is a column vector of generalized velocities. The second-order additional mass perturbation matrix is ​​pre-calculated and kept constant; For the previous time step The discrete rate of change vector of the lower generalized momentum with respect to time; This is the second operator that converts the discrete rate of change of generalized momentum into generalized acceleration, and its overall output is a column vector of generalized acceleration.

[0078] This model overcomes the computational limitations of fixed-boundary lookup tables, achieving real-time dynamic adaptive adjustment of the transient motion state of components to the hydrodynamic loads they experience. Regarding the aforementioned general model functionality, in its specific implementation, matrix multiplication and vector addition / subtraction instructions within the formulas are encapsulated into a low-level basic linear algebra subroutine library. The multi-core CPU of the edge computing industrial control computer schedules these subroutine libraries for concurrent thread computation, thereby ensuring the output of the actual hydrodynamic action vector within milliseconds in each discrete computation step.

[0079] For numerical error control of the first-order finite difference method and conventional algebraic algorithms for matrix inversion, those skilled in the art can use standard numerical analysis methods to allocate computational resources. The basic operational rules of matrix algebra and the handling of truncation errors are well-known technologies in this field and will not be elaborated here.

[0080] After calculating the actual hydrodynamic action vector that includes the radiation effect of component motion, the closed-loop coupled solution module performs the construction of the Hamiltonian of the multibody system and the formulation of the canonical equations. This module treats the prefabricated wave-breaking wall components and the lifting cables as a mutually coupled holistic mechanical system, and establishes the underlying dynamic evolution rules through energy conservation and variational principles.

[0081] The closed-loop coupled solution module constructs the total kinetic energy of the system based on the generalized momentum and generalized mass matrix. The total kinetic energy of the system is defined as half multiplied by the transpose of the generalized momentum column vector, the inverse of the generalized mass matrix, and the product of the generalized momentum column vector. The momentum-based kinetic energy representation directly bypasses the velocity parameter in traditional Newtonian mechanics and satisfies the data format requirements for conjugate variables in symplectic geometric operations.

[0082] The closed-loop coupled solution module calculates the total potential energy of the system in the current spatial pose. The total potential energy of the system is composed of the superposition of the gravitational potential energy of the components and the elastic deformation potential energy of the cables. The gravitational potential energy of the components is obtained by multiplying the inherent physical mass of the components, the gravitational acceleration constant, and the vertical height coordinates of the component's center of mass in the global geodetic coordinate system.

[0083] The elastic deformation potential energy of the cables is calculated based on the linear Hooke's law. The system extracts the real-time spatial length of each cable corresponding to the generalized coordinates of the current time step, and subtracts its initial stress-free length to obtain the deformation elongation. The elastic potential energy of a single cable is configured as half multiplied by the product of the cable's equivalent tensile stiffness coefficient and the square of the deformation elongation. The system sums the elastic potential energy of all cables connected to the components to obtain the total elastic deformation potential energy of the system.

[0084] The closed-loop coupled solution module adds the total kinetic energy and total potential energy calculated above to construct a Hamiltonian function characterizing the total energy distribution of the component and cable system. This Hamiltonian function is a scalar energy function with generalized coordinates and generalized momentum as independent scalar variables.

[0085] After establishing the Hamiltonian function, the closed-loop coupled solution module introduces the Hamiltonian canonical equations to describe the dynamic evolution trajectory of the system in phase space. For the position dimension evolution, the first canonical equation of the system configuration stipulates that the derivative of the generalized coordinate column vector with respect to time is strictly equal to the partial derivative of the Hamiltonian function with respect to the generalized momentum column vector.

[0086] For momentum dimension evolution, the closed-loop coupled solution module introduces the actual hydrodynamic vector output from the aforementioned algebraic compensation calculation as a generalized non-conservative external force term into the canonical system. The second canonical equation configured in the system stipulates that the time derivative of the generalized momentum column vector is equal to the sum of the actual hydrodynamic vector and the negative values ​​of the partial derivatives of the Hamiltonian function with respect to the generalized coordinate column vector. This computational logic couples the fluid load into a conservative Hamiltonian dynamics framework, generating the continuous-time governing equations required for subsequent numerical integration derivation.

[0087] For the partial derivatives of the generalized coordinates of the system with respect to the spatial geometric relationship of the cable length, and the experimental calibration of the equivalent tensile stiffness of multi-strand steel wire ropes, those skilled in the art can use conventional kinematic derivations in analytical mechanics and material mechanics testing benchmarks to configure physical quantities. The basic functional expansion and partial differential calculation of the system potential energy are well-known techniques in this field and will not be elaborated here.

[0088] After constructing the Hamiltonian and formulating the canonical equations for the multibody system, the closed-loop coupled solution module employs an explicit symplectic integral algorithm to perform discretization and progressive calculations on the continuous-time control equations. This progressive calculation aims to recursively solve the evolution trajectory of the prefabricated wave-breaking wall components in the generalized phase space within the time domain, according to a given discrete time step.

[0089] The closed-loop coupled solution module extracts the generalized coordinates and generalized momentum of the current time step, and calculates the conservative restoring force inside the system by combining the partial derivative of the Hamiltonian function with respect to the generalized coordinates. The closed-loop coupled solution module then performs vector superposition of this conservative restoring force with the actual hydrodynamic action vector output from the previous stage to obtain the sum of the total generalized non-conservative external forces and conservative forces acting on the system under the current generalized coordinates.

[0090] The closed-loop coupled solution module performs a numerical integration update operation on the momentum dimension. The system multiplies the total force vector obtained above by a preset discrete time step to obtain the momentum increment in the current time step. The system then vector-adds this momentum increment to the generalized momentum of the previous time step to obtain the updated generalized momentum for the current time step after time evolution.

[0091] After obtaining the updated generalized momentum, the closed-loop coupled solution module performs a numerical integration update operation on the coordinate dimensions. The system calls the inverse of the generalized mass matrix and performs a linear matrix multiplication operation with the updated generalized momentum to extract the updated generalized velocity of the component in the current time step. The system multiplies this updated generalized velocity by the discrete time step to obtain the position and attitude space increment within the current time step. The system adds this increment to the generalized coordinates of the previous time step to obtain the updated generalized coordinates of the current time step.

[0092] In the specific implementation of the functional generalizations described above, the value of the discrete time step parameter is controlled by the system's highest natural frequency. The edge computing industrial control computer calculates the system's high-frequency response period based on the maximum equivalent tensile stiffness of the lifting cable and the minimum principal moment of inertia of the components. The discrete time step is configured to be strictly less than one-tenth of this high-frequency response period to meet the conditional stability requirements of explicit numerical integration. An explicit symplectic integral algorithm directly bypasses the iterative solution process of the nonlinear algebraic equation system, eliminating the high computational overhead of traditional implicit algorithms in matrix inversion and Jacobian matrix updates, thus ensuring millisecond-level real-time output under the condition of limited computing power for the industrial control computer.

[0093] The closed-loop coupled solution module outputs the calculated updated generalized coordinates and updated generalized momentum as the final state data at the current discrete calculation moment. This state data is unidirectionally transmitted to the phase space early warning and intervention module for collision avoidance assessment. The closed-loop coupled solution module then feeds the updated generalized momentum back to the system memory buffer, overwriting the historical data from the previous time step, and uses this as input for hydrodynamic correction calculations in the next discrete calculation step, thus forming a continuous closed loop in the time series.

[0094] For the proof of the geometric property of the explicit symplectic integral algorithm in maintaining the uncompressible phase volume in long-term time-domain solutions, and the frequency-domain derivation method of the numerical stability boundary conditions, those skilled in the art can perform convergence determination and step size optimization based on the computational multibody dynamics specification. Its symplectic preservation characteristics and numerical dispersion control mechanism are well-known technologies in this field and will not be elaborated here.

[0095] After the closed-loop coupled solution module outputs the generalized coordinates and generalized momentum of the current time step, the phase space early warning and intervention module receives the state output results of the above dynamic evolution calculation and performs the splicing and mapping operation of the reduced phase space state vector.

[0096] The phase space early warning and intervention module calls the inverse matrix of the generalized mass matrix pre-stored during system initialization. The system multiplies the received six-dimensional generalized momentum column vector by this inverse matrix on the left, and transforms the momentum space parameter mapping back to a six-dimensional generalized velocity column vector in Cartesian coordinates through linear algebraic multiplication. This generalized velocity column vector contains the three-dimensional translational velocity component of the component's center of mass and the three-dimensional rotational angular velocity component about the local connected coordinate axes. This transformation operation restores the abstract symplectic geometric momentum parameters to intuitive kinematic velocity parameters, providing a data foundation for subsequent physical collision boundary assessment.

[0097] After acquiring the generalized coordinates and generalized velocity containing six degrees of freedom, the system performs dimensionality reduction extraction on the high-dimensional state space. During the underwater hoisting and assembly of prefabricated wave-breaking walls, the lowering displacement and velocity of the components along the vertical Z-axis are typically rigidly constrained and strictly controlled by the hoisting equipment's winch. Simultaneously, the roll and pitch attitudes of the components are strongly constrained by the tension of multiple slings, resulting in a low probability of significant overturning. Based on these construction physics characteristics, the phase space early warning and intervention module extracts the component's position coordinates along the horizontal X and Y axes in the global geodetic coordinate system, as well as the yaw angle around the vertical Z-axis, from the six-dimensional generalized coordinates. These three parameters constitute the dimensionality-reduced position vector, representing the component's absolute position and orientation in the horizontal plane.

[0098] The phase space early warning and intervention module simultaneously extracts the translational velocity components along the horizontal X and Y axes, and the yaw angular velocity component around the vertical Z axis, from the six-dimensional generalized velocity column vector. These three parameters constitute a reduced-dimensional velocity vector, representing the instantaneous speed and rotational tendency of the component in the horizontal plane.

[0099] The phase space early warning and intervention module sequentially concatenates the extracted dimensionality-reduced position vector and velocity vector. In the specific implementation of functional lower-level features, the system allocates a contiguous storage area in the computing device's memory and sequentially writes floating-point values ​​for the horizontal X-axis coordinate, horizontal Y-axis coordinate, yaw angle, horizontal X-axis velocity, horizontal Y-axis velocity, and yaw angle velocity. Through the above multidimensional array slicing extraction and memory address concatenation, the system constructs a six-dimensional dimensionality-reduced phase space state vector.

[0100] This dimensionality-reduced phase space state vector fully preserves the main collision-sensitive degrees of freedom within the horizontal plane of the component while eliminating computational redundancy in non-sensitive controlled dimensions. This dimensionality reduction significantly reduces the geometric complexity of subsequent polyhedral intersection operations, thereby meeting the computational timeliness requirements for real-time collision avoidance warning.

[0101] For the projection transformation algorithm from a six-degree-of-freedom space to a specific planar motion dimension, and the memory allocation and array concatenation mechanism for variable types in the underlying software system, those skilled in the art can use conventional computer data structure operations to write programs. The matrix dimension pruning and array recombination rules are well-known technologies in this field and will not be elaborated here.

[0102] After acquiring the reduced-dimensional phase space state vector, the phase space early warning and intervention module performs phase space bounding box sequence construction and interference topology determination calculation. By introducing a lead factor in the velocity dimension, this module transforms the traditional static spatial geometric collision avoidance assessment into a dynamic phase space topology detection that considers motion trends.

[0103] The system presets a collision avoidance prediction time window and divides this window into multiple discrete future prediction moments according to a fixed prediction time step. For each future prediction moment, the phase space early warning and intervention module uses the current position coordinates and current velocity parameters in the reduced phase space state vector to perform kinematic forward inference.

[0104] The system extracts the horizontal X-axis coordinates, horizontal Y-axis coordinates, and yaw angle from the reduced-dimensional position vector, and the horizontal X-axis velocity, horizontal Y-axis velocity, and yaw angle velocity from the reduced-dimensional velocity vector. The system multiplies the velocity parameters by the predicted time difference to calculate the position increment and angle increment corresponding to that predicted moment. The system adds these increments to the initial position and initial attitude at the current moment to deduce the predicted centroid coordinates and predicted yaw angle of the component at a specific future predicted moment.

[0105] Based on the calculated predicted centroid coordinates and predicted yaw angle, the system constructs a directed bounding box for the prefabricated wave-breaking wall component within the horizontal plane of the global geodetic coordinate system. The geometric dimensions of the directed bounding box are determined by the outer contour boundary of the component's 3D CAD model projected onto the horizontal plane, specifically representing a rectangular planar region with length and width. The system constructs a two-dimensional rotation matrix by predicting the yaw angle, mapping the local coordinate system vertices of the bounding box to the global geodetic coordinate system, thereby determining the coordinates of the four global vertices of the bounding box at a specific prediction time. The series of directed bounding boxes generated sequentially over the entire prediction time window together constitute the phase space bounding box sequence of the component.

[0106] After constructing the phase space bounding box sequence, the system extracts the geometric position data of adjacent installed components or construction boundaries to construct static obstacle bounding boxes. The phase space early warning and intervention module performs interference topology determination for each dynamic bounding box and static obstacle bounding box in the phase space bounding box sequence. To quantify the amplification effect of component motion velocity on collision risk, the system constructs a phase space topological interference index model, the specific expression of which is: ; in, To predict at a specific time The phase space topological interference index of the output; Predicting moments for the future; For the preset set of separation axes The unit normal vector of the separation axis in the middle; Indicates a pre-defined set of separation axes All separation axis unit normal vectors Iterate through the results and take the minimum value. For the component at the predicted time The directed bounding box; For static obstacle bounding boxes; The independent variables of the projection spacing function include the dynamic bounding box, the static obstacle bounding box, and the selected separation axis; Indicates the oriented bounding box of the component With obstacle enclosure unit normal vector of the separation axis Geometric projection spacing on; The preset phase space velocity sensitivity coefficient is used to adjust the physical weight of velocity parameters in interference and determination. Let V be the horizontal velocity vector of the component relative to the obstacle. Represents the relative velocity vector unit normal vector of the separation axis The absolute value of the projection on; This indicates the velocity compensation term on the specified separation axis, determined by the phase space velocity sensitivity coefficient. The result, obtained by multiplying the absolute value of the relative velocity projection above, is used to subtract the dynamic approaching amount generated by the motion trend from the static geometric spacing.

[0107] The system iterates through all potential separation axes in the separation axis set, calculating the difference between the projected spacing and the velocity compensation term on each axis. The system extracts the minimum value among all axial differences as the phase space topological interference index for that prediction moment. When the index is greater than zero, it indicates the existence of a clear topological separation boundary in phase space, and the system determines there is no collision risk at the current prediction moment. When the index is less than or equal to zero, it indicates that the bounding box has projected overlap considering velocity compensation on all separation axes, and the system determines that topological interference has occurred, triggering an interference warning signal.

[0108] For the specific implementation of the aforementioned interference topology determination features, the phase space early warning and intervention module calls the underlying floating-point arithmetic unit to perform vector dot product and extreme value comparison. In the loop structure that traverses the prediction time window, once the system detects that the phase space topological interference index is less than or equal to zero at a certain prediction time, it executes the loop interruption jump instruction and outputs the specific prediction time point of interference and the minimum projection distance, thereby avoiding subsequent redundant calculations and ensuring the effective allocation of industrial control computer resources.

[0109] For the geometric projection algorithm of the separating axis theorem in polygon collision detection, the trigonometric function calculation method of the two-dimensional rotation matrix, and the selection rules of the separating axis normal vector set, those skilled in the art can use the conventional interferometric detection library of computer graphics to compile the code. The polygon vertex mapping and axial projection scalar calculation are well known technologies in this field and will not be elaborated here.

[0110] After constructing the phase space bounding box sequence and determining the interference topology, the phase space early warning and intervention module executes the risk warning classification triggering and intervention logic. This logic aims to output differentiated control commands and alarm information to construction lifting equipment and operators based on the urgency of the predicted interference occurrence.

[0111] The system extracts the specific predicted time point of interference from the interference topology determination calculation and defines it as the remaining collision time. The phase space early warning and intervention module pre-configures three time threshold parameters in memory, defined as the prompt intervention time threshold, the warning intervention time threshold, and the extreme intervention time threshold, respectively. These three time threshold parameters exhibit a strictly decreasing algebraic relationship, corresponding to different collision risk stages of urgency.

[0112] The system compares the calculated remaining collision time with the three time threshold parameters mentioned above step by step. If the remaining collision time is greater than the intervention time threshold, the system determines that the current construction status is within a safe range, maintains the normal operating permissions of the lifting equipment, and displays a green safety status indicator on the host computer monitoring interface.

[0113] If the remaining collision time is less than or equal to the prompt intervention time threshold, but greater than the warning intervention time threshold, the system triggers a primary warning. The phase space warning intervention module sends a trigger command to the hardware alarm terminal in the crane cab via the field communication network. Upon receiving the command, the alarm terminal activates a flashing yellow warning light and triggers a buzzer at a preset first frequency to alert the operator to closely monitor the lowering posture of the component.

[0114] If the remaining collision time is less than or equal to the warning intervention time threshold, but greater than the extreme intervention time threshold, the system triggers a medium-level warning. The system switches the warning light on the alarm terminal to red and triggers a continuous buzzer at a higher second frequency. The phase-space warning intervention module sends a speed-limiting command message to the hoist frequency converter of the lifting equipment via industrial Ethernet. This speed-limiting command restricts the maximum output frequency of the hoist motor at the control layer, forcibly reducing the lowering and translating speed of the prefabricated wave barrier components at the hardware level.

[0115] If the remaining collision time is less than or equal to the critical intervention time threshold, the system determines that there is an extremely high risk of immediate collision and triggers the highest level of intervention logic. The phase space early warning intervention module bypasses the operator's regular operating handle command port and directly writes the emergency braking control word to the main programmable logic controller of the lifting equipment. To ensure that the component comes to a complete stop before collision, the system synchronously runs a braking deceleration calculation model to calculate the minimum braking acceleration required. The specific calculation expression is as follows: ; in, To calculate the minimum braking acceleration required for the system output; The scalar magnitude of the instantaneous composite velocity of the component at the current horizontal plane, extracted from the state vector in the reduced phase space. The predicted remaining collision time extracted by the system; The inherent mechanical delay time experienced by the lifting equipment from receiving the braking network electrical signal to the complete locking of the mechanical brake device; It represents the effective time window that can actually be used to perform the physical deceleration and braking process after deducting the fixed mechanical delay.

[0116] After receiving the calculated minimum braking acceleration parameters and the emergency braking control word, the programmable logic controller (PLC) drives the servo drives of the crane winch and slewing mechanism to perform electrical reverse braking. Simultaneously, the PLC outputs a relay closing signal, activating the mechanical hydraulic brake circuit and locking the prefabricated wave-breaking wall component in its current spatial position. This intervention cuts off the kinematic path of the component towards the obstacle, avoiding an underwater rigid collision. After the intervention is completed, the system records the current locked state data to a hard disk log file and displays a reset confirmation window on the host computer interface. The system can only resume normal operation after on-site manual verification to eliminate the risk and input of the authorization password.

[0117] For the underlying communication message format of the programmable logic controller of the lifting equipment, the hardware braking circuit design of the servo motor, and the relay wiring method of the audible and visual alarm, those skilled in the art can use industrial automation control standards to network the equipment and program it. The parsing instructions of the electrical control word and the execution principle of the mechanical brake are well-known technologies in this field and will not be elaborated here.

[0118] Specific application examples: Example: Specific application of underwater hoisting project of prefabricated wave barrier in a deep-water port; Scenario Deployment: This embodiment is applied to the underwater hoisting operation of a prefabricated wave barrier in a deep-water port breakwater construction section. The average operating water depth in the target sea area is 25 meters, the hoisting component is a 1200-ton reinforced concrete caisson, and the hoisting operation is carried out by a large crane vessel with a maximum rated lifting capacity of 2000 tons.

[0119] Hardware configuration: Offline computing cluster: The computing server is deployed in the shore-based computing center and has a built-in hydrodynamic solution engine with five million grid nodes.

[0120] Multi-source sensing equipment: wave buoys deployed in the construction area, acoustic Doppler current profilers deployed underwater, inertial measurement units and high-precision wire encoders installed on the boom and caisson components of the crane vessel.

[0121] Edge control equipment: An edge computing industrial control computer installed in the cockpit of the crane ship is directly connected to the main control programmable logic controller (PLC) of the crane equipment via industrial Ethernet, and is connected to an external audible and visual alarm terminal and a mechanical hydraulic brake execution circuit.

[0122] Evolution of hoisting operations and data interaction process: Step 1: Offline preprocessing and state initialization; The offline computing server pre-calculates and generates a high-dimensional hydrodynamic response tensor field based on the set translational and rotational pose intervals and historical statistical boundaries of the marine environment. The system simultaneously calculates the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix, and serializes and loads them into the memory of the edge computing industrial control computer.

[0123] Environmental awareness: At a specific moment of hoisting and lowering, the edge computing industrial control computer reads real-time environmental data, with an effective wave height of 1.8 meters, a wave period of 6.5 seconds, and a horizontal flow velocity of 1.2 meters per second.

[0124] State extraction: The system extracts the initial yaw angle and centroid coordinates of the components measured by the inertial measurement unit of the crane ship.

[0125] Algebraic mapping: The system combines the generalized mass matrix pre-stored in memory to convert the above position and attitude data into the underlying initial generalized coordinates and initial generalized momentum.

[0126] Step 2: Closed-loop coupling solution and phase space dimensionality reduction; the edge computing industrial control computer starts the multibody dynamics numerical integration with a discrete time step of 10 milliseconds.

[0127] Addressing and Correction: The system addresses and extracts the reference hydrodynamic vector in the high-dimensional hydrodynamic response tensor field, extracts the generalized momentum of the previous time step, and calculates the actual hydrodynamic action vector including the radiation effect by combining the first-order radiation damping and the second-order additional mass perturbation matrix.

[0128] Integral propulsion: The system constructs the Hamiltonian and canonical equations for the components and cables, and calls an explicit symplectic integral algorithm to output the generalized coordinates and generalized momentum updated at the current time step. Influenced by strong ocean currents, the dynamic feedback from the underlying layers causes the components to drift laterally at a speed of 0.3 meters per second along the horizontal X-axis.

[0129] Feature dimensionality reduction: The system extracts the horizontal coordinates, yaw angles and corresponding translation and rotation velocity components of the components from the six-dimensional space output, and then concatenates them in memory to generate a six-dimensional reduced phase space state vector.

[0130] Step 3: Interference topology determination and hierarchical early warning intervention; the phase space early warning intervention module uses the reduced-dimensional phase space state vector to establish a collision avoidance prediction mechanism.

[0131] Forward extrapolation: Within the next 5-second prediction time window, the system calculates the predicted centroid coordinates and predicted yaw angle of the components based on the velocity lead, and constructs a dynamic bounding box sequence in phase space.

[0132] Topological interference: The system calculates the geometric projection distance between the dynamic bounding box and the installed static wave-breaking wall bounding box on the separation axis, deducting the velocity compensation term. At the predicted time of 2.5 seconds, due to the lateral drift trend, the phase space topological interference index changes from positive to negative, and the system determines that phase space topological interference has occurred.

[0133] Control Intervention: The system extracts the remaining collision time as 2.5 seconds, determining it to be within the warning intervention time threshold range. The industrial control computer sends a speed limit command message to the winch inverter to force a speed reduction, and simultaneously triggers the cab warning light to sound continuously.

[0134] Ultimate braking: As the ocean current continues to impact, the remaining collision time rapidly shortens, reaching the 1.5-second ultimate intervention time threshold. The system bypasses the manual control handle, calculates the minimum braking acceleration required to be 0.5 meters per square second, writes the emergency braking control word to the PLC, and instantly activates the mechanical hydraulic brake circuit, rigidly locking the component at the current water depth.

[0135] Experimental verification and effect comparison: To verify the actual effect of the invention, a 60-minute continuous hoisting simulation comparative experiment was conducted in the aforementioned deep-water port construction area under extreme sea conditions (wave height suddenly increased to 2.5 meters, and current velocity suddenly increased to 1.8 meters / second).

[0136] Control group: The dynamic equations were solved using the traditional fourth-order Runge-Kutta numerical integration algorithm. The collision avoidance mechanism relied on the static three-dimensional spatial distance of the component at the current moment (the preset safe distance threshold was 1.0 meter).

[0137] Experimental group: The system based on Hamiltonian explicit symplectic integral and dynamic bounding box interference topology determination of the present invention was activated.

[0138] Experimental data presentation: See attached document Figure 3 Chart description: The horizontal axis represents the duration of evolution (0-60 minutes), and the vertical axis represents the total energy drift rate of the system (percentage).

[0139] Data Interpretation: The dashed line marked with a hollow triangle (control group) in the graph exhibits numerical dissipation when handling high-frequency wave loads. By the 45-minute integral evolution period, the energy drift rate had exceeded the tolerance limit of 5%, showing a nonlinear divergence trend. The solid black line (experimental group) strictly maintained the volume-preserving characteristics of the phase space. Throughout the entire 60-minute evolution calculation period, the total system energy exhibited bounded, minute oscillations, and the global energy drift rate was consistently limited to an extremely low error range of 0.1%.

[0140] Experimental results: The experimental group effectively overcame the artificial damping defects of traditional algorithms by using explicit symplectic integrals, and achieved higher dynamic calculation accuracy and numerical stability under long-term continuous simulation conditions.

[0141] See attached document Figure 4 The chart describes the evolution time series of the approaching obstacle (in seconds) on the horizontal axis and the minimum relative distance between the component and the outer enclosure of the static obstacle (in meters) on the vertical axis. The chart shows the distance change over time during the collision avoidance braking process under extreme sea conditions.

[0142] Data Interpretation: The light gray dashed line marked with a hollow circle (control group) shows that the system only triggered the alarm when the physical distance to the component dropped to the 1.0-meter threshold. Due to the neglect of the component's significant hydrodynamic inertia, the component continued to slide forward after the mechanical brake action was executed, eventually causing the broken line to cross the zero-distance line and drop to a negative value, indicating physical overlap (collision). The black solid line marked with a solid square (experimental group) shows that the phase space topological interference index translates the motion trend into a warning lead time, triggering the emergency braking control word 2.5 meters before the physical distance. The broken line gradually flattens after braking, eventually stabilizing at 0.8 meters above the zero-distance line, providing sufficient deceleration buffer window for the brake and successfully avoiding interference.

[0143] Experimental results: The experimental group solved the failure hysteresis problem of the static collision avoidance model under inertial slip by using phase space dimensionality reduction and dynamic velocity compensation.

[0144] Effect Comparison Summary Table

[0145] Conclusion: Experimental results show that this invention overcomes the bottleneck of real-time fluid-structure interaction calculation under complex marine boundaries by using a hydrodynamic perturbation tensor field and Hamiltonian explicit symplectic integral architecture; and achieves accurate topological prediction of physical collision risks by constructing a phase space dynamic bounding box model. The closed-loop risk intervention system built upon this system can output high-confidence early warning and intervention commands in extreme hydrological environments, improving the safety and intelligence level of underwater construction of large prefabricated wave barriers.

Claims

1. A simulation and risk warning system for the construction process of a prefabricated wave barrier wall, characterized in that, include: The offline data preprocessing module constructs a hydrodynamic response tensor field based on spatial pose parameters and marine environmental parameters, and calculates the first-order radiation damping perturbation matrix and the second-order added mass perturbation matrix. The environmental perception and state initialization module collects the marine environmental parameters and the initial motion state of the components, and combines them with a preset generalized mass matrix to convert them into initial generalized coordinates and initial generalized momentum. The closed-loop coupled solution module extracts the reference hydrodynamic vector from the hydrodynamic response tensor field based on the marine environmental parameters and the generalized coordinates of the previous time step. It then combines the generalized momentum of the previous time step, the first-order radiation damping perturbation matrix, and the second-order additional mass perturbation matrix to obtain the actual hydrodynamic action vector. The module establishes the Hamiltonian function and the regular equation, and outputs the generalized coordinates and generalized momentum of the current step. The phase space early warning and intervention module constructs a reduced-dimensional phase space state vector based on the current step generalized coordinates and the generalized velocity vector mapped from the current step generalized momentum, generates a phase space dynamic bounding box sequence, and performs an intersection operation with a preset static obstacle bounding box. When the result is not empty, it sends a braking signal to the programmable logic controller.

2. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The spatial pose parameters include the linear translation coordinates of the component's centroid along the X, Y, and Z axes in the global geodetic coordinate system, and the yaw, pitch, and roll angles of the component's rotation around the X, Y, and Z axes in the local connected coordinate system, forming a six-dimensional spatial pose state vector. The marine environmental parameters include the effective wave height, the peak period of the wave spectrum, and the average horizontal current velocity within the construction water depth range extracted from the sea state statistical model, forming a three-dimensional marine environmental parameter vector.

3. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The specific steps for constructing the hydrodynamic response tensor field are as follows: Based on the preset spatial step size, the physical boundary limits of the component translation dimension and the angular limits of yaw angle, pitch angle and roll angle are divided into equidistant or variable distances to generate a set of discrete nodes in spatial pose. Set upper and lower limits for wave height, wave period, and flow velocity to generate a set of discrete environmental nodes; A nine-dimensional global computing network is constructed by performing a Cartesian product operation on the set of discrete nodes in spatial pose and the set of discrete nodes in the environment. The computational fluid dynamics (CFD) engine is invoked, and the spatial coordinates and attitude angles corresponding to the discrete nodes are input into the numerical water tank model as fixed rigid body boundary conditions. Wave height, wave period, and flow velocity values ​​are input as wave inlet boundary conditions and flow field loads. Translational hydrodynamic components and rotational hydrodynamic torque components are extracted and combined to form the reference hydrodynamic vector corresponding to the discrete nodes. The reference hydrodynamic vectors corresponding to all discrete nodes are then summarized to construct the hydrodynamic response tensor field.

4. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The specific steps for calculating the first-order radiation damping perturbation matrix and the second-order additional mass perturbation matrix are as follows: Based on the three-dimensional linear potential flow theory, the boundary element method is used to solve the Laplace equation in the water space domain to obtain the radiation velocity potential distribution of the flow field around the component. The radiation velocity potential is calculated by pressure integration on the wet surface of the component, and the fluid reaction force obtained by integration is decomposed into a component in phase with the component's acceleration and a component in phase with the component's velocity. The components in phase with the component's acceleration are combined according to a six-degree-of-freedom cross-coupling relationship to form the second-order additional mass perturbation matrix, and the components in phase with the component's velocity are combined according to a six-degree-of-freedom cross-coupling relationship to form the first-order radiation damping perturbation matrix.

5. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The process of extracting the baseline hydrodynamic vector is as follows: The marine environmental parameters and the spatial pose parameters corresponding to the generalized coordinates of the previous time step constitute a nine-dimensional target independent variable; The numerical values ​​of the nine-dimensional target independent variables are compared with the preset discrete node sequence in the hydrodynamic response tensor field to determine the lower and upper discrete nodes surrounding the target values. Extract the difference between the target value and the lower limit discrete node value, and calculate the ratio with the interval span between the upper limit discrete node value and the lower limit discrete node value to obtain the one-dimensional interpolation weight coefficient. In the hydrodynamic response tensor field, a hypercube mesh cell that surrounds the nine-dimensional target independent variable is determined. The reference hydrodynamic vector corresponding to the hypercube mesh cell is extracted, and multilinear interpolation is performed using the single-dimensional interpolation weight coefficient to output the current reference hydrodynamic vector.

6. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The specific process for obtaining the actual hydrodynamic action vector is as follows: The difference between the generalized momentum of the previous time step and the generalized momentum retained in the time step before that is calculated and the ratio is calculated with the preset discrete time step parameter to obtain the discrete rate of change vector of the generalized momentum with respect to time. The inverse of the preset generalized mass matrix is ​​multiplied by the generalized momentum of the previous time step and the discrete rate of change vector, respectively, to map them into a generalized velocity vector and a generalized acceleration vector. The radiation damping compensation term is calculated by left-multiplying the generalized velocity vector by the first-order radiation damping perturbation matrix, and the additional mass compensation term is calculated by left-multiplying the generalized acceleration vector by the second-order additional mass perturbation matrix. The radiation damping compensation term and the additional mass compensation term are superimposed on the reference hydrodynamic vector to output the actual hydrodynamic action vector.

7. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The specific steps for establishing the Hamiltonian function and the canonical equation include: The total kinetic energy is defined as half of the product of the transpose of the generalized momentum, the inverse of the preset generalized mass matrix, and the generalized momentum. The total potential energy is defined as the sum of the gravitational potential energy of the component and the elastic deformation potential energy of the cable calculated based on the linear Hooke's law. The total kinetic energy and total potential energy are added together to construct a scalar energy function with generalized coordinates and generalized momentum as independent scalar variables, which is then used as the Hamiltonian function. The first regular equation is set to be the derivative of the generalized coordinate column vector with respect to time, which is exactly the same as the partial derivative of the Hamiltonian function with respect to the generalized momentum. The second regular equation is set to be the same as the sum of the derivative of the generalized momentum column vector with respect to time, the actual hydrodynamic action vector, and the negative values ​​of the partial derivatives of the Hamiltonian function with respect to the generalized coordinate column vector.

8. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The specific steps for constructing the dimension-reduced phase space state vector are as follows: Multiply the current step generalized momentum by the inverse of the preset generalized mass matrix to map it into the generalized velocity vector; The position coordinates along the horizontal X-axis and Y-axis of the global geodetic coordinate system and the yaw angle around the vertical Z-axis are extracted from the current step generalized coordinates to form a dimension-reduced position vector; The translational velocity component and yaw angular velocity component along the horizontal X-axis and Y-axis are extracted from the generalized velocity vector to form a dimension-reduced velocity vector; The reduced position vector and the reduced velocity vector are concatenated in sequence, and the horizontal X-axis coordinate, horizontal Y-axis coordinate, yaw angle, horizontal X-axis velocity, horizontal Y-axis velocity and yaw angle velocity values ​​are written into memory in sequence to form the six-dimensional reduced phase space state vector.

9. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 1, characterized in that, The logic of the intersection operation is as follows: Construct a phase space topological interference index model and calculate the geometric projection distance between each dynamic bounding box in the phase space dynamic bounding box sequence and the preset static obstacle bounding box on the separation axis unit normal vector; The velocity compensation term, obtained by multiplying the phase space velocity sensitivity coefficient by the absolute value of the projection of the relative velocity vector onto the unit normal vector of the separation axis, is subtracted from the geometric projection spacing. The minimum value among all axial differences is extracted as the phase space topological interference index. When the index is not greater than zero, topological interference is determined to have occurred, and the specific predicted time point of interference is output.

10. The prefabricated wave barrier construction process simulation and risk early warning system according to claim 9, characterized in that, The logic for sending the braking signal is as follows: The specific predicted time point output by the intersection operation is defined as the remaining collision time, and compared with the preset warning intervention time threshold and the extreme intervention time threshold. When the remaining collision time is not greater than the warning intervention time threshold but greater than the limit intervention time threshold, a speed limit command message is sent to the hoist frequency converter of the lifting equipment. When the remaining collision time is not greater than the limit intervention time threshold, the minimum braking acceleration is calculated by quotienting the instantaneous composite velocity of the component at the current horizontal plane extracted from the reduced phase space state vector with the effective time window, and an emergency braking control word containing the minimum braking acceleration is written to the programmable logic controller of the lifting equipment.

Images